Next: , Previous: , Up: C API   [Contents][Index]


5.47 libsys

5.47.1 Overview

View lcov test coverage results on http://www.ufoot.org/liquidwar/v6/doc/coverage/src/lib/sys/index.html.

5.47.2 API

Function: int lw6sys_arg_match (const char * keyword, const char * argv_string)

keyword: the option to match, without the prefix "-" or "–"

argv_string: the argv value, for instance argv[1]

This is an utility function which allow the program to handle options in a uniform manner. Key comparison is insensitive, that is, –option and –OPTION are equivalent. Besides, -option and –OPTION are equivalent too. Liquid War 6 documentation mentions options in lowercase with a double dash (–option) by default, but it’s a fact, the program supports variants. This is just for convenience, the philosophy behind this behavior is "be as permissive as possible when interpreting input, and as strict as possible when generating output". In fact, it’s even said that Liquid War 6 will accept the argument without any prefix dash as being valid... This is to say running "liquidwar6 –option" is the same as running "liquidwar6 option". But, this is a secret 8-)

Return value: non zero if it matches, 0 if it doesn’t.

Function: int lw6sys_arg_exists (int argc, const char * [] argv, const char * keyword)

argc: the number of arguments, as passed to main

argv: an array of arguments, as passed to main

keyword: the keyword to match

Parses all command-line arguments, searching for one precise "–key[=...]" entry.

Return value: 1 if key is present, 0 if not.

Function: char * lw6sys_arg_get_value (int argc, const char * [] argv, const char * keyword)

argc: the number of arguments, as passed to main

argv: an array of arguments, as passed to main

keyword: the keyword to match

Parses all command-line arguments, searching for one precise "–key=value" pair, and returns the value.

Return value: a pointer to the value. May be NULL. Must be freed.

Function: char * lw6sys_arg_get_value_with_env (int argc, const char * [] argv, const char * keyword)

argc: the number of arguments, as passed to main

argv: an array of arguments, as passed to main

keyword: the keyword to match

Parses all command-line arguments, searching for one precise "–key=value" pair, and returns the value. If a corresponding environment variable is available, but no command-line parameter was passed, the environment variable is intepreted. Such environment variables are uppercased, prefixed by "LW6_" and "_" replaces "-". The environment variable will be overriden if the command-line parameter is present.

Return value: a pointer to the value. May be NULL. Must be freed.

Function: int lw6sys_arg_test_mode (int argc, const char * [] argv)

argc: argc as passed to main

argv: argv as passed to main

Chooses between the two test modes "check" or "test" and also reports wether one should run "batch" or "interactive" tests. This is done by using the bit mask defined in LW6SYS_TEST_MODE_FULL_TEST and LW6SYS_TEST_MODE_INTERACTIVE.

Return value: a bit mask one can pass to test functions

Function: lw6sys_assoc_t * lw6sys_assoc_new (lw6sys_free_func_t free_func)

free_func: optional callback used to free memory when stored date is a pointer. Can be NULL when one stores non dynamically allocated data, such as an integer or a static array.

Creates an empty assoc. There’s a difference between NULL and an empty assoc. The empty assoc would (in Scheme) be ’() whereas NULL corresponds to undefined "is not a assoc and will generate errors if you ever call assoc functions on it". Such created assoc are not performant hash tables but slowish "strcmp me for each key" associative arrays, the key being a "char *" string and the value a "void *" pointer.

Return value: a pointer to the newly allocated associative array. Must be freed with lw6sys_assoc_free.

Function: void lw6sys_assoc_free (lw6sys_assoc_t * assoc)

assoc: the assoc to be freed.

The function will cascade delete all elements, using (if not NULL...) the callback passed when first creating the assoc.

Return value: void

Function: int lw6sys_assoc_has_key (lw6sys_assoc_t * assoc, const char * key)

assoc: the assoc to test

key: the key to search

Not a very fast function, since on a "big" assoc, strcmp will be called internally until the key is found.

Return value: non-zero if there’s an entry with the corresponding key.

Function: void * lw6sys_assoc_get (lw6sys_assoc_t * assoc, const char * key)

assoc: the assoc to query

key: the key of which we want the value

Return value: a void pointer to the data contained in the assoc. Note that the pointer on the actual data is returned, that is, if it’s static data, you must not try to free it... As long as memory management is concerned, destroying the assoc will actually free the data if needed.

Function: void lw6sys_assoc_set (lw6sys_assoc_t ** assoc, const char * key, void * value)

assoc: the assoc to modify

key: the key we want to updated

value: the new value

Sets a value in an associative array. The key pointer need not be persistent, it can be freed after affectation. In fact a new string will be created internally. This is not true for the value, it’s hard to find way to copy "any object". So if you want an associative array of strings, key can disappear after calling this function, but not value. The function passed as free_func when creating the assoc will be used to free stuff whenever needed (unset or free).

Return value: void

Function: void lw6sys_assoc_unset (lw6sys_assoc_t * assoc, const char * key)

assoc: the assoc concerned

key: the key to unset

Clears an entry in an associative array. The callback passed when creating the assoc will be called if needed, to free the data automatically.

Return value: void

Function: lw6sys_list_t * lw6sys_assoc_keys (lw6sys_assoc_t * assoc)

assoc: the assoc to work on

Returns a list containing all the keys of the assoc. The list must be free with lw6sys_list_free by the caller. This list copies all the keys of the assoc, so it is safe to use it once the assoc is deleted. However the keys will of course be of little interest in this case. But the program won’t segfault.

Return value: the list of keys.

Function: void lw6sys_assoc_map (lw6sys_assoc_t * assoc, lw6sys_assoc_callback_func_t func, void * func_data)

assoc: the assoc to work on

func: a callback to call on each entry

func_data: a pointer on some data which will be passed to the callback

Executes a function on all assoc items. The func_data parameter allows you to pass extra values to the function, such as a file handler or any variable which can not be inferred from list item values, and you of course do not want to make global...

Return value: void

Function: void lw6sys_assoc_sort_and_map (lw6sys_assoc_t * assoc, lw6sys_assoc_callback_func_t func, void * func_data)

assoc: the assoc to work on

func: a callback to call on each entry, may be NULL

func_data: a pointer on some data which will be passed to the callback

Executes a function on all assoc items, like lw6sys_assoc_sort_and_map but befor doing so, sorts all entries in alphabetical order.

Return value: void

Function: lw6sys_assoc_t * lw6sys_assoc_dup (lw6sys_assoc_t * assoc, lw6sys_dup_func_t dup_func)

assoc: the assoc to duplicate, can be NULL

dup_func: the function which will be called to duplicate data

Duplicates an assoc. All keys will be copied so that if the first assoc is deleted, the duplicated one is fine. Additionnaly, dup_func will be called with all data fields. If dup_func is NULL, then data values will simply be copied. This is likely to be usefull when data is not dynamically allocated.

Returned value: a newly allocated assoc.

Function: char * lw6sys_backtrace (int skip, int detailed)

skip: number of calls to skip

detailed: 0 for light output, 1 for complete, detailed messages

Returns the current backtrace as a comma separated list. This can typically be used for debugging purposes. Not available on some platforms, including mingw32, it requires backtrace_symbols to be defined. Note that this function calls internal string functions so it makes usage of the sys module in many ways, therefore should be used only in other modules, it can’t be used for debugging of internal memory functions for instance. To debug those, use backtrace_symbols_fd directly (or maybe just gdb...). The skip parameter allows you to skip caller’s stack, 0 will display everything but lw6sys_backtrace itself.

Return value: dynamically allocated string

Function: int lw6sys_default_memory_bazooka ()

Will set up a default memory bazooka, a slow yet convenient tool to track down and hopefully kill memory leaks. Named bazooka after a night wasted to track down an unfoundable leak... BAZOOOOOOKA!!!

Return value: 1 if success, 0 if failed.

Function: void lw6sys_clear_memory_bazooka ()

Clears the memory bazooka.

Return value: none.

Function: int lw6sys_set_memory_bazooka_size (int size)

size: number of items (calls to malloc) to keep

Resizes, the memory bazooka. What’s this? It’s an inelegant yet efficient tool to track down memory leak. Memory bazooka will keep track of every call to malloc, keeping a trace of what has been malloced, where it has been called (from which file, which line), how much memory was allocated, it will even show you what’s at the address in a 0-terminated string-friendly fashion. Of course this slows down the program, so in production, you might set this to 0, but for debugging, a million bazooka is worth the megabytes and CPU cycles it wastes.

Return value: 1 if success, 0 if failure.

Function: int lw6sys_get_memory_bazooka_size ()

The companion of lw6sys_set_memory_bazooka_size. This function will return how many calls to malloc can be traced. A return value of 0 indicates that feature is disabled.

Return value: size of the bazooka array.

Function: int lw6sys_set_memory_bazooka_eraser (int state)

state: the state of the eraser

Sets the memory bazooka eraser state. Note that to really work, it requires the memory bazooka to be "big enough".

Return value: 1 if activated, 0 if not. Note that the main reason for it not to be activated is if the memory bazooka has zero size.

Function: int lw6sys_get_memory_bazooka_malloc_count ()

Provided you have always called the LW6SYS_MALLOC an LW6SYS_CALLOC to allocate memory, this function will tell you how many times malloc has been called.

Return value: the number of calls to lw6sys_malloc or lw6sys_calloc since program was started.

Function: int lw6sys_get_memory_bazooka_free_count ()

Provided you have always called the LW6SYS_FREE macro to free memory, this function will tell you how many times free has been called.

Return value: the number of calls to lw6sys_free since program was started.

Function: int lw6sys_get_memory_bazooka_malloc_current_count ()

Provided you have always called the LW6SYS_MALLOC an LW6SYS_CALLOC to allocate memory, this function will tell you the current number of pointer returned by LW6SYS_MALLOC an LW6SYS_CALLOC, currently alive on the heap.

Return value: the number of calls to lw6sys_malloc or lw6sys_calloc since program was started.

Function: int lw6sys_get_memory_bazooka_malloc_max_count ()

Provided you have always called the LW6SYS_MALLOC an LW6SYS_CALLOC to allocate memory, this function will tell you the maximum of pointers returned by malloc that were present at the same time on the heap.

Return value: the number of calls to lw6sys_malloc or lw6sys_calloc since program was started.

Function: int lw6sys_get_memory_bazooka_malloc_megabytes ()

Provided you have always called the LW6SYS_MALLOC an LW6SYS_CALLOC to allocate memory, this function will tell you how many bytes malloc has reserved.

Return value: the number of calls to lw6sys_malloc or lw6sys_calloc since program was started.

Function: int lw6sys_get_memory_bazooka_free_megabytes ()

Provided you have always called the LW6SYS_FREE macro to free memory, this function will tell you how many bytes free has freed.

Return value: the number of calls to lw6sys_free since program was started.

Function: int lw6sys_get_memory_bazooka_malloc_current_bytes ()

Provided you have always called the LW6SYS_MALLOC an LW6SYS_CALLOC to allocate memory, this function will tell you the current number of bytes returned by LW6SYS_MALLOC an LW6SYS_CALLOC, currently alive on the heap.

Return value: the number of calls to lw6sys_malloc or lw6sys_calloc since program was started.

Function: int lw6sys_get_memory_bazooka_malloc_max_bytes ()

Provided you have always called the LW6SYS_MALLOC an LW6SYS_CALLOC to allocate memory, this function will tell you the maximum bytes returned by malloc that were present at the same time on the heap.

Return value: the number of calls to lw6sys_malloc or lw6sys_calloc since program was started.

Function: int lw6sys_is_memory_bazooka_trustable ()

Returns true if memory bazooka data are perfectly trustable, that is, it has never been resetted or resized.

Return value: 1 if trustable, 0 if not.

Function: int lw6sys_memory_bazooka_report ()

Reports memory bazooka diagnostics on the console. Carefull, this one is not reentrant, call at the end of your program when all threads are joined.

Return value: 1 if no allocated stuff left, 0 if there are still malloc’ed stuff

Function: char * lw6sys_build_get_package_tarname ()

Returns the name of the package. This is the PACKAGE_TARNAME constant defined by the GNU Autoconf ./configure script. While it’s always possible to use the defined constant directly, using this function will return the value defined when compiling the binary, not the one you’re using when compiling another program relying on Liquid War as a library.

Return value: a non-NULL string "liquidwar6", must not be freed.

Function: char * lw6sys_build_get_package_name ()

Returns the name of the package, in a user friendly form, which can include spaces, for instance. This is the PACKAGE_NAME constant defined by the GNU Autoconf ./configure script. While it’s always possible to use the defined constant directly, using this function will return the value defined when compiling the binary, not the one you’re using when compiling another program relying on Liquid War as a library.

Return value: a non-NULL string "Liquid War 6", must not be freed.

Function: char * lw6sys_build_get_package_string ()

Returns the description of the package. This is the PACKAGE_STRING constant defined by the GNU Autoconf ./configure script. It’s the concatenation of PACKAGE_NAME and VERSION. While it’s always possible to use the defined constant directly, using this function will return the value defined when compiling the binary, not the one you’re using when compiling another program relying on Liquid War as a library.

Return value: a non-NULL string "Liquid War 6 <version>", must not be freed.

Function: char * lw6sys_build_get_package_id ()

Returns the id of the package. This is not an autotools standard ID, in fact it’s just PACKAGE_TARNAME concatenated with VERSION, that is liquidwar6-<version>.

Return value: a non-NULL string "liquidwar6-<version>", must not be freed.

Function: char * lw6sys_build_get_version ()

Returns the version of the program. This is the VERSION constant defined by the GNU Autoconf ./configure script. Same as PACKAGE_VERSION. Note that while using a function to get PACKAGE_TARNAME might seem useless, having both ways to get the version, that is, a function and a constant, is very usefull. Think, for instance, that a dynamically loaded shared library might need to check its own version against the version of the core program.

Return value: a non-NULL string, which must not be freed.

Function: char * lw6sys_build_get_codename ()

Returns the the program codename. This is the little name of the version. It’s been decided that all LW6 releases would take the name of a famous general, warrior, whatever. For instance, it could be "Napoleon".

Return value: a non-NULL string, traditionnally the name of a famous general, someone which has been involved in war. Must not be freed (I mean, the string, not the general).

Function: char * lw6sys_build_get_version_base ()

Returns the program base version number. If version is X.Y.Z, this is X.Y, think of it as MAJOR.MINOR and globally determines the level of compatibility of the program. Two program exposing the same version should be network compatible and also be able to use the same ressource files (graphics, maps, sounds...) as well as being capable of using the same binary modules (graphics backends, bots and so on).

Return value: a non-NULL string, typically "0.1" (beta release) or "6.0" (stable). Must not be freed.

Function: char * lw6sys_build_get_version_major ()

Returns the program major version number. If version is X.Y.Z, this is X. It’s mainly used to make the difference between alpha/beta releases (with "0" here) and stable releases using "6" as we are talking about LW6, after all.

Return value: a non-NULL string, typically "0" (beta release) or "6" (stable). Must not be freed.

Function: char * lw6sys_build_get_version_minor ()

Returns the program minor version number. If version is X.Y.Z, this is Y. This one should increase manually at each significant/public release of the game.

Return value: a non-NULL string like "42", which must not be freed.

Function: char * lw6sys_build_get_stamp ()

Returns the program stamp. This is like a serial number. It’s is not the same as the version. The version is meant to be set to something readable. This is just a cryptic thing, incremented at each ./configure or each developper’s "I feel like it needs to be incremented". The idea is just to keep (one more...) track of which source code is build. Ideally, this would be plugged to the source revision control system but this has some drawbacks, including that it would require it to modify files before commiting them, which is not safe, and almost impossible if you sign archives. One more point: this is a string. It’s true the return value is actually a string containing the representation of an integer, but because all other build parameters are strings, and because we don’t know what the future reserves, it’s a string. If version is X.Y.Z, this is Z. Also called revision.

Return value: a non-NULL string like "666", which must not be freed.

Function: char * lw6sys_build_get_md5sum ()

Returns an md5 checkum which is caculated from C (.c and .h) source files. This is complementary with the build stamp. By default the stamp will be enough to check what has been compiled, but one can always imagine a case where Bob compiles something a little different than Alice, with the same stamp, incremented by 1 from a common source tree. They apply their own patches, for instance. This md5sum double-checks that two binaries have been built from the same sources. Note that this is not the md5 checksum of the generated binary. Nor does it include any information about scheme scripts and data.

Return value: a non-NULL string, which must not be freed.

Function: char * lw6sys_build_get_copyright ()

Returns a (very) short copyright information about the program.

Return value: a non-NULL string, single line whithout ’\n’ at the end. Must not be freed.

Function: char * lw6sys_build_get_license ()

Returns the license for the program (GNU GPL v3 or later).

Return value: a non-NULL string, single line whithout ’\n’ at the end. Must not be freed.

Function: char * lw6sys_build_get_home_url ()

Returns the URL of the game, its homepage.

Return value: a non-NULL string, single line whithout ’\n’ at the end. Must not be freed.

Function: char * lw6sys_build_get_bugs_url ()

Returns the URL for bugs, the bug reports page.

Return value: a non-NULL string, single line whithout ’\n’ at the end. Must not be freed.

Function: char * lw6sys_build_get_configure_args ()

Returns the arguments passed to the GNU Autoconf ./configure script when buildling the game. Very usefull to know how the binary was generated, that is, what kind of optimizations are peculiar settings it uses.

Return value: a non-NULL string, which, passed to ./configure again, would hopefully generate the same binary. Must not be freed.

Function: char * lw6sys_build_get_gcc_version ()

Returns __VERSION__ GCC preprocessor value, that is, the human readable version of the compiler.

Return value: a non-NULL string, must not be freed.

Function: char * lw6sys_build_get_cflags ()

Returns the arguments which would allow another program to use liquidwar6 as a library. Typically, pass this to gcc when compiling your sources. Basically contains "-I" switches which tell where the headers are.

Return value: a non-NULL string, which must not be freed.

Function: char * lw6sys_build_get_ldflags ()

Returns the arguments which would allow another program to link against liquidwar6. Pass this to gcc or libtool when compiling your program. Basically contains a "-L" option which says where the library is. Note that this will only allow you to link against the main libliquidwar6 library, but not the dynamically loaded modules.

Return value: a non-NULL string, which must not be freed.

Function: char * lw6sys_build_get_hostname ()

Returns the value return by the standard shell hostname command on the machine where the game has been built. Usefull to track binaries and know where do they come from.

Return value: a non-NULL string, must not be freed.

Function: char * lw6sys_build_get_date ()

Returns the compilation date. While this information can easily be obtained with the C __DATE__ macro, having this function is convenient for it returns a value which is the same for the whole program, and does not possibly change in every file.

Return value: a non-NULL string, must not be freed.

Function: char * lw6sys_build_get_time ()

Returns the compilation date. While this information can easily be obtained with the C __TIME__ macro, having this function is convenient for it returns a value which is the same for the whole program, and does not possibly change in every file.

Return value: a non-NULL string, must not be freed.

Function: char * lw6sys_build_get_host_cpu ()

Returns the CPU this program is designed for. Convenient on i386 compatible CPUs to know which flavor (i386, i586...) the binary is made for.

Return value: a non-NULL string, must not be freed.

Function: char * lw6sys_build_get_endianness ()

Returns the endianness of the computer.

Return value: ’little’ (x86-like) or ’big’ (ppc-like), as a string. Must not be freed.

Function: int lw6sys_build_get_pointer_size ()

Returns the system pointer size, in bytes.

Return value: 4 for 32-bit, 8 for 64-bit.

Function: int lw6sys_build_is_x86 ()

Tells wether CPU belongs to x86 family or not.

Return value: 1 if x86, 0 if not

Function: char * lw6sys_build_get_host_os ()

Returns the OS this program is designed for. Usefull for bug reports.

Return value: a non-NULL string, must not be freed.

Function: int lw6sys_build_is_gnu ()

Tells wether the program was compiled for a GNU system, or not.

Return value: 1 if compiled on windows, 0 if not

Function: int lw6sys_build_is_unix ()

Tells wether the program was compiled for a UNIX system, or not.

Return value: 1 if compiled on windows, 0 if not

Function: int lw6sys_build_is_ms_windows ()

Tells wether the program was compiled for Microsoft Windows, or not.

Return value: 1 if compiled on windows, 0 if not

Function: int lw6sys_build_is_mac_os_x ()

Tells wether the program was compiled for Mac OS X, or not.

Return value: 1 if compiled on OS X, 0 if not

Function: int lw6sys_build_is_gp2x ()

Tells wether the program was compiled for GP2X, or not.

Return value: 1 if compiled on OS X, 0 if not

Function: char * lw6sys_build_get_top_srcdir ()

Returns the top source directory, when the game was built. This can seem useless and non relevant on the end-user’s machine, but... it’s a must-have for developpers and packagers. Without this, binaries would never find their associated data, especially when building outside the source tree. Or, testing the game would be impossible without installing it, given the fact that most of the code is in scripts that are stored in /usr/local by default, this would be painfull. So this function is here to help finding data within the source tree when the game is not installed yet. Note that the function is rather clever, since it will automatically try to remove useless ’../’ sequences at the beginning of a possibly relative path. Note that the equivalent abs_srcdir function is somewhat more reliable.

Return value: a non-NULL string, must not be freed.

Function: char * lw6sys_build_get_abs_srcdir ()

Returns top_srcdir as an absolute path, this is most of the time more usable than the relative path.

Return value: a non-NULL string, must not be freed.

Function: char * lw6sys_build_get_prefix ()

Returns the prefix value as given to the GNU Autoconf ./configure script. Used to deduce the path to other directories and files.

Return value: a non-NULL string, "/usr/local" by default. Must not be freed.

Function: char * lw6sys_build_get_datadir ()

Returns the datadir value defined by the GNU Autoconf ./configure script. This is not the value which can be overriden by the Liquid War 6 specific. "–data-dir" option. datadir is usually something like "/usr/local/share" while the actual Liquid War 6 defined data dir is a more profound path which includes the name of the package, its version, and so on.

Return value: a non-NULL string, "/usr/local/share" by default. Must not be freed.

Function: char * lw6sys_build_get_libdir ()

Returns the libdir value defined by the GNU Autoconf ./configure script. This is not the value which can be overriden by the Liquid War 6 specific. "–mod-dir" option. libdir is usually something like "/usr/local/lib" while the actual Liquid War 6 defined module dir is a more profound path which includes the name of the package, its version, and so on.

Return value: a non-NULL string, "/usr/local/lib" by default. Must not be freed.

Function: char * lw6sys_build_get_includedir ()

Returns the includedir value defined by the GNU Autoconf ./configure script. As for other options, it’s interesting to have this value, this enables the program to inform people who want to hack the game of the place headers are supposed to be installed.

Return value: a non-NULL string, "/usr/local/include" by default. Must not be freed.

Function: char * lw6sys_build_get_localedir ()

Returns the localedir value defined by the GNU Autoconf ./configure script. Used as an argument for gettext / libintl functions.

Return value: a non-NULL string, "/usr/local/share/locale" by default. Must not be freed.

Function: char * lw6sys_build_get_docdir ()

Returns the docdir value defined by the GNU Autoconf ./configure script. Used to write consistent XML file headers.

Return value: a non-NULL string, "/usr/local/share/doc/liquidwar6" by default. Must not be freed.

Function: char * lw6sys_build_get_enable_console ()

Tells wether console is enabled or not.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_gtk ()

Tells wether gtk is enabled or not.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_mod_gl1 ()

Tells wether the graphical mod-gl1 backend was compiled.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_mod_gles2 ()

Tells wether the graphical mod-gles2 backend was compiled.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_mod_soft ()

Tells wether the graphical mod-soft backend was compiled.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_mod_caca ()

Tells wether the graphical mod-caca backend was compiled.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_mod_csound ()

Tells wether the audio mod-csound backend was compiled.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_mod_ogg ()

Tells wether the audio mod-ogg backend was compiled.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_mod_http ()

Tells wether the network mod-http backend was compiled.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_openmp ()

Tells wether the game was compiled with openmp support.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_optimize ()

Tells wether the game was compiled in optimize mode.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_allinone ()

Tells wether the game was compiled in allinone mode.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_fullstatic ()

Tells wether the game was compiled in fullstatic mode.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_paranoid ()

Tells wether the game was compiled with paranoid memory management.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_gprof ()

Tells wether the game was compiled with suitable informations for gprof.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_instrument ()

Tells wether the game was compiled with the ’-finstrument-fonctions’ GCC flag.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_profiler ()

Tells wether the game was compiled for later use with Google Profiler support.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_gcov ()

Tells wether the game was compiled with suitable informations for gcov.

Return value: "yes" or "no", must no be freed.

Function: char * lw6sys_build_get_enable_valgrind ()

Tells wether the game was compiled for later use with valgrind.

Return value: "yes" or "no", must no be freed.

Function: int lw6sys_build_get_bin_id ()

Returns the internal bin-id value, which does not mean anything but changes at each build.

Return value: an integer

Function: void lw6sys_build_log_all ()

Dumps in the log file the whole program pedigree, host, modules, that is, what are the values of all the build options. Usefull for bug reports.

Return value: none.

Function: lw6sys_cache_t * lw6sys_cache_new (lw6sys_free_func_t free_func, int size, int delay_msec)

free_func: optional callback used to free memory when stored date is a pointer. Can be NULL when one stores non dynamically allocated data, such as an integer or a static array.

size: the estimated size of the cache table. This is required because, internally, the object uses a hash. Note that this is an estimation only. You could theorically fit 1000000 objects in a 3-sized cache. Problem -> this is inefficient, you’d better use an assoc or a bigger cache. If you store 3 elements in a 1000000-sized cache, you’ll waste memory. It might be wise to use a prime number as the estimated size. 421 is prime ;)

Creates an empty cache. There’s a difference between NULL and an empty cache.

Return value: a pointer to the newly allocated cache table. Must be freed with lw6sys_cache_free.

Function: void lw6sys_cache_free (lw6sys_cache_t * cache)

cache: the cache to be freed.

The function will cascade delete all elements, using (if not NULL...) the callback passed when first creating the cache.

Return value: void

Function: void lw6sys_cache_free_callback (void * data)

data: data to free, this is normally an cache item

This is a wrapper, used as the actual free callback for the internal hash. What it does is that it just runs the real free callback (the one given by the user) on the member value. This indirection is required since we use the intermediate item object to store the timestamp along with the key and data.

Return value: none.

Function: int lw6sys_cache_has_key (lw6sys_cache_t * cache, const char * key)

cache: the cache to test

key: the key to search

Tells wether the key is present or not. If key is here but too old (expired) then will return 0 and key will be deleted on the fly.

Return value: non-zero if there’s an entry with the corresponding key.

Function: void * lw6sys_cache_get (lw6sys_cache_t * cache, const char * key)

cache: the cache to query

key: the key of which we want the value

Gets the value corresponding to a given key. Note that the value might be NULL, even if the key exists. If the key has expired, NULL will be returned, and the entry deleted on the fly.

Return value: a void pointer to the data contained in the cache. Note that the pointer on the actual data is returned, that is, if it’s static data, you must not try to free it... As long as memory management is concerned, destroying the cache will actually free the data if needed.

Function: void lw6sys_cache_set (lw6sys_cache_t * cache, const char * key, void * value)

cache: the cache to modify

key: the key we want to updated

value: the new value

Sets a value in a cache table. The key pointer need not be persistent, it can be freed after affectation. In fact a new string will be created internally. This is not true for the value, it’s hard to find way to copy "any object". So if you want a cache of strings, key can disappear after calling this function, but not value. The function passed as free_func when creating the cache will be used to free stuff whenever needed (unset or free).

Return value: void

Function: void lw6sys_cache_unset (lw6sys_cache_t * cache, const char * key)

cache: the cache concerned

key: the key to unset

Clears an entry in a cache table. The callback passed when creating the cache will be called if needed, to free the data automatically.

Return value: void

Function: u_int32_t lw6sys_checksum (unsigned char * data, int len)

data: the data to process

len: the length, in bytes, of the data to process

Creates a checksum from a byte array. This could be mapped on any standard CRC-32 and/or MD5 algorithm, but licence issues for those are such a headache that for the sake of simplicity, it’s wrapped here. In LW6 context, we do not really really fear any attack for these checksums are used internally to track bugs and check, for instance, that two game states are actually the same on two distant computers in a network game. Data encryption and security of network links is another debate. Additionnally, this function returns an integer, easier to handle in standard C than any malloc’ed stuff.

Return value: the checksum, as an integer.

Function: u_int32_t lw6sys_checksum_str (const char * value)

value: the string to process

Creates a checksum from a string. This is a convenience function to save the programmer the hassle of calling strlen before any checksum calculation.

Return value: the checksum, as an integer.

Function: u_int32_t lw6sys_checksum_int32 (u_int32_t value)

value: the integer to process

Creates a checksum from an integer. This is a convenience function to save the programmer the hassle of passing a pointer to the integer with the size of it each time there’s a checksum to do. Additionnally, with this one you can pass an int8 or an int16, and function will work just the same indenpendantly of endianness.

Return value: the checksum, as an integer.

Function: u_int32_t lw6sys_checksum_int64 (u_int64_t value)

value: the integer to process

Creates a checksum from an integer. This is a convenience function to save the programmer the hassle of passing a pointer to the integer with the size of it each time there’s a checksum to do. This function handles 64-bit long long integers..

Return value: the checksum, as an integer.

Function: u_int32_t lw6sys_checksum_whd (lw6sys_whd_t * whd)

whd: a pointer to the wh struct to be processed

Creates a checksum from the given structure. Convenience function to save the hassle of passing a pointer to and the size of the lw6sys_wh_t struct each time, knowing that there are very often checksums calculated on it. Also avoids endianess issues.

Return value: the checksum, as an integer.

Function: u_int32_t lw6sys_checksum_xyz (lw6sys_xyz_t * xyz)

xyz: a pointer to the xy struct to be processed

Creates a checksum from the given structure. Convenience function to save the hassle of passing a pointer to and the size of the lw6sys_xy_t struct each time, knowing that there are very often checksums calculated on it. Also avoids endianess issues.

Return value: the checksum, as an integer.

Function: void lw6sys_checksum_update (u_int32_t * checksum, unsigned char * data, int len)

checksum: a pointer to the previous checksum

data: the data to process

len: the length, in bytes, of the data to process

Creates a checksum from the given data. The difference with lw6sys_checksum is that this one updates an existing checksum, thus enabling the programmer to call it sequentially and get a global checksum on different sources.

Return value: none.

Function: void lw6sys_checksum_update_str (u_int32_t * checksum, const char * value)

checksum: a pointer to the previous checksum

value: the string to process

Creates a checksum from the given string. The difference with lw6sys_checksum_str is that this one updates an existing checksum, thus enabling the programmer to call it sequentially and get a global checksum on different sources.

Return value: none.

Function: void lw6sys_checksum_update_int32 (u_int32_t * checksum, int32_t value)

checksum: a pointer to the previous checksum

value: the integer to process

Creates a checksum from the given integer. The difference with lw6sys_checksum_int32 is that this one updates an existing checksum, thus enabling the programmer to call it sequentially and get a global checksum on different sources.

Return value: none.

Function: void lw6sys_checksum_update_int64 (u_int32_t * checksum, int64_t value)

checksum: a pointer to the previous checksum

value: the integer to process

Creates a checksum from the given integer. The difference with lw6sys_checksum_int64 is that this one updates an existing checksum, thus enabling the programmer to call it sequentially and get a global checksum on different sources.

Return value: none.

Function: void lw6sys_checksum_update_whd (u_int32_t * checksum, const lw6sys_whd_t * whd)

checksum: a pointer to the previous checksum

whd: a pointer to the wh struct to be processed

Creates a checksum from the given structure. The difference with lw6sys_checksum_whd is that this one updates an existing checksum, thus enabling the programmer to call it sequentially and get a global checksum on different sources.

Return value: none.

Function: void lw6sys_checksum_update_xyz (u_int32_t * checksum, const lw6sys_xyz_t * xyz)

checksum: a pointer to the previous checksum

xyz: a pointer to the xy struct to be processed

Creates a checksum from the given structure. The difference with lw6sys_checksum_xyz is that this one updates an existing checksum, thus enabling the programmer to call it sequentially and get a global checksum on different sources.

Return value: none.

Function: u_int8_t lw6sys_color_float2char (float f)

f: the value to convert, from 0.0f to 1.0f

Converts a floating point value between 0.0f and 1.0f to its 8-bit equivalent between 0 and 255. Usefull in color conversion.

Return value: an integer between 0 and 255.

Function: float lw6sys_color_char2float (u_int8_t i)

i: the value to convert, from 0 to 255

Converts an 8-bit value between 0 and 255 to its floating-point equivalent between 0.0f and 1.0f. Usefull in color conversion.

Return value: a float between 0.0f and 1.0f.

Function: lw6sys_color_8_t lw6sys_color_f_to_8 (const lw6sys_color_f_t * color_f)

color_f: the color to convert

Converts a color from floating point format to the integer "0 to 255" common format. All fields (RGBA) are converted.

Return value: the color in 8-bit format.

Function: void lw6sys_color_8_to_f (lw6sys_color_f_t * color_f, lw6sys_color_8_t color_8)

color_f: the converted color (pointer must point to writable memory)

color_8: the color to convert

Converts a color from the integer "0 to 255" common format to floating point format. All fields (RGBA) are converted.

Return value: none.

Function: u_int32_t lw6sys_color_f_to_irgba (const lw6sys_color_f_t * color_f)

color_f: the color to convert

Converts a color from floating point format to a single integer, where all fields (RGBA) are serialized. This serialization is endianess independant. Could be used directly by low-level libraries such as SDL.

Return value: the color serialized in an integer.

Function: u_int32_t lw6sys_color_f_to_ibgra (const lw6sys_color_f_t * color_f)

color_f: the color to convert

Converts a color from floating point format to a single integer, where all fields (BGRA) are serialized. This serialization is endianess independant. Could be used directly by low-level libraries such as SDL.

Return value: the color serialized in an integer.

Function: u_int32_t lw6sys_color_f_to_iargb (const lw6sys_color_f_t * color_f)

color_f: the color to convert

Converts a color from floating point format to a single integer, where all fields (ARGB) are serialized. This serialization is endianess independant. Could be used directly by low-level libraries such as SDL.

Return value: the color serialized in an integer.

Function: u_int32_t lw6sys_color_f_to_iabgr (const lw6sys_color_f_t * color_f)

color_f: the color to convert

Converts a color from floating point format to a single integer, where all fields (ABGR) are serialized. This serialization is endianess independant. Could be used directly by low-level libraries such as SDL.

Return value: the color serialized in an integer.

Function: u_int32_t lw6sys_color_8_to_irgba (lw6sys_color_8_t color_8)

color_8: the color to convert

Converts a color from common "0 to 255" structured format to a single integer, where all fields (RGBA) are serialized. This serialization is endianess independant. Could be used directly by low-level libraries such as SDL.

Return value: the color serialized in an integer.

Function: u_int32_t lw6sys_color_8_to_ibgra (lw6sys_color_8_t color_8)

color_8: the color to convert

Converts a color from common "0 to 255" structured format to a single integer, where all fields (BGRA) are serialized. This serialization is endianess independant. Could be used directly by low-level libraries such as SDL.

Return value: the color serialized in an integer.

Function: u_int32_t lw6sys_color_8_to_iargb (lw6sys_color_8_t color_8)

color_8: the color to convert

Converts a color from common "0 to 255" structured format to a single integer, where all fields (ARGB) are serialized. This serialization is endianess independant. Could be used directly by low-level libraries such as SDL.

Return value: the color serialized in an integer.

Function: u_int32_t lw6sys_color_8_to_iabgr (lw6sys_color_8_t color_8)

color_8: the color to convert

Converts a color from common "0 to 255" structured format to a single integer, where all fields (ABGR) are serialized. This serialization is endianess independant. Could be used directly by low-level libraries such as SDL.

Return value: the color serialized in an integer.

Function: void lw6sys_color_irgba_to_f (lw6sys_color_f_t * color_f, u_int32_t color_i)

color_f: the converted color (point must point to writable memory)

color_i: the color to convert

Converts a color from a serialized integer format (RGBA) to a floating point structure.

Return value: none.

Function: void lw6sys_color_ibgra_to_f (lw6sys_color_f_t * color_f, u_int32_t color_i)

color_f: the converted color (point must point to writable memory)

color_i: the color to convert

Converts a color from a serialized integer format (BGRA) to a floating point structure.

Return value: none.

Function: void lw6sys_color_iargb_to_f (lw6sys_color_f_t * color_f, u_int32_t color_i)

color_f: the converted color (point must point to writable memory)

color_i: the color to convert

Converts a color from a serialized integer format (ARGB) to a floating point structure.

Return value: none.

Function: void lw6sys_color_iabgr_to_f (lw6sys_color_f_t * color_f, u_int32_t color_i)

color_f: the converted color (point must point to writable memory)

color_i: the color to convert

Converts a color from a serialized integer format (ABGR) to a floating point structure.

Return value: none.

Function: lw6sys_color_8_t lw6sys_color_irgba_to_8 (u_int32_t color_i)

color_i: the color to convert

Converts a color from a serialized integer format (RGBA) to a "0 to 255" based structure.

Return value: the converted color (structure).

Function: lw6sys_color_8_t lw6sys_color_ibgra_to_8 (u_int32_t color_i)

color_i: the color to convert

Converts a color from a serialized integer format (BGRA) to a "0 to 255" based structure.

Return value: the converted color (structure).

Function: lw6sys_color_8_t lw6sys_color_iargb_to_8 (u_int32_t color_i)

color_i: the color to convert

Converts a color from a serialized integer format (ARGB) to a "0 to 255" based structure.

Return value: the converted color (structure).

Function: lw6sys_color_8_t lw6sys_color_iabgr_to_8 (u_int32_t color_i)

color_i: the color to convert

Converts a color from a serialized integer format (ABGR) to a "0 to 255" based structure.

Return value: the converted color (structure).

Function: lw6sys_color_8_t lw6sys_color_a_to_8 (const char * ascii)

ascii: the color to convert

Converts a color from a human readable string to a "0 to 255" based structure. The string must be of the form "#RRGGBBAA" or "#RGB", in a general manner any HTML-valid value should work.

Return value: the converted color (structure).

Function: void lw6sys_color_a_to_f (lw6sys_color_f_t * color_f, const char * ascii)

color_f: the converted color (pointer must point to writable memory)

ascii: the color to convert

Converts a color from a human readable string to a float based structure. The string must be of the form "#RRGGBBAA" or "#RGB", in a general manner any HTML-valid value should work.

Return value: none.

Function: char * lw6sys_color_8_to_a (lw6sys_color_8_t color_8)

color_8: the color to convert

Converts a color from a "0 - 255" integer based structure to its readable form "#RRGGBBAA". If alpha is 255 (0xFF), that is, if it’s opaque, then the "AA" part is ommitted.

Return value: a newly allocated string.

Function: void lw6sys_color_rgb_to_hsv (lw6sys_color_hsv_t * color_hsv, lw6sys_color_8_t color_8)

color_hsv: the target color, in HSV format

color_8: the source color, in RGB 256 format

Converts from HSV to RGB. Usefull for color manipulation, since most colors are stored in RGB but HSV is convenient for transformation. Alpha layer is kept as is.

Return value: none.

Function: lw6sys_color_8_t lw6sys_color_hsv_to_rgb (const lw6sys_color_hsv_t * color_hsv)

color_hsv: the source color, in HSV format

Converts from RGB to HSV. Usefull to make colors transformed in HSV format usable again by all display routines, which consume RGB. Alpha layer is kept as is.

Return value: the RGB color.

Function: void lw6sys_color_hsv_invert (lw6sys_color_hsv_t * color_hsv, int invert_h, int invert_s, int invert_v)

color_hsv: the source color, in HSV format

invert_h: wether to invert the hue

invert_s: wether to invert the saturation

invert_v: wether to invert the value

Inverts an HSV color, calling it with 1,0,0 the color will become a color with opposite hue but same saturation and same value.

Return value: none.

Function: int lw6sys_color_is_grey (lw6sys_color_8_t color)

color: the color to test

Tells wether a color is pure grey or not. This is interesting for such colors have no hue and sometimes need special handling.

Return value: 1 if grey, 0 if colored

Function: lw6sys_color_8_t lw6sys_color_average (int size, const lw6sys_color_8_t * colors)

size: number of the color array (number of items)

colors: the colors to compute

Tries to find out the "average" color from an array of colors. The algorithm is far from perfect, but should output a color which reflects the colors passed in.

Return value: the (inexact) average color.

Function: lw6sys_color_8_t lw6sys_color_ponderate (lw6sys_color_8_t color1, lw6sys_color_8_t color2, float coeff)

color1: first color

color2: second color

coeff: the ponderation coefficient

Tries to find a color between the two colors passed as an argument. The coefficient can be used, to set the relative weight of each color. Using 0 will return color1, 1 will return color2 and 0.5 will make an average between the two colors. Any value between 0 and 1 can be used.

Return value: the (inexact) ponderated color.

Function: float lw6sys_color_distance (lw6sys_color_8_t color1, lw6sys_color_8_t color2)

color1: first color

color2: second color

Calculates the distance between two colors. The unit is arbitrary, a big value means "colors are different", 0 means they are the same. A distance of 1 corresponds to colors which have barely anything in common, but the result can still be greater than 1. Alpha layer is not taken in account.

Return value: the distance.

Function: int lw6sys_color_is_same (lw6sys_color_8_t color1, lw6sys_color_8_t color2)

color1: the first color to compare

color2: the second color to compare

Compares two colors.

Return value: 1 if they are the same, 0 if not.

Function: void lw6sys_color_8_solid (lw6sys_color_8_t * color)

color: the color to modify

Make a color "solid" that is make it not transparent at all.

Return value: none.

Function: void lw6sys_color_f_solid (lw6sys_color_f_t * color)

color: the color to modify

Make a color "solid" that is make it not transparent at all.

Return value: none.

Function: int lw6sys_atoi (const char * str)

str: string to convert

Just a plain wrapper on atoi, it’s here for API consistency. Will check if str is NULL (and in this case return 0).

Return value: an integer.

Function: int64_t lw6sys_atoll (const char * str)

str: string to convert

Wrapper on atoll, it’s here for API consistency. Will check if str is NULL (and in this case return 0).

Return value: a 64-bit integer.

Function: int lw6sys_atob (const char * str)

str: string to convert

Transform a string into a boolean value. Accepts "0"/"1" in input, but also y/n, yes/no, true/false, on/off. Will check if str is NULL (and in this case return 0).

Return value: an integer, 0 or 1.

Function: float lw6sys_atof (const char * str)

str: string to convert

A wrapper on atof, makes sure the locale used is C (default) and won’t change the decimal separator whatsoever. Usefull for serialization for instance. Will check if str is NULL (and in this case return 0).

Return value: a float.

Function: char * lw6sys_itoa (int value)

value: the integer to convert

Converts an integer to a string, the advantage of this function is it allocates memory, and does the dirty job.

Return value: a newly allocated pointer, must be freed, may be NULL.

Function: char * lw6sys_lltoa (int64_t value)

value: the integer to convert

Converts a 64-bit integer to a string, the advantage of this function is it allocates memory, and does the dirty job.

Return value: a newly allocated pointer, must be freed, may be NULL.

Function: char * lw6sys_btoa (int value)

value: the boolean to convert

Converts a boolean to a string, the advantage of this function is it allocates memory, and does the dirty job.

Return value: a newly allocated pointer, must be freed, may be NULL.

Function: char * lw6sys_ftoa (float value)

value: the float to convert

Converts a float to a string, the advantage of this function is it allocates memory, and does the dirty job.

Return value: a newly allocated pointer, must be freed, may be NULL.

Function: int lw6sys_cunit_run_tests (int mode)

mode: mode passed to program (bit mask)

Run all registered suites and their tests, will interpret mode and call the right CUnit function (Batch, Console, NCurses...).

Return value: 1 if tests or OK, 0 if not.

Function: char * lw6sys_daemon_pid_file (int argc, const char * [] argv)

argc: argc as passed to main

argv: argv as passed to main

Get the default pid file, used to lock daemon and avoid 2 daemons running at the same time.

Return value: newly allocated string

Function: int lw6sys_daemon_start (char * pid_file)

pid_file: the pid file used for the daemon

Calls fork internally to put the process in the program, make it a daemon. Note this won’t work on all platforms, for instance it won’t work on MS-Windows but this is rarely an issue as MS-Windows users are rarely concerned with detaching a program from a tty. Note that this isn’t a wrapper on fork, the return value is different,

Return value: 1 on success, 0 on failure.

Function: int lw6sys_daemon_stop (char * pid_file)

pid_file: the pid file used for the daemon

Removes the daemon pid file. Can be called safely even if daemon wasn’t started.

Return value: 1 on success, 0 on failure

Function: int lw6sys_debug_get ()

Gets the debug mode.

Function: void lw6sys_debug_set (int mode)

mode: the debug mode, 1 if set, 0 if not.

Sets the debug mode.

Function: void lw6sys_dump_clear (char * user_dir)

user_dir: the user directory, where user can write data.

Clears the dump file. That is, resets it to a "0 byte" file.

Return value: none.

Function: int lw6sys_dump (char * user_dir, char * content)

user_dir: the user directory, where user can write data.

content: the content to be written in the dump file.

Writes the dump file onto the disk. The dump is used for special error messages which do not really fit in the standard log, and require a special treatment. In pratice, it’s used to log fatal script (Guile) errors.

Return value: 1 if success, 0 if failure.

Function: char lw6sys_env_separator_char ()

Gets the ENV separator, that is, for instance, the character used to separate paths in environment variables. Typically, this would be ":" on GNU and ";" on Microsft platforms.

Return value: the ascii character code.

Function: char * lw6sys_env_separator_str ()

Gets the ENV separator, that is, for instance, the character used to separate paths in environment variables. Typically, this would be ":" on GNU and ";" on Microsft platforms.

Return value: a pointer to a single 0-terminated character string which contains the character. Must not be freed.

Function: char * lw6sys_env_concat (const char * value1, const char * value2)

value1: the left part to be concatenated

value2: the right part to be concatenated

Concatenates two values and puts the ENV separator, as returned by lw6sys_env_separator_char between them.

Return value: the concatenated string, must be freed.

Function: int lw6sys_env_exists_prefixed (const char * keyword)

keyword: the keyword to be searched in the environment variables.

Searches environment variables for the given keyword. The keyword will be fixed so that all dashes "-" characters are replaced by underscores "_" characters. Characters will be changed to uppercase. Any non alphanumeric character will be replaced by "_". Finally, an "LW6_" prefix will be added. That is to say, calling this function with "my-param" will search for the "LW6_MY_PARAM" environment variable.

Return value: 1 if the environment variable exists, 0 if not.

Function: char * lw6sys_getenv (const char * key)

key: the environment variable to get.

Searches environment variables for the given value. This is a wrapper over the standard C getenv, the difference is it will return a dynamically allocated pointer, and on some platforms will query specific OS functions.

Return value: the value for the given keyword. May be NULL. Must be freed.

Function: char * lw6sys_getenv_prefixed (const char * keyword)

keyword: the keyword to be searched in the environment variables.

Searches environment variables for the given value. The keyword will be fixed so that all dashes "-" characters are replaced by underscores "_" characters. Characters will be changed to uppercase. Any non alphanumeric character will be replaced by "_". Finally, an "LW6_" prefix will be added. That is to say, calling this function with "my-param" will search for the "LW6_MY_PARAM" environment variable.

Return value: the value for the given keyword. May be NULL. Must be freed.

Function: int lw6sys_setenv (const char * keyword, const char * value)

keyword: the environment variable to set

value: the value of the environment variable to set

Sets the environment variable to a given value. If value is NULL, variable is unset. Note that unlike lw6sys_getenv_prefixed, this function does not transform the keyword into "LW6_..." before setting the value, so it’s your responsability to call "lw6sys_keyword_as_env" if needed.

Return value: 1 if success, 0 if failed

Function: int lw6sys_setenv_prefixed (const char * keyword, const char * value)

keyword: the keyword to be searched in the environment variables.

value: the value of the environment variable to set

Sets the environment variable to the given value. The keyword will be fixed so that all dashes "-" characters are replaced by underscores "_" characters. Characters will be changed to uppercase. Any non alphanumeric character will be replaced by "_". Finally, an "LW6_" prefix will be added. That is to say, calling this function with "my-param" will set the "LW6_MY_PARAM" environment variable.

Return value: 1 if success, 0 if failure

Function: lw6sys_list_t * lw6sys_env_split (const char * value)

value: the value, a list of item separated by... the separator

Splits the environment value into a list of strings containing each element. All strings are dynamically allocated, but they will be freed automatically when the list is freed.

Return value: a list of strings.

Function: char * lw6sys_get_home ()

Gets the home directory of the user. Used internally to calculate the user-dir value. Note that Liquid War 6, by default, never stores files under ’$HOME’, instead it put things in ’$HOME/.liquidwar6’, that is ’user-dir’. If the environment variable ’HOME’ is not set, will return ’.’.

Return value: a newly allocated pointer, must be freed.

Function: char * lw6sys_get_username ()

Gets the name of the current user. Difference with the standard function getlogin is that this function will returned a dynamically allocated pointer, and provide a default value if it’s undefined. Also, if will look at the content of the ’LOGNAME’ environment variable if needed, and will even provide a default value.

Return value: a newly allocated pointer, must be freed.

Function: char * lw6sys_get_hostname ()

Gets the name of the current host. The name of the computer. Might not work perfectly, this function is just used to provide default values for player names and such things.

Return value: a newly allocated pointer, must be freed.

Function: char * lw6sys_escape_http_uri (const char * src)

src: the string to escape

Transforms a string so that it does not contain any non-valid URL chars, it will mostly convert chars over 128 into their an hexadecimal code which replaces them in URLs. Note that this function is non really standard compliant for it won’t encode ’%’ but keep it the same. This is to allow using it several times on the same string and avoid double-triple encoding of ’%’. In practice it’s not recommended to have public_url for nodes with ’%’ in them, and the program will never generate such url when guessing urls.

Return value: newly allocated string.

Function: char * lw6sys_escape_html_attribute (const char * src)

src: the string to escape

Transforms a string so that it can fit in a html field, this is typically for alt="" or title="" fields so it will convert a double quote into its equivalent escaped char.

Return value: newly allocated string.

Function: char * lw6sys_escape_sql_value (const char * src)

src: the string to escape

Transforms a string so that it can fit as an SQL parameter, it will get rid URL chars, it will mostly convert chars over 128 into an hexadecimal form which replaces them in URLs.

Return value: newly allocated string.

Function: char * lw6sys_exec_find_myself (int argc, const char * [] argv)

argc: number of args as passed to main

argv: array of args as passed to main

Finds the path of the program currently run, this is typically to pass it to lw6sys_exec_again and run it again.

Return value: the path (newly allocated string).

Function: int lw6sys_is_executed_again (int argc, const char * [] argv)

argc: number of args as passed to main

argv: array of args as passed to main

Tells wether the program is already executed by itself by lw6sys_exec_again function. Based on environment and command switches.

Return value: 1 if executed again, 0 if not.

Function: int lw6sys_exec_again (int argc, const char * [] argv)

argc: number of args as passed to main

argv: array of args as passed to main

Runs the program from itsef, that is fires a new program (the same running) and ends up the current one. This is used to fix some environment variable issues. If LW6_EXECUTED_AGAIN (environment variable) is set, will not run the program so this is not really like exec as in the C standard library, this function will actually return and be successfull even if no other process was started. It’s just designed to bootstrap/launch the process once.

Return value: 1 on success, 0 on failure (always fail)

Function: int lw6sys_exec_restart (int argc, const char * [] argv)

argc: number of args as passed to main

argv: array of args as passed to main

Restart the program with exactly the same arguments it was given the first time.

Return value: 1 on success, 0 on failure (always fail)

Function: int lw6sys_clear_file (const char * filename)

filename: absolute or relative filename

Clears a file, that is, make it a 0 byte file, empty, ready to be filled if needed. If this function is called successfully, program can reasonnably assume file will be writable during its execution.

Return value: 1 if success, 0 if failure.

Function: char * lw6sys_read_file_content (const char * filename)

filename: absolute or relative filename

Reads the content of a file, and returns it as a string. Note that content might or might not be ascii or binary, the function will however put a tailing 0 character at the end so that low-level standard C functions do not segfault when used with the returned value.

Return value: a newly allocated pointer, must be freed.

Function: void * lw6sys_read_file_content_bin (int * filesize, const char * filename)

filesize: will contain the file size, in bytes

filename: absolute or relative filename

Reads the content of a file, and returns it as a binary buffer. Even if not ascii or binary, the function will however put a tailing 0 character at the end so that low-level standard C functions do not segfault when used with the returned value. This 0 character is not included in filesize so if there are 4 bytes in the file the 5 bytes will be allocated, this is just for string functions not to explode if called by accident. The filesize can be NULL, in that case function is just like the lw6sys_read_file_content function.

Return value: a newly allocated pointer, must be freed.

Function: int lw6sys_write_file_content (const char * filename, const char * content)

filename: absolute or relative filename

content: the content to be written.

Writes the content into the file. Content is assumed to be a string, function will segfault if it’s not correctly 0 terminated as in C string convention. So this function will not allow you to write down arbitrary binary data, however LW6 uses mostly text files to store information, and opaque binary data usage is not recommended.

Function: lw6sys_hash_t * lw6sys_hash_new (lw6sys_free_func_t free_func, int size)

free_func: optional callback used to free memory when stored date is a pointer. Can be NULL when one stores non dynamically allocated data, such as an integer or a static array.

size: the estimated size of the hash table. Note that this is an estimation only. You could theorically fit 1000000 objects in a 3-sized hash. Problem -> this is inefficient, you’d better use an assoc or a bigger hash. If you store 3 elements in a 1000000-sized hash, you’ll waste memory. It might be wise to use a prime number as the estimated size. 421 is prime ;)

Creates an empty hash. There’s a difference between NULL and an empty hash.

Return value: a pointer to the newly allocated hash table. Must be freed with lw6sys_hash_free.

Function: void lw6sys_hash_free (lw6sys_hash_t * hash)

hash: the hash to be freed.

The function will cascade delete all elements, using (if not NULL...) the callback passed when first creating the hash.

Return value: void

Function: int lw6sys_hash_has_key (lw6sys_hash_t * hash, const char * key)

hash: the hash to test

key: the key to search

Tells wether the key is present or not.

Return value: non-zero if there’s an entry with the corresponding key.

Function: void * lw6sys_hash_get (lw6sys_hash_t * hash, const char * key)

hash: the hash to query

key: the key of which we want the value

Gets the value corresponding to a given key. Not that the value can be NULL, even if the key exitsts.

Return value: a void pointer to the data contained in the hash. Note that the pointer on the actual data is returned, that is, if it’s static data, you must not try to free it... As long as memory management is concerned, destroying the hash will actually free the data if needed.

Function: void lw6sys_hash_set (lw6sys_hash_t * hash, const char * key, void * value)

hash: the hash to modify

key: the key we want to updated

value: the new value

Sets a value in a hash table. The key pointer need not be persistent, it can be freed after affectation. In fact a new string will be created internally. This is not true for the value, it’s hard to find way to copy "any object". So if you want a hash table of strings, key can disappear after calling this function, but not value. The function passed as free_func when creating the hash will be used to free stuff whenever needed (unset or free).

Return value: void

Function: void lw6sys_hash_unset (lw6sys_hash_t * hash, const char * key)

hash: the hash concerned

key: the key to unset

Clears an entry in a hash table. The callback passed when creating the hash will be called if needed, to free the data automatically.

Return value: void

Function: lw6sys_list_t * lw6sys_hash_keys (lw6sys_hash_t * hash)

hash: the hash to work on

Returns a list containing all the keys of the hash. The list must be free with lw6sys_list_free by the caller. This list copies all the keys of the hash, so it is safe to use it once the hash is deleted. However the keys will sometimes be of little interest in this case. But the program won’t segfault.

Return value: the list of keys.

Function: void lw6sys_hash_map (lw6sys_hash_t * hash, lw6sys_assoc_callback_func_t func, void * func_data)

hash: the hash to work on

func: a callback to call on each entry

func_data: a pointer on some data which will be passed to the callback

Executes a function on all hash items. The func_data parameter allows you to pass extra values to the function, such as a file handler or any variable which can not be inferred from list item values, and you of course do not want to make global...

Return value: void

Function: void lw6sys_hash_sort_and_map (lw6sys_hash_t * hash, lw6sys_assoc_callback_func_t func, void * func_data)

hash: the hash to work on

func: a callback to call on each entry, may be NULL

func_data: a pointer on some data which will be passed to the callback

Executes a function on all hash items, like lw6sys_hash_sort_and_map but befor doing so, sorts all entries in alphabetical order.

Return value: void

Function: lw6sys_hash_t * lw6sys_hash_dup (lw6sys_hash_t * hash, lw6sys_dup_func_t dup_func)

hash: the hash to duplicate, can be NULL

dup_func: the function which will be called to duplicate data

Duplicates an hash. All keys will be copied so that if the first hash is deleted, the duplicated one is fine. Additionnaly, dup_func will be called with all data fields. If dup_func is NULL, then data values will simply be copied. This is likely to be usefull when data is not dynamically allocated.

Returned value: a newly allocated hash.

Function: lw6sys_hexa_serializer_t * lw6sys_hexa_serializer_new (const char * hexa_string)

hexa_string: an initialization string, can be NULL.

Creates an hexa serializer object. It can be initialized or not, if an initialization string is provided it must of course be valid hexadecimal ascii code, and all serialized content will simply be appended to it.

Return value: a newly allocated object.

Function: void lw6sys_hexa_serializer_free (lw6sys_hexa_serializer_t * hexa_serializer)

hexa_serializer: an hexa serializer object

Frees an hexa serializer object.

Return value: none.

Function: void lw6sys_hexa_serializer_rewind (lw6sys_hexa_serializer_t * hexa_serializer)

hexa_serializer: an hexa serializer object

Rewinds the serializer pointer, that is, make it point to start. Usefull before calling pop functions, when one wants to be sure to get the first object.

Return value: none.

Function: int lw6sys_hexa_serializer_eof (lw6sys_hexa_serializer_t * hexa_serializer)

hexa_serializer: an hexa serializer object

Tests wether we’re at EOF. Usefull when one wants to know if there’s still some data or if all objects have been correctly popped.

Return value: 1 if at end of file, 0 if not.

Function: char * lw6sys_hexa_serializer_as_string (lw6sys_hexa_serializer_t * hexa_serializer)

hexa_serializer: an hexa serializer object

Exports the current content of the serializer as a string. String can then safely be sent on the network, for instance. String is copied from internal value, so it’s safe to use it after serializer has been freed or modified.

Return value: a newly allocated string, must be freed.

Function: int lw6sys_hexa_serializer_push_int64 (lw6sys_hexa_serializer_t * hexa_serializer, int64_t value)

hexa_serializer: an hexa serializer object

value: value to push

Pushes a 64 bit integer in the serializer object.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_push_int32 (lw6sys_hexa_serializer_t * hexa_serializer, int32_t value)

hexa_serializer: an hexa serializer object

value: value to push

Pushes a 32 bit integer in the serializer object.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_push_int16 (lw6sys_hexa_serializer_t * hexa_serializer, int16_t value)

hexa_serializer: an hexa serializer object

value: value to push

Pushes a 16 bit integer in the serializer object.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_push_int8 (lw6sys_hexa_serializer_t * hexa_serializer, int8_t value)

hexa_serializer: an hexa serializer object

value: value to push

Pushes an 8 bit integer in the serializer object.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_push_float (lw6sys_hexa_serializer_t * hexa_serializer, float value)

hexa_serializer: an hexa serializer object

value: value to push

Pushes a floating point value in the serializer object.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_push_str (lw6sys_hexa_serializer_t * hexa_serializer, const char * value)

hexa_serializer: an hexa serializer object

value: value to push

Pushes a string in the serializer object. Note that the string is not directly copied in the serializer, instead all its characters are converted to their ASCII equivalent, then appended.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_push_xyz (lw6sys_hexa_serializer_t * hexa_serializer, lw6sys_xyz_t value)

hexa_serializer: an hexa serializer object

value: value to push

Pushes a lw6sys_xyz_t structure in the serializer object. Calling this avoids calling push for 2 integers separately.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_push_whd (lw6sys_hexa_serializer_t * hexa_serializer, lw6sys_whd_t value)

hexa_serializer: an hexa serializer object

value: value to push

Pushes a lw6sys_whd_t structure in the serializer object. Calling this avoids calling push for 2 integers separately.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_push_color (lw6sys_hexa_serializer_t * hexa_serializer, lw6sys_color_8_t value)

hexa_serializer: an hexa serializer object

value: value to push

Pushes a color structure in the serializer object.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_pop_int64 (lw6sys_hexa_serializer_t * hexa_serializer, int64_t * value)

hexa_serializer: an hexa serializer object

value: value to pop (returned value)

Pops a 64 bit integer from the serializer object.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_pop_int32 (lw6sys_hexa_serializer_t * hexa_serializer, int32_t * value)

hexa_serializer: an hexa serializer object

value: value to pop (returned value)

Pops a 32 bit integer from the serializer object.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_pop_int16 (lw6sys_hexa_serializer_t * hexa_serializer, int16_t * value)

hexa_serializer: an hexa serializer object

value: value to pop (returned value)

Pops a 16 bit integer from the serializer object.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_pop_int8 (lw6sys_hexa_serializer_t * hexa_serializer, int8_t * value)

hexa_serializer: an hexa serializer object

value: value to pop (returned value)

Pops an 8 bit integer from the serializer object.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_pop_float (lw6sys_hexa_serializer_t * hexa_serializer, float * value)

hexa_serializer: an hexa serializer object

value: value to pop (returned value)

Pops a floating point value from the serializer object.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_pop_str (lw6sys_hexa_serializer_t * hexa_serializer, char ** value)

hexa_serializer: an hexa serializer object

value: value to pop (returned value)

Pops a string from the serializer object. The returned value is a newly allocated pointer, which must be freed, you don’t need to provide a buffer, just a valid pointer on a NULL pointer.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_pop_xyz (lw6sys_hexa_serializer_t * hexa_serializer, lw6sys_xyz_t * value)

hexa_serializer: an hexa serializer object

value: value to pop (returned value)

Pops a lw6sys_xyz_t structure from the serializer object. Avoids calling two integer pops.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_pop_whd (lw6sys_hexa_serializer_t * hexa_serializer, lw6sys_whd_t * value)

hexa_serializer: an hexa serializer object

value: value to pop (returned value)

Pops a lw6sys_whd_t structure from the serializer object. Avoids calling two integer pops.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_serializer_pop_color (lw6sys_hexa_serializer_t * hexa_serializer, lw6sys_color_8_t * value)

hexa_serializer: an hexa serializer object

value: value to pop (returned value)

Pops a color from the serializer object.

Return value: 1 if success, 0 if failure

Function: int lw6sys_hexa_str_to_buf (void * buf, int size, const char * str)

buf: binary buffer to convert

size: binary buffer length

str: the source string

Converts the stringified hexa representation of a string to its source binary buffer. Buffer must be exactly strlen(str)/2

Return value: 1 on success

Function: char * lw6sys_hexa_buf_to_str (void * buf, int size)

buf: the buffer to stringify

size: the length of the buffer

Transforms a binary buffer into its hexa representation.

Return value: newly allocated string.

Function: void * lw6sys_hexa_str_to_ptr (const char * str)

str: the string containing an hexa representation of pointer

Transforms a string into a pointer, this is typically used to store pointers in temporary agnostic storage such as a database. Beware not to use that to exchange data with other computers and/or use it for persistent data. This is a high-risk function as it lets you do real dirty stuff but it really does save time compared to using a key returned by the database engine and then search this key in a user-space hash table. Direct pointer access is definitely faster.

Return value: the pointer, or NULL is str is invalid.

Function: char * lw6sys_hexa_ptr_to_str (void * ptr)

ptr: pointer to convert into string representation

Transforms a pointer into a string, this is typically used to store pointers in temporary agnostic storage such as a database. Beware not to use that to exchange data with other computers and/or use it for persistent data. This is a high-risk function as it lets you do real dirty stuff but it really does save time compared to using a key returned by the database engine and then search this key in a user-space hash table. Direct pointer access is definitely faster.

Return value: the string, can be NULL on errror, must be freed.

Function: void lw6sys_history_init ()

Initializes the history system. Not initializing won’t cause any segfault, but data will be inconsistent.

Return value: none.

Function: void lw6sys_history_register (char * msg)

msg: the message to register.

Registers a message in the history log, that is, adds it.

Return value: none.

Function: char_ptr_t * lw6sys_history_get (int64_t timeout)

timeout: the message age limit.

Get all the messages that are younger than timeout (in seconds).

Return value: a pointer on string pointers. May be NULL. Last pointer is NULL too, that’s how you know the array is over.

Function: void lw6sys_history_free (char ** history)

history: the data to free

Frees a pointer returned by lw6sys_history_get.

Return value: none.

Function: char * lw6sys_locale_to_utf8 (const char * string)

string: the string to convert

Used to force strings into UTF-8 mode, this is basically to match the TTF font settings used when displaying stuff on OpenGL. Indeed, in this case, the standard _ gettext function won’t work, we need to force UTF-8 mode. If the locale is UTF-8, then function does nothing, but at least it’s transparent usage won’t hurt.

Returned value: a newly allocated string, always in UTF-8 no matter what the locale is.

Function: u_int16_t lw6sys_generate_id_16 ()

Long 16-bit ID generator, calls the random function internally. As usual, those are not perfect random numbers, however the function implementation emphasizes more on ’real randomness’ rather than relying on performance. Generating twice the same number should be fairly rare.

Function: u_int32_t lw6sys_generate_id_32 ()

Long 32-bit ID generator, calls the random function internally. As usual, those are not perfect random numbers, however the function implementation emphasizes more on ’real randomness’ rather than relying on performance. Generating twice the same number should be fairly rare.

Function: u_int64_t lw6sys_generate_id_64 ()

Long 64-bit ID generator, calls the random function internally. As usual, those are not perfect random numbers, however the function implementation emphasizes more on ’real randomness’ rather than relying on performance. Generating twice the same number should be fairly rare.

Function: int lw6sys_check_id_16 (u_int16_t id_16)

id_16: the id to check

Checks wether the given id is a valid 16-bit id.

Return value: 1 if OK, 0 if not a valid id.

Function: int lw6sys_check_id_32 (u_int32_t id_32)

id_32: the id to check

Checks wether the given id is a valid 32-bit id.

Return value: 1 if OK, 0 if not a valid id.

Function: int lw6sys_check_id_64 (u_int64_t id_64)

id_64: the id to check

Checks wether the given id is a valid 64-bit id.

Return value: 1 if OK, 0 if not a valid id.

Function: int lw6sys_check_id (u_int64_t id)

id: the id to check

Checks wether the given id is a valid id (16, 32 or 64-bit).

Return value: 1 if OK, 0 if not a valid id.

Function: char * lw6sys_id_ltoa (u_int64_t id)

id: the id to convert

Transform an id into its string representation. Error checking is done, if the id is invalid, returned value is NULL. All ids (16, 32 and 64-bit) are handled.

Return value: a newly allocated string, might be NULL.

Function: u_int64_t lw6sys_id_atol (char * id)

id: the id to convert

Transform an id into a long integer. Error checking is done, if the id is invalid, returned value is 0. All ids (16, 32 and 64-bit) are handled.

Return value: the id as a long integer, 0 if incorrect source id.

Function: char * lw6sys_keyword_as_key (const char * keyword)

keyword: the keyword to transform

Transforms a keyword into a "key", that is, removes all heading dashes, switches to lowercase, and other stuff. This is used internally to match options and config file parameters, for instance.

Return value: a newly allocated pointer, must be freed.

Function: char * lw6sys_keyword_as_arg (const char * keyword)

keyword: the keyword to transform

Transforms a keyword into a command-line parameter to be matched. Does the same as lw6sys_keyword_as_key, and adds a "–" prefix.

Return value: a newly allocated pointer, must be freed.

Function: char * lw6sys_keyword_as_env (const char * keyword)

keyword: the keyword to transform

Transforms a keyword into the corresponding environment variable name. It will uppercase the name, replace "-" by "_", and add a "LW6_" prefix. "my-param" will become "LW6_MY_PARAM".

Return value: a newly allocated pointer, must be freed.

Function: char * lw6sys_keyword_as_xml (const char * keyword)

keyword: the keyword to transform

Transforms a keyword into the corresponding config file entry. In practice, just the same as lw6sys_keyword_as_key.

Return value: a newly allocated pointer, must be freed.

Function: lw6sys_list_t * lw6sys_list_new (lw6sys_free_func_t free_func)

free_func: a callback which will be called on data when freeing the list

Creates an empty list. There’s a difference between NULL and an empty list. The empty list would (in Scheme) be ’() whereas NULL corresponds to undefined "is not a list and will generate errors if you ever call list functions on it".

Return value: a pointer to the created object, may be NULL.

Function: void lw6sys_list_free (lw6sys_list_t * list)

list: the list to delete.

Delete a list, this will cascade delete all the following items in the list.

Return value: none.

Function: lw6sys_list_t * lw6sys_list_next (lw6sys_list_t * list)

list: the current position in the list

It’s safer to call this rather than dig right into the internals of the list.

Return value: a new position in the list, may be NULL.

Function: int lw6sys_list_is_empty (lw6sys_list_t * list)

list: the list we want informations about

Checks wether the list is empty or not. Note that being empty and being NULL is not the same. An empty list is a valid pointer on a list where there’s no item, a NULL pointer is not a list at all. Do *NOT* call this function on NULL.

Return value: 1 if empty, 0 if there is at list one item.

Function: int lw6sys_list_length (lw6sys_list_t * list)

list: the list we want informations about

Calculates the length of the list. This is a performance killer for lists are inadapted to this. But it can still be usefull.

Return value: the number of elements, 0 is none (empty list).

Function: void lw6sys_list_map (lw6sys_list_t * list, lw6sys_list_callback_func_t func, void * func_data)

list: the list where elements will be taken

func: the function which will be executed

func_data: additionnal data to be passed to func

Executes a function on all list items. The func_data parameter allows you to pass extra values to the function, such as a file handler or any variable which can not be inferred from list item values, and you of course do not want to make global... Not as convenient as a real "for each" construct as can be found in any modern langage, but does the job. No return value, if you really want one, pass a structure in func_data and modify something in it on success, failure, whatever.

Return value: none.

Function: void lw6sys_list_filter (lw6sys_list_t ** list, lw6sys_list_filter_func_t func, void * func_data)

list: the list where elements will be taken

func: the function which will be executed

func_data: additionnal data to be passed to func

Executes a function on all list items and keeps only those for which the function returned non zero (true). The func_data parameter allows you to pass extra values to the function, such as a file handler or any variable which can not be inferred from list item values, and you of course do not want to make global...

Return value: none.

Function: void lw6sys_list_push_front (lw6sys_list_t ** list, void * data)

list: a pointer to the list (pointer on pointer, read/write value)

data: the data to be pushed

Pushes data on the list. The free_func function is copied from the previous element. The pointer on the list is changed "in place" (in/out). Note that if there’s a malloc problem it might end-up being NULL... This should be rare but it *can* happen. You cannot push something else than a pointer, pushing an int is a very bad idea. Push a pointer on the integer, and ensure it’s always there, or malloc it and pass lw6sys_free_callback when creating the list. If you think you can cast an integer into a pointer, think 64-bit machines...

Return value: none.

Function: void * lw6sys_list_pop_front (lw6sys_list_t ** list)

list: a pointer to the list (pointer on pointer, read/write value)

Pops data from the list, the returned value is what was passed to list_push. The pointer on the list is changed "in place" (in/out). When data is popped, that needs some freeing (i.e. free_func was not NULL when creating the list) then it’s the responsibility of the caller to free it when popping it. One popped it’s not freed, but it’s out of the list scope. Of course the lw6sys_list_t is freed, but not the data. If you happen to store non-NULL data in your list, you can call this function without bothering calling lw6sys_list_is_empty and assume that when you get NULL, there’s no data left. At this stage, the list won’t exist anymore BTW, you won’t even need to free it. The idea is: popping a list which has no elements left (empty list) destroys the list and returns NULL.

Return value: a pointer on the popped data, whatever you pushed.

Function: void lw6sys_list_push_back (lw6sys_list_t ** list, void * data)

list: a pointer to the list (pointer on pointer, read/write value)

data: the data to be pushed

Pushes data on the list. The free_func function is copied from the previous element. The pointer on the list is changed "in place" (in/out). Note that if there’s a malloc problem it might end-up being NULL... This should be rare but it *can* happen. You cannot push something else than a pointer, pushing an int is a very bad idea. Push a pointer on the integer, and ensure it’s always there, or malloc it and pass lw6sys_free_callback when creating the list. If you think you can cast an integer into a pointer, think 64-bit machines...

Return value: none.

Function: void * lw6sys_list_pop_back (lw6sys_list_t ** list)

list: a pointer to the list (pointer on pointer, read/write value)

Pops data from the list, the returned value is what was passed to list_push. The pointer on the list is changed "in place" (in/out). When data is popped, that needs some freeing (i.e. free_func was not NULL when creating the list) then it’s the responsibility of the caller to free it when popping it. One popped it’s not freed, but it’s out of the list scope. Of course the lw6sys_list_t is freed, but not the data. If you happen to store non-NULL data in your list, you can call this function without bothering calling lw6sys_list_is_empty and assume that when you get NULL, there’s no data left. At this stage, the list won’t exist anymore BTW, you won’t even need to free it. The idea is: popping a list which has no elements left (empty list) destroys the list and returns NULL.

Return value: a pointer on the popped data, whatever you pushed.

Function: lw6sys_list_t * lw6sys_list_dup (lw6sys_list_t * list, lw6sys_dup_func_t dup_func)

list: the list to duplicate, can be NULL

dup_func: the function which will be called to duplicate data

Duplicates a list. All data will be copied so that if the first list is deleted, the duplicated one is fine. Addtionnally, dup_func will be called to filter all data, and possibly allocated new pointers if needed, for instance. If dup_func is NULL, then data values will simply be copied. This is likely to be usefull when data is not dynamically allocated.

Returned value: a newly allocated list.

Function: const char * lw6sys_log_errno_str (int errno_int)

errno_int: the error code, typically errno

Convenience fonction which returns the "macro" corresponding to an errno code. I find it easier to use this than bothering declaring a buffer for strerror_r... Besides LW6b has its own error messages. Wil never return NULL, if error does not exists just returns "?".

Return value: static string, must not be freed

Function: void lw6sys_log_set_file (const char * filename)

filename: the name of the log file.

Sets up the log file. Until you call this function, messages all logged to the default log file, as returned by the lw6sys_get_default_log_file function.

Return value: void

Function: void lw6sys_log_clear (const char * filename)

filename: the name of the log file.

Clears the log file, this function would typically be called at the beginning of the program. If filename is NULL, then the default log file is cleared.

Return value: void

Function: int lw6sys_log_set_dialog_timeout (int timeout_sec)

timeout_sec: number of seconds to wait before alert dialogs disappear

By default, alert boxes will stay out forever unless one clicks on them, however, this parameter will force the dialog shutdown after some time. Mostly used for testing, to allow tests being blocked on a dialog.

Return value: 1 if timeout is supported on platform, 0 if not

Function: void lw6sys_log (int level_id, const char * file, int line, const char * fmt, ...)

level_id: the log level to use. Possible values are, by order, LW6SYS_LOG_ERROR_ID (0), LW6SYS_LOG_WARNING_ID (1), LW6SYS_LOG_NOTICE_ID (2), LW6SYS_LOG_INFO_ID (3), LW6SYS_LOG_DEBUG_ID (4) and LW6SYS_LOG_TMP_ID (5).

file: the name of the source file where the function is called, one can use __FILE__

line: the line in the source file where the function is called, one can use __LINE__

fmt: a printf-like format string ...: printf-like arguments, corresponding to fmt.

This function is usually called with the first three arguments packed into a single macro. For instance the LW6SYS_LOG_WARNING macro expands and fills the first 3 args, so there’s no need to type __FILE__ and __LINE__ again and again. Note that this function will reset errno. The idea is to call it whenever there’s something to do with errno (if you deal with errno, it’s a good habit to log it) then errno is cleared so that it won’t interfere with next log messages.

Function: void lw6sys_log_critical (const char * fmt, ...)

fmt: a printf-like format string ...: printf-like arguments, corresponding to fmt.

This function is a special log function which will dump informations on the console only, without opening any log file whatsoever. The idea is that it’s a "never fail" function. Additionnally, it will never return but quit the program. This can be used as an ultimate emergency function, use it when the program won’t run for sure, and displaying an immediate error message is the only issue.

Function: int lw6sys_log_get_level ()
Function: void lw6sys_log_set_level (int level)

level: the log level, integer between 0 & 4. 4 is very verbose (debug), 0 displays errors only.

Function: int lw6sys_log_get_console_state ()

Get the console output state. This is important, for instance to set the console "back in the state it was" after setting it on or off.

Return value: 1 if enabled, 0 if not.

Function: void lw6sys_log_set_console_state (int state)

state: 1 to activate console output, 0 to disable it.

Enables or disables console output. By console output, we basically mean stderr (and possibly stdout). If console output is enabled (the default) all output is copied to stderr. If it’s disabled, only the log file will contain the information.

Return value: none.

Function: void lw6sys_math_poly_wy1y2s1 (float * y, float * s, float x, float w, float y1, float y2, float s1)

y: the return value (position, may be NULL)

s: the return value (speed, may be NULL)

x: the x parameter, the value to iterate on

w: the width, that is, the x value after which output is constant

y1: the initial value, when v is s1 and x=0

y2: the target value, when v=0 and x>=w

s1: the initial speed, that is dy/dx at x=0

A function which can be used to implement smooth moving. It will extrapolate, for values of x>=0, an y position with a continuous derivate (dy/dx is continuous, function is 2nd order polynom) and which ends up at x=w with a constant value, that is dy/dx=v=0. Typically an item set with an initial speed of v with this function

Function: float lw6sys_math_angle_360 (int x, int y)

x: x coordinate

y: y coordinate

This is a wrapper over the standard atan function which will handle internally the special x == 0 case and the various positive/negative values of x and y.

Return value: the angle, in degrees

Function: float lw6sys_math_heartbeat (int64_t x, int period, float y1, float y2)

x: the parameter (typically a timestamp)

period: the period (typically something like 1000 milliseconds)

y1: the low value (heart at rest)

y2: the high value (when bumping)

A heartbeat function, typically usefull to power up eye-candy, but it could do something else.

Function: int lw6sys_math_blink (int64_t x, int period)

x: the parameter (typically a timestamp)

period: the period (typically something like 1000 milliseconds)

This function will alternatively return 1 or 0, usefull to handle blinking for instance.

Return value: 0 or 1

Function: float lw6sys_math_lin2log (int lin_value, int base)

lin_value: value on a linear scale

base: the base to use, 10 for decibel-like scale

Converts a linar-scale value to a logarithmic one. The scale is done so that base in linear mode is base in scaled mode, and it uses a log-base conversion, so that with a base 10 it behaves the way the decibel sound-volume unit works.

Return value: value on a logarithmic scale.

Function: int lw6sys_math_log2lin (float log_value, int base)

base: the base to use, 10 for decibel-like scale

Converts a linar-scale value to a logarithmic one. The scale is done so that 10 in linear mode is 10 in scaled mode, and it uses a log-base conversion, so that with a base 10 it behaves the way the decibel sound-volume unit works.

Return value: value on a linear scale.

Function: void * lw6sys_malloc (int size, const char * file, int line)

size: number of bytes to allocate.

file: name of the file calling the function, use __FILE__x_

line: line in the file calling the function, use __LINE__x_

This is a wrapper over the standard malloc function. Additionnally it will keep track of the call with an internal program-wide counter, thus enabling memory leak checks. You should not use this function directly but use the macro LW6SYS_MALLOC which has the same syntax, without the last two parameters, which are automatically provided by macro expansion.

Return value: the newly allocated pointer. Data is not initialized.

Function: void * lw6sys_calloc (int size, const char * file, int line)

size: number of bytes to allocate.

file: name of the file calling the function, use __FILE__x_

line: line in the file calling the function, use __LINE__x_

This is a wrapper over the standard calloc function. Additionnally it will keep track of the call with an internal program-wide counter, thus enabling memory leak checks. You should not use this function directly but use the macro LW6SYS_CALLOC which has the same syntax, without the last two parameters, which are automatically provided by macro expansion.

Return value: the newly allocated pointer. Data is filled with zeros.

Function: void * lw6sys_realloc (void * ptr, int size, const char * file, int line)

ptr: the pointer to reallocate.

size: number of bytes to allocate.

file: name of the file calling the function, use __FILE__x_

line: line in the file calling the function, use __LINE__x_

This is a wrapper over the standard realloc function. You should not use this function directly but use the macro LW6SYS_REALLOC which has the same syntax, without the last two parameters, which are automatically provided by macro expansion.

Return value: the newly allocated pointer.

Function: void lw6sys_free (void * ptr, const char * file, int line)

ptr: the pointer to free.

file: name of the file calling the function, use __FILE__x_

line: line in the file calling the function, use __LINE__x_

This is a wrapper over the standard free function. Additionnally it will keep track of the call with an internal program-wide counter, thus enabling memory leak checks. You should not use this function directly but use the macro LW6SYS_FREE which has the same syntax, without the last two parameters, which are automatically provided by macro expansion.

Return value: none.

Function: void lw6sys_free_callback (void * ptr)

ptr: the pointer to free.

This is a callback to be used when the lw6sys_free does not fit. A good example is a list, which, to free its elements, requires you to provide a callback that only takes 1 arg, the pointer to free. Problem, lw6sys_free takes 3 args. And the LW6SYS_FREE macro is not usable in such a context. And you can’t use standard free either for it would mess up the malloc / free automatical count which is so convenient to track memory leaks. So this callback is here, it’s only drawback is that in case of an error, the error will not be reported with the real file and line parameters. It’s still better than nothing.

Return value: none.

Function: int lw6sys_megabytes_available ()

Gives a raw approximation of available memory, in megabytes. Value is to be taken with distance, but it can give good hints when system is running short of ressources.

Return value: number of megabytes (physical memory) available.

Function: int lw6sys_is_big_endian ()

Checks the endianess of the machine. PPC is big endian, for instance.

Return value: 1 if system is big endian, 0 if little endian.

Function: int lw6sys_is_little_endian ()

Checks the endianess of the machine. x86 is little endian, for instance.

Return value: 1 if system is little endian, 0 if big endian.

Function: int lw6sys_check_types_size ()

Checks of common types and usefull structures, this is a debugging function which helps finding compiler strange behaviors and programmer’s bad intuitions.

Return value: 1 if everything is OK, 0 if error.

Function: lw6sys_mutex_t * lw6sys_mutex_create ()

Creates a mutex object.

Return value: newly allocated pointer.

Function: void lw6sys_mutex_destroy (lw6sys_mutex_t * mutex)

mutex: the mutex to destroy.

Destroys a mutex object.

Return value: none.

Function: int lw6sys_mutex_lock (lw6sys_mutex_t * mutex)

mutex: the mutex to use

Locks the mutex. Note that this should never fail unless there’s a serious initialization problem, instead, function will wait forever until mutex is released.

Return value: 1 if success, 0 if failure.

Function: int lw6sys_mutex_trylock (lw6sys_mutex_t * mutex)

mutex: the mutex to use

Tries to locks the mutex. That is, tells wether mutex can be locked immediately or not. Note that this does not mean there’s 100% chance next call to lock will terminated immediately, since lock can still be acquired by another thread.

Return value: 1 if mutex unlocked, 0 if locked or error.

Function: int lw6sys_mutex_unlock (lw6sys_mutex_t * mutex)

mutex: the mutex to use

Unlocks a mutex.

Return value: 1 if sucess, 0 if error.

Function: int lw6sys_get_mutex_lock_count ()

Returns how many mutexes have been locked since program start. Usefull for sanity checking when debugging.

Return value: number of calls to lock

Function: int lw6sys_get_mutex_unlock_count ()

Returns how many mutexes have been unlocked since program start. Usefull for sanity checking when debugging.

Return value: number of calls to unlock

Function: int lw6sys_check_mutex_count ()

Checks wether unlock has been called as many times as lock. Usefull for sanity checking when debugging.

Return value: 1 if OK, 0 if inconsistency.

Function: int lw6sys_true ()

Function which returns always true, that is, something different than 0.

Function: int lw6sys_false ()

Function which returns always false, that is, 0. This can seem totally useless but it does have some utility. It’s used for instance to "fool" the compiler and force it to compile and link functions in binaries, so that, afterwards, dynamically loaded .so files can find in the main binary some functions which would otherwise be stripped during the final link.

Function: int lw6sys_openmp_get_num_procs ()

Wrapper on omp_get_num_procs the advantage of this is that it’s always defined, wether OpenMP supported is compiled in or not, will returned 1 if no OpenMP support.

Return value: number of procs

Function: char * lw6sys_get_default_user_dir ()

Returns the default user directory. Note that this value is not static, it can depend, for instance, of the environment variable HOME.

Return value: a newly allocated string.

Function: char * lw6sys_get_default_config_file ()

Returns the default config file. Note that this value is not static, it can depend, for instance, of the environment variable HOME.

Return value: a newly allocated string.

Function: char * lw6sys_get_default_log_file ()

Returns the default log file. Note that this value is not static, it can depend, for instance, of the environment variable HOME.

Return value: a newly allocated string.

Function: char * lw6sys_get_default_prefix ()

Returns the default prefix, could be /usr/local for instance.

Return value: a newly allocated string.

Function: char * lw6sys_get_default_mod_dir ()

Returns the default module directory (dynamically loaded libraries).

Return value: a newly allocated string.

Function: char * lw6sys_get_default_data_dir ()

Returns the default data directory.

Return value: a newly allocated string.

Function: char * lw6sys_get_default_music_dir ()

Returns the default music directory.

Return value: a newly allocated string.

Function: char * lw6sys_get_default_music_path ()

Returns the default music path, which can be composed of several directories.

Return value: a newly allocated string.

Function: char * lw6sys_get_default_map_dir ()

Returns the default map directory.

Return value: a newly allocated string.

Function: char * lw6sys_get_default_map_path ()

Returns the default map path, which can be composed of several directories.

Return value: a newly allocated string.

Function: char * lw6sys_get_default_script_file ()

Returns the default script file.

Return value: a newly allocated string.

Function: void lw6sys_options_log_defaults ()

Logs all default values to log file. Usefull for debugging, to know where the program is searching for its informations.

Function: char * lw6sys_get_cwd ()

Returns the current working directory (absolute path).

Return value: a newly allocated string.

Function: char * lw6sys_get_run_dir (int argc, const char * [] argv)

argc: argc, number of arguments, as given to main

argv: argv, pointer to arguments, as given to main

Returns the binary directory, that is, the directory the binary is stored in. This is calculated dynamically, by interpreting command-line arguments.

Return value: a newly allocated string.

Function: char * lw6sys_get_user_dir (int argc, const char * [] argv)

argc: argc, number of arguments, as given to main

argv: argv, pointer to arguments, as given to main

Returns the user dir, taking in account command-line and environment variables. However config file content has no impact on the result.

Return value: a newly allocated string.

Function: char * lw6sys_get_config_file (int argc, const char * [] argv)

argc: argc, number of arguments, as given to main

argv: argv, pointer to arguments, as given to main

Returns the config file, taking in account command-line and environment variables. However config file content has no impact on the result.

Return value: a newly allocated string.

Function: char * lw6sys_get_log_file (int argc, const char * [] argv)

argc: argc, number of arguments, as given to main

argv: argv, pointer to arguments, as given to main

Returns the log file, taking in account command-line and environment variables. However config file content has no impact on the result.

Return value: a newly allocated string.

Function: char * lw6sys_get_prefix (int argc, const char * [] argv)

argc: argc, number of arguments, as given to main

argv: argv, pointer to arguments, as given to main

Returns the prefix, taking in account command-line and environment variables. However config file content has no impact on the result.

Return value: a newly allocated string.

Function: char * lw6sys_get_mod_dir (int argc, const char * [] argv)

argc: argc, number of arguments, as given to main

argv: argv, pointer to arguments, as given to main

Returns the mod dir (modules, shared .so), taking in account command-line and environment variables. However config file content has no impact on the result.

Return value: a newly allocated string.

Function: char * lw6sys_get_data_dir (int argc, const char * [] argv)

argc: argc, number of arguments, as given to main

argv: argv, pointer to arguments, as given to main

Returns the data dir, taking in account command-line and environment variables. However config file content has no impact on the result.

Return value: a newly allocated string.

Function: char * lw6sys_get_music_dir (int argc, const char * [] argv)

argc: argc, number of arguments, as given to main

argv: argv, pointer to arguments, as given to main

Returns the music dir, taking in account command-line and environment variables. However config file content has no impact on the result.

Return value: a newly allocated string.

Function: char * lw6sys_get_music_path (int argc, const char * [] argv)

argc: argc, number of arguments, as given to main

argv: argv, pointer to arguments, as given to main

Returns the music path, taking in account command-line and environment variables. However config file content has no impact on the result. Music path can contain several directories.

Return value: a newly allocated string.

Function: char * lw6sys_get_map_dir (int argc, const char * [] argv)

argc: argc, number of arguments, as given to main

argv: argv, pointer to arguments, as given to main

Returns the map dir, taking in account command-line and environment variables. However config file content has no impact on the result.

Return value: a newly allocated string.

Function: char * lw6sys_get_map_path (int argc, const char * [] argv)

argc: argc, number of arguments, as given to main

argv: argv, pointer to arguments, as given to main

Returns the map path, taking in account command-line and environment variables. However config file content has no impact on the result. Map path can contain several directories.

Return value: a newly allocated string.

Function: char * lw6sys_get_script_file (int argc, const char * [] argv)

argc: argc, number of arguments, as given to main

argv: argv, pointer to arguments, as given to main

Returns the script file, taking in account command-line and environment variables. However config file content has no impact on the result.

Return value: a newly allocated string.

Function: void lw6sys_options_log (int argc, const char * [] argv)

argc: argc, number of arguments, as given to main

argv: argv, pointer to arguments, as given to main

Logs all the main options values which are not config-file dependant but depend on built-in defaults, command-line arguments and environment variables. Usefull to debug and know where the program is searching for things.

Function: int lw6sys_file_exists (const char * filename)

filename: the file to test

Tests the existence of a file on the filesystem. File is considered to exists if it’s at least readable.

Return value: 1 if OK, 0 if file doesn’t exist or can’t be read.

Function: int lw6sys_dir_exists (const char * dirname)

dirname: the directory to test

Tests the existence of a directory on the filesystem.

Return value: 1 if OK, 0 if directory doesn’t exist.

Function: int lw6sys_dir_exists_with_readme (const char * dirname)

dirname: the directory to test

Tests the existence of a directory on the filesystem, will also validate that it contains a README or readme.txt file.

Return value: 1 if OK, 0 if directory doesn’t exist.

Function: int lw6sys_dir_exists_with_readme_containing_text (const char * dirname, const char * needle)

dirname: the directory to test

needle: string to search, can be NULL

Tests the existence of a directory on the filesystem, with a README or readme.txt file, which contains the string needle.

Return value: 1 if OK, 0 if directory doesn’t exist.

Function: int lw6sys_create_dir (const char * dirname)

dirname: the directory to create

Creates a directory, performing sanity checks such as verifying the directory really exists after being created.

Return value: 1 if OK, 0 if error.

Function: int lw6sys_create_dir_silent (const char * dirname)

dirname: the directory to create

Creates a directory like lw6sys_create_dir but this function is silent in the sense that it won’t log any error. Usefull to create the log directory itself, for instance, and avoid infinite loops on error.

Return value: 1 if OK, 0 if error.

Function: char * lw6sys_path_add_slash (const char * path)

path: a path

Adds a slash, or in a general manner, a directory separator, at the end of a path, if needed. So /foo/bar will become /foo/bar/ but /bar/foo/ will remain /bar/foo/.

Return value: a newly allocated string, must be freed.

Function: char * lw6sys_path_strip_slash (const char * path)

path: a path

Strips the slash, or in a general manner, the directory separator, at the end of a path, if needed. So /foo/bar/ will become /foo/bar but /bar/foo will remain /bar/foo.

Return value: a newly allocated string, must be freed.

Function: char * lw6sys_path_concat (const char * path1, const char * path2)

path1: left part of the path

path2: right part of the path

Concatenates 2 parts of a path. Function will try to avoid stupid "double-slash" when concatenating /foo/ with /bar/ and conversely insert a directory separator when concatenating /foo with bar/.

Return value: a newly allocated string, must be freed.

Function: lw6sys_list_t * lw6sys_path_split (const char * path)

path: a path

Splits a path into all its parts. For instance /boo/bar/foo2/bar2 returns a 4 elements list. This is more than a plain split, for heading and tailing slashes will be ignored, and various path separators will be interpreted (depends on platform).

Return value: a list containing 0-terminated strings.

Function: char * lw6sys_path_file_only (const char * path)

path: a path

Returns the file name only, without heading directories.

Return value: file name, must be freed

Function: int lw6sys_path_is_relative (const char * path)

path: a path

Checks wether a path is relative or absolute.

Return value: 1 if relative, 0 if absolute.

Function: int lw6sys_path_is_cwd (const char * path)

path: a path

Checks wether a path is "." or not. Will also trap "" and "./".

Return value: 1 if relative, 0 if absolute.

Function: char * lw6sys_path_parent (const char * path)

path: a path

Returns the parent path. That will return /foo when given /foo/bar in input.

Return value: a newly allocated string, must be freed.

Function: char * lw6sys_path_unparent (const char * path)

path: a path

Given the ../foo/bar path, will return foo/bar. Usefull to get rid of heading ../ when a path is known to start with it.

Return value: a newly allocated string, must be freed.

Function: char * lw6sys_path_unparent_no_malloc (char * path)

path: a path

Given the ../foo/bar path, will return foo/bar. Usefull to get rid of heading ../ when a path is known to start with it. This is different from lw6sys_path_unparent just because the result is not dynamically allocated and copied from source.

Return value: a pointer which points somewhere within the string passed as an argument.

Function: lw6sys_list_t * lw6sys_dir_list (const char * dir, lw6sys_dir_list_filter_func_t filter_func, void * func_data, int * n)

dir: the path of the directory to list

filter_func: a function which will filter entries, can be NULL

func_data: additionnal data passed to filter_func

n: will contain the number of items found

This list a directory. The filter will be passed the file path as an argument. If it returns 1, the file is kept, if it returns 0 it’s suppressed from the list.

Return value: a list containing strings (file paths).

Function: lw6sys_list_t * lw6sys_path_list (const char * path, lw6sys_dir_list_filter_func_t filter_func, void * func_data, int * n)

path: the path of the path to list

filter_func: a function which will filter entries, can be NULL

func_data: additionnal data passed to filter_func

n: will contain the number of items found

This list a directory. By path we mean here a list of separated directories, separated by : for instance. The filter will be passed the file path as an argument. If it returns 1, the file is kept, if it returns 0 it’s suppressed from the list. It’s like performing a call to lw6sys_dir_list on each of the path members.

Return value: a list containing strings (file paths).

Function: char * lw6sys_find_in_dir_and_path (const char * dir, const char * path, const char * file)

dir: a directory, when to search the file first

path: the path to search too, a separated list of dirs

file: the filename to search for

Tries to find a file in the given paths. The function is typically used to find music files. First it tries to find the file in dir, then it tries to find it in each dir of path. file must be only a file name and not contain any directory. The function will use the filename only anyway.

Return value: the full path of the found file.

Function: void lw6sys_print_xml_header (FILE * f, char * comment)

f: file to output content to

Prints a standard Liquid War compliant XML header in the given file.

Return value: none.

Function: void lw6sys_print_xml_footer (FILE * f)

f: file to output content to

Prints a standard Liquid War 6 compliant XML footer in the given file.

Return value: none.

Function: int lw6sys_process_is_fully_supported ()

Tells wether functions related to fork and pids are likely to work allright or not. Typically, those functions will return false (0) systematically if called on a platform that does not support them fully. In practice this is only for a few tests, so it’s not that bad if it does not really work. Would be better if, but well, it’s OK.

Return value: 1 if supported, 0 if not.

Function: u_int64_t lw6sys_process_fork_and_call (lw6sys_fork_func_t func, void * data)

data: pointer on arbitrary data used by func

This is not a standard fork function, it will return to the caller (parent) with something >0 if a child has been created, or 0 if failed. On the child it will launch the callback func, run it and exit right away.

Return value: a process ID on success, 0 on failure.

Function: int lw6sys_process_kill_1_9 (u_int64_t pid)

pid: pid to kill

Kills a process with the given PID. The kill will first use a signal 1 SIGTERM the a signal 9 SIGKILL. This is mostly for testing, the idea is to be sure to vaccum after we’re done. We use 64-bit for PIDs, yes, they are very likely 32 only, but had to choose (and pid_t is a pain because for logging one would needed to check the length before calling printf like functions...).

Return value: 1 on success, 0 if failed

Function: int lw6sys_profiler_check (int verbose)

verbose: wether to display informations on the console

Checks wether Google Profiler support has been built, and if it’s set, outputs the log file. If CPUPROFILE is defined but binary has no support for it, will display a warning message.

Return value: 1 if google profile enabled and activated, 0 if not

Function: void lw6sys_progress_bind (lw6sys_progress_t * progress, float * value)

progress: the progress struct to initialize

value: the value to point to

Sets a progress struct to default values, that is, ranging from 0.0f to 1.0f, does not touch the value.

Return value: none.

Function: void lw6sys_progress_default (lw6sys_progress_t * progress, float * value)

progress: the progress struct to initialize

value: the value to point to

Sets a progress struct to default values, that is, ranging from 0.0f to 1.0f. This function won’t touch the value, one needs to call begin or update or end to do that.

Return value: none.

Function: void lw6sys_progress_update (lw6sys_progress_t * progress, int min, int max, int value)

progress: the progress struct to update

min: the min value

max: the max value

value: the current value

Updates a progress struct. This is typically the function used by a callback to show the progress of a process. Note that this is note an initializer. Rather, the progress struct was initialized before, and this call is done in a loop with min being 0, max being the last value in the loop, and value the current index in the loop. NULL pointers correctly handled internally, so call this with any parameters, it’s safe.

Return value: none.

Function: void lw6sys_progress_split (lw6sys_progress_t * progress1, lw6sys_progress_t * progress2, lw6sys_progress_t * progress_src)

progress1: the first part of the splitted progress progress2: the second part of the splitted progress progress_src: the progress to split

Utility function to split a progress struct, that is, if a progress was ranging from a to b, make 2 progress structs, ranging from a to c and from c to b, c being between a and b.

Return value: none

Function: void lw6sys_progress_split_here (lw6sys_progress_t * progress1, lw6sys_progress_t * progress2, lw6sys_progress_t * progress_src, float here)

progress1: the first part of the splitted progress progress2: the second part of the splitted progress progress_src: the progress to split here: where to split

Utility function to split a progress struct, that is, if a progress was ranging from a to b, make 2 progress structs, ranging from a to c and from c to b, c being between a and b. The here value controls what c is. If here=0, then c=a. If here=1, then c=b.

Return value: none

Function: void lw6sys_progress_split3 (lw6sys_progress_t * progress1, lw6sys_progress_t * progress2, lw6sys_progress_t * progress3, lw6sys_progress_t * progress_src)

progress1: the first part of the splitted progress progress2: the second part of the splitted progress progress3: the third part of the splitted progress progress_src: the progress to split

Utility function to split a progress struct, this one will split it into 3 equal parts.

Return value: none

Function: void lw6sys_progress_split4 (lw6sys_progress_t * progress1, lw6sys_progress_t * progress2, lw6sys_progress_t * progress3, lw6sys_progress_t * progress4, lw6sys_progress_t * progress_src)

progress1: the first part of the splitted progress progress2: the second part of the splitted progress progress3: the third part of the splitted progress progress4: the fourth part of the splitted progress progress_src: the progress to split

Utility function to split a progress struct, this one will split it into 4 equal parts.

Return value: none

Function: void lw6sys_progress_split5 (lw6sys_progress_t * progress1, lw6sys_progress_t * progress2, lw6sys_progress_t * progress3, lw6sys_progress_t * progress4, lw6sys_progress_t * progress5, lw6sys_progress_t * progress_src)

progress1: the first part of the splitted progress progress2: the second part of the splitted progress progress3: the third part of the splitted progress progress4: the fourth part of the splitted progress progress5: the fourth part of the splitted progress progress_src: the progress to split

Utility function to split a progress struct, this one will split it into 5 equal parts.

Return value: none

Function: void lw6sys_progress_begin (lw6sys_progress_t * progress)

progress: the progress to update

Sets the progress to its min value, NULL values correctly handled.

Return value: none

Function: void lw6sys_progress_half (lw6sys_progress_t * progress)

progress: the progress to update

Sets the progress to the average between min and max, NULL values correctly handled.

Return value: none

Function: void lw6sys_progress_end (lw6sys_progress_t * progress)

progress: the progress to update

Sets the progress to its max value, NULL values correctly handled.

Return value: none

Function: u_int32_t lw6sys_random (u_int32_t range)

range: the high limit for random generated numbers. If you want random numbers between 0 and 5, set this to 6.

Wrapper over standard random function. This one is thread safe. This idea is not to provide cryptographic-proof random numbers, rather generate sequences which are random enough to generate unique server ids and such things. The function is initialized on its first call, and results depend on timestamp, host name, user name, and memory available.

Function: float lw6sys_random_float (float min, float max)

min: the min value, as a float

max: the max value, as a float

Returns a random float number between min & max. Can be equal to min or max.

Function: int lw6sys_sdl_register ()

Function used to avoid initializing SDL several times in a program. AFAIK Allegro has a was_init function, but SDL doesn’t. With this function - which every LW6 sub-module should use - one can know globally, for the whole program, wether SDL has been initialized or not.

Function: int lw6sys_sdl_unregister ()

Call this whenever you are done with SDL and exit it, so that the lw6sys_sdl_register function works correctly.

Return value: 1 if SDL needs to be unregistered, that is, if it has already been initialized, else 0.

Function: void lw6sys_serialize_int64 (unsigned char * data, int64_t value)

data: pointer to the data, must contain at least 8 bytes of writable space

value: the integer to serialize

Serializes a 64-bit integer in a byte buffer. Result is not dependant on machine endianess. Typically used for checksums or high-level serializations.

Function: int64_t lw6sys_unserialize_int64 (unsigned char * data)

data: pointer to the data, must contain at least 8 bytes

Recovers a 64-bit integer from a byte buffer created, for instance, with lw6sys_serialize_int64.

Function: void lw6sys_serialize_int32 (unsigned char * data, int32_t value)

data: pointer to the data, must contain at least 4 bytes of writable space

value: the integer to serialize

Serializes a 32-bit integer in a byte buffer. Result is not dependant on machine endianess. Typically used for checksums or high-level serializations.

Function: int32_t lw6sys_unserialize_int32 (unsigned char * data)

data: pointer to the data, must contain at least 4 bytes

Recovers a 32-bit integer from a byte buffer created, for instance, with lw6sys_serialize_int32.

Function: void lw6sys_serialize_int16 (unsigned char * data, int16_t value)

data: pointer to the data, must contain at least 2 bytes of writable space

value: the integer to serialize

Serializes a 16-bit integer in a byte buffer. Result is not dependant on machine endianess. Typically used for checksums or high-level serializations.

Function: int16_t lw6sys_unserialize_int16 (unsigned char * data)

data: pointer to the data, must contain at least 2 bytes

Recovers a 16-bit integer from a byte buffer created, for instance, with lw6sys_serialize_int16.

Function: int lw6sys_shape_check_min_max_whd (const lw6sys_whd_t * shape, const lw6sys_whd_t * min, const lw6sys_whd_t * max)

shape: the dimensions to control

min: the minimum shape allowed

max: the maximum shape allowed

Will check wether the given shape respects some basic constraints, being not to small and not too big.

Return value: 1 if OK, 0 if not.

Function: int lw6sys_shape_check_pos (const lw6sys_whd_t * shape, const lw6sys_xyz_t * pos)

shape: the boundary box

pos: the position

Checks wether position is within the given boundary box.

Return value: 1 if OK, 0 if not.

Function: int lw6sys_shape_is_same (const lw6sys_whd_t * shape_a, const lw6sys_whd_t * shape_b)

shape_a: the first shape to compare

shape_b: the other shape to compare

Compares two shapes.

Return value: 1 if same, 0 if not.

Function: int lw6sys_shape_is_same_xy (const lw6sys_whd_t * shape_a, const lw6sys_whd_t * shape_b)

shape_a: the first shape to compare

shape_b: the other shape to compare

Compares two shapes, but ignores the z (d) parameter.

Return value: 1 if same_xy, 0 if not.

Function: int lw6sys_shape_volume_whd (const lw6sys_whd_t * shape)

Gives the volume (w * h * d) for a given shape.

Return value: the volume.

Function: int lw6sys_shape_surface_wh (const lw6sys_whd_t * shape)

Gives the surface (w * h) for a given shape.

Return value: the surface.

Function: void lw6sys_signal_custom (int trap_errors)

trap_errors: set to 1 if you want to trap SIGSEGV and SIGFPE

Set up our signal handlers. This will probably be overrided later by other libs such as libSDL, but at least in pure server mode it gives a way to treat SIGTERM the right way.

Return value: none.

Function: void lw6sys_signal_default ()

Restore default signal handlers for those modified by lw6sys_signal_custom.

Return value: none.

Function: void lw6sys_signal_term_handler (int signum)

signum: SIGTERM

The own TERM signal handler, will basically call the lw6sys_signal_send_quit function, which will set a flag used later by lw6sys_signal_poll_quit.

Return value: none.

Function: void lw6sys_signal_int_handler (int signum)

signum: SIGINT

The own INT signal handler, will basically call the lw6sys_signal_send_quit function, which will set a flag used later by lw6sys_signal_poll_quit.

Return value: none.

Function: void lw6sys_signal_hup_handler (int signum)

signum: SIGTERM

The own HUP signal handler, will basically do something that shows the program is alive, typically display a NOTICE message.

Return value: none.

Function: void lw6sys_signal_segv_handler (int signum)

signum: SIGTERM

The own SEGV signal handler, will display a backtrace and exit.

Return value: none.

Function: void lw6sys_signal_fpe_handler (int signum)

signum: SIGTERM

The own FPE signal handler, will display a backtrace and exit.

Return value: none.

Function: void lw6sys_signal_send_quit ()

Sets the quit flag to 1, so that lw6sys_signal_poll_quit returns true, that is, tells the polling loop to stop.

Return value: none.

Function: int lw6sys_signal_poll_quit ()

Tests wether we need to stop right now.

Return value: 1 if we need to stop now, 0 if program can continue.

Function: int lw6sys_sort_int_callback (const lw6sys_list_t ** list_a, const lw6sys_list_t ** list_b)

list_a: pointer to a list of int item

list_b: pointer to a list of int item

A typicall sort callback function, can be passed to lw6sys_sort to sort a list of integers.

Return value: -1 if list_a < list_b , 0 if list_a == list_b, 1 if list_a > list_b

Function: int lw6sys_sort_int_desc_callback (const lw6sys_list_t ** list_a, const lw6sys_list_t ** list_b)

list_a: pointer to a list of int item

list_b: pointer to a list of int item

A typicall sort callback function, can be passed to lw6sys_sort to sort a list of integers. This one will sort in reverse mode.

Return value: 1 if list_a < list_b , 0 if list_a == list_b, -1 if list_a > list_b

Function: int lw6sys_sort_float_callback (const lw6sys_list_t ** list_a, const lw6sys_list_t ** list_b)

list_a: pointer to a list of float item

list_b: pointer to a list of float item

A typicall sort callback function, can be passed to lw6sys_sort to sort a list of floating point numbers.

Return value: -1 if list_a < list_b , 0 if list_a == list_b, 1 if list_a > list_b

Function: int lw6sys_sort_float_desc_callback (const lw6sys_list_t ** list_a, const lw6sys_list_t ** list_b)

list_a: pointer to a list of float item

list_b: pointer to a list of float item

A typicall sort callback function, can be passed to lw6sys_sort to sort a list of floating point numbers. This one will sort in reverse mode.

Return value: 1 if list_a < list_b , 0 if list_a == list_b, -1 if list_a > list_b

Function: int lw6sys_sort_str_callback (const lw6sys_list_t ** list_a, const lw6sys_list_t ** list_b)

list_a: pointer to a list of string item

list_b: pointer to a list of string item

A typicall sort callback function, can be passed to lw6sys_sort to sort a list of 0-terminated strings.

Return value: -1 if list_a < list_b , 0 if list_a == list_b, 1 if list_a > list_b

Function: int lw6sys_sort_str_desc_callback (const lw6sys_list_t ** list_a, const lw6sys_list_t ** list_b)

list_a: pointer to a list of string item

list_b: pointer to a list of string item

A typicall sort callback function, can be passed to lw6sys_sort to sort a list of 0-terminated strings. This one will sort in reverse mode.

Return value: 1 if list_a < list_b , 0 if list_a == list_b, -1 if list_a > list_b

Function: void lw6sys_sort (lw6sys_list_t ** list, lw6sys_sort_callback_func_t sort_func)

list: the list to sort, might be modified by the function

sort_func: the callback function used to sort

A general sorting function. Internally, will use the glibc qsort function, but this one is adapted to the LW6 specific data structures, more exactly, the lw6sys_list structure. Several default sort callbacks are defined, but one is free to use any callback, provided it has the right prototype.

Function: lw6sys_spinlock_t * lw6sys_spinlock_create ()

Creates a spinlock object.

Return value: newly allocated pointer.

Function: void lw6sys_spinlock_destroy (lw6sys_spinlock_t * spinlock)

spinlock: the spinlock to destroy.

Destroys a spinlock object.

Return value: none.

Function: int lw6sys_spinlock_lock (lw6sys_spinlock_t * spinlock)

spinlock: the spinlock to use

Locks the spinlock. Note that this should never fail unless there’s a serious initialization problem, instead, function will wait forever until spinlock is released.

Return value: 1 if success, 0 if failure.

Function: int lw6sys_spinlock_trylock (lw6sys_spinlock_t * spinlock)

spinlock: the spinlock to use

Tries to locks the spinlock. That is, tells wether spinlock can be locked immediately or not. Note that this does not mean there’s 100% chance next call to lock will terminated immediately, since lock can still be acquired by another thread.

Return value: 1 if spinlock unlocked, 0 if locked or error.

Function: int lw6sys_spinlock_unlock (lw6sys_spinlock_t * spinlock)

spinlock: the spinlock to use

Unlocks a spinlock.

Return value: 1 if sucess, 0 if error.

Function: char * lw6sys_str_copy (const char * src)

src: the string to copy

Duplicate a string, creating a new pointer on it, which must be freed afterwards. The main difference with strdup is that here we use the LW6SYS_MALLOC macro to track down possible memory leaks.

Return value: a newly allocated pointer, must be freed.

Function: char * lw6sys_str_concat (const char * str1, const char * str2)

str1: the left part to be concatenated

str2: the right part to be concatenated

Concatenate 2 strings, and put the result in a newly allocated string. Unlike strcat which uses the same pointer.

Return value: a newly allocated pointer, must be freed.

Function: char * lw6sys_new_sprintf (const char * fmt, ...)

fmt: a format string, like the one you would pass to printf ...: optional arguments, like the ones you would pass to printf

An sprintf like function, except it allocates a new string automatically, with "enough space". This is not a highly optimized function, it will allocate plenty of memory, possibly several times, and thus consume time and resources. But it has the great advantage of freeing the programmer of the dirty work of guessing "how log will the sprintf’ed string be?" before even generating it. So it’s a time saver for the programmer. Additionnally, helps avoiding memory leaks and buffer overflows.

Return value: a new allocated string, must be freed.

Function: int lw6sys_buf_sprintf (char * buf, int len, const char * fmt, ...)

buf: a buffer of len+1 chars

len: the max length of string

fmt: a format string, like the one you would pass to printf ...: optional arguments, like the ones you would pass to printf

Almost like snprintf except that it will *always* append a char 0 (’\0’) at the end of the string. Therefore buf must be of size len+1.

Return value: 1 if success, 0 if failed.

Function: int lw6sys_str_is_blank (const char * str)

str: the string to test

Tests wether a string is blank, that is, if it’s composed of space, tabs, or carriage returns only.

Return value: 1 if blank, 0 if not.

Function: int lw6sys_str_is_null_or_empty (const char * str)

str: the string to test

Tests wether a string is NULL or empty (string with 0 chars "").

Return value: 1 if NULL or empty, 0 if contains something.

Function: const char * lw6sys_str_empty_if_null (const char * str)

str: the string to test

Returns always a non-NULL string, if string is NULL, returns "" The argument str is not passed as const else this function would equate to a disguised cast from const to non-const.

Return value: source string or "" if it was NULL

Function: int lw6sys_str_is_same (const char * str_a, const char * str_b)

str_a: 1st string to compare, can be NULL

str_b: 2nd string to compare, can be NULL

Compares two strings for equality. Difference with strcmp is that this one won’t check for alphabetical order and return -1 or +1, but will check for NULL args. of space, tabs, or carriage returns only.

Return value: 1 if same, 0 if not.

Function: int lw6sys_str_is_same_no_case (const char * str_a, const char * str_b)

str_a: 1st string to compare, can be NULL

str_b: 2nd string to compare, can be NULL

Compares two strings for equality. Difference with strcmp is that this one won’t check for alphabetical order and return -1 or +1, but will check for NULL args. of space, tabs, or carriage returns only. This function is not case sensitive.

Return value: 1 if same, 0 if not.

Function: int lw6sys_str_starts_with (const char * str, const char * beginning)

str: the string to analyse

beginning: the pattern to search

Tells wether string starts with a given beginning.

Return value: 1 if str starts with beginning, 0 if not

Function: int lw6sys_str_starts_with_no_case (const char * str, const char * beginning)

str: the string to analyse

beginning: the pattern to search

Tells wether string starts with a given beginning. This function is not case sensitive.

Return value: 1 if str starts with beginning, 0 if not

Function: int lw6sys_skip_blanks (char ** str_ptr)

str_ptr: a pointer to a string pointer (read/write parameter).

Skips blanks at the beginning of a string. The passed parameter is modifed in place. Usefull for parsing.

Return value: 1 if blanks were found, else 0.

Function: void lw6sys_str_cleanup (char * str)

str: a pointer to the string, which will be modified in-place.

Used to clean up some strings, for instance if they come from the network, we don’t necessarly want system chars to be displayed on the console. Basically it removes all characters with an ASCII code inferior to 32, that is, all system characters. This way, there won’t be any tab, linefeed, or any of such characters left.

Return value: none.

Function: void lw6sys_str_cleanup_ascii7 (char * str)

str: a pointer to the string, which will be modified in-place.

Used to clean up some strings, for instance if they come from the network, we don’t necessarly want system chars to be displayed on the console. Basically it removes all characters with an ASCII code inferior to 32, that is, all system characters. This way, there won’t be any tab, linefeed, or any of such characters left. This function will even remove any character above ASCII 127.

Return value: none.

Function: char * lw6sys_str_reformat (const char * str, const char * prefix, int nb_columns)

str: a pointer to the string we want to modify

prefix: a prefix to put before each line

Reformats a string, that is, insert newline characters in the right places to that it fits in a given number of columns. A prefix is appended at the beginning of each line. Will not handle strings which already contain newline characters perfectly.

Return value: a newly allocated string, must be freed.

Function: void lw6sys_str_reformat_this (char * str, int nb_columns)

str: a pointer to the string we want to modify

Reformats a string, that is, insert newline characters in the right places to that it fits in a given number of columns. This function will modify the buffer so str must be writeable. Will not handle strings which already contain newline characters perfectly.

Return value: none

Function: char * lw6sys_eol ()

Returns the value of EOL, that is, the "end of line" sequence. Will simply return "\n" on UNIX and "\r\n" on Microsoft platforms. Note that while this is convenient to write config and example files, for instance, it’s a bad idea to use this to generate network messages, because this kind of message needs to be platform independant. Thus any network protocol oriented string would use chr(10) and char(13) directly.

Return value: the EOL string, must not be freed.

Function: lw6sys_list_t * lw6sys_str_split (const char * str, char c)

str: a string

c: the delimiter to split with

Splits a string, for instance ’foo,bar’ splited with ’o’ will return ’f’, ” and ’,bar’.

Return value: a list containing 0-terminated strings.

Function: lw6sys_list_t * lw6sys_str_split_no_0 (const char * str, char c)

str: a string

c: the delimiter to split with

Splits a string, ignoring empty ’0-length’ members. For instance ’foo,bar’ splited with ’o’ will return ’f’ and ’,bar’.

Return value: a list containing 0-terminated strings.

Function: lw6sys_list_t * lw6sys_str_split_config_item (const char * str)

str: a string

Splits a string, ignoring empty ’0-length’ members, and using the comma ’,’ as a separator. This is typically usefull for config elements such as backend lists. Only paths need another separator (platform-dependant).

Return value: a list containing 0-terminated strings.

Function: char * lw6sys_str_join (lw6sys_list_t * list, const char * glue)

list: list of strings to join

glue: string to add in-between

Companion function of lw6sys_str_split which will do the contrary and join the string. Here we use a string as the glue/separator, more flexible than a simple char in this case.

Return value: dynamically allocated string

Function: void lw6sys_str_toupper (char * str)

str: the string to modify

Transforms a string to upper case, the pointer must point to modifiable data.

Return value: none, str pointed data modified in-place

Function: void lw6sys_str_tolower (char * str)

str: the string to modify

Transforms a string to lower case, the pointer must point to modifiable data.

Return value: none, str pointed data modified in-place

Function: void lw6sys_str_truncate (char * str, int len)

str: the string to truncate

len: the new length

Truncates a string to the max given length. If truncated to 3, "abcdef" becomes "abc".

Return value: none, str pointed data modified in-place

Function: void lw6sys_str_truncate_middle (char * str, int len, const char * middle)

str: the string to truncate

len: the new length

middle: the string to add in the middle

Truncates a string to the max given length, by truncating the middle of the string, and putting the string middle at this place. Calling it with "abcdefghijk",5,"X" will give "abXjk".

Return value: none, str pointed data modified in-place

Function: char * lw6sys_str_random (int len)

len: the length of the random string to generate.

Generates a random string, this is usefull for testing.

Return value: newly allocated string

Function: char * lw6sys_str_random_words (int len)

len: the length of the random string to generate.

Generates a random string, this is usefull for testing. This version only generates words with alpha-numerical content (letters and digits plus spaces).

Return value: newly allocated string

Function: char * lw6sys_str_random_word (int len)

len: the length of the random string to generate.

Generates a random string, this is usefull for testing. This version generates on single word with alpha-numerical content (letters and digits but no spaces).

Return value: newly allocated string

Function: int lw6sys_str_is_bin (const char * buf, int len)

buf: the buffer to test

len: the length of the buffer

Tests wether a buffer is likely to contain a string. This is not a bulletproof function, just a simple heuristic based estimator.

Return value: 1 if probably binary, 0 if probably text

Function: char * lw6sys_stream_file_to_str (FILE * f)

f: file to get input from, typically stdin

Will read file/stream and return it as a string. This is not for serious stream operation since it will return only when stream is closed, and read all file into memory before doing anything. It’s also limited in size since it uses a fixed length buffer, so this is just for quick testing, typically used by command line switches which are used to test encoding/decoding functions. Do not use it to read a filesystem file, lw6sys_read_file_content is much better.

Return value: newly allocated string.

Function: void lw6sys_stream_str_to_file (FILE * f, char * str)

f: file to receive the string

str: the string to output

Here only for API consistency, will just put string to file (just a simple fprint).

Return value: none.

Function: int lw6sys_test_register (int mode)

mode: test mode (bitmask)

Registers all tests for the libsys module.

Return value: 1 if test is successfull, 0 on error.

Function: int lw6sys_test_run (int mode)

mode: test mode (bitmask)

Runs the sys module test suite, testing most (if not all...) functions. Note that some tests perform file system operations and might therefore fail on a read-only filesystem, or if user permissions are not sufficient.

Return value: 1 if test is successfull, 0 on error.

Function: int lw6sys_test_exec (int argc, const char * [] argv, int mode)

argc: number of args as passed to main

argv: array of args as passed to main

mode: 0 for check only, 1 for full test

Runs the sys module test which is specific to exec functions, these ones require argc and argv to be correctly set so the extra argument justifies putting it outside lw6sys_test. Additionnally, it’s not fool proof... Moreover, it should be run at the beginning of the program, running it afterwards could give unpredictable results. So it’s safer to use it outside the CUnit standard mechanisms.

Return value: 1 if test is successfull, 0 on error.

Function: lw6sys_thread_handler_t * lw6sys_thread_create (lw6sys_thread_callback_func_t callback_func, lw6sys_thread_callback_func_t callback_join, void * callback_data)

callback_func: the main callback, the function that will run the thread

callback_join: function which will be called when joining, at the end

callback_data: data which will be passed to the callback

Creates a thread. All threads must be joined. This is because we really do not want the game to leak, and detached threads are typically the kind of thing that leaves stuff in the heap. Note that callback_func is just something which will be called when joining it can be NULL. The idea is to put in it free & delete functions, which you can’t call before joining when you want the main thread to get the results of the callback_func.

Return value: an opaque pointer on the thread. Can be NULL if failed.

Function: int lw6sys_thread_is_callback_done (lw6sys_thread_handler_t * thread_handler)

thread_handler: thread to work on

Tells wether the callback is done, that is to say, wether the results are available, and we can join.

Return value: 1 if done, else 0.

Function: int lw6sys_thread_wait_callback_done (lw6sys_thread_handler_t * thread_handler)

thread_handler: thread to work on

Waits until the callback of the thread is done, this does not necessarly mean it’s freed, in fact it’s not at this stage, the join callback can still be yet to call, but at least the main stuff is done.

Return value: 1 if done, 0 on error

Function: int lw6sys_thread_get_id (lw6sys_thread_handler_t * thread_handler)

thread_handler: thread to query

Returns the id of the thread, this is an internal value, unique for each process, which can help identifying the thread.

Return value: the id, should be >0.

Function: void * lw6sys_thread_get_data (lw6sys_thread_handler_t * thread_handler)

thread_handler: thread to query

Returns the data associated to the thread, that is, the pointer which was passed to the callback function.

Return value: a pointer.

Function: void lw6sys_thread_join (lw6sys_thread_handler_t * thread_handler)

thread_handler: thread to end

Joins the thread, that’s to say wait until the thread is over, and destroys the ressources associated to it. Note that if the thread is looping forever, this function will just wait forever. This is the only way to end a thread.

Return value: none.

Function: int lw6sys_get_thread_create_count ()

Utility function used to check how many threads where created and joined.

Return value: how many threads were created.

Function: int lw6sys_get_thread_join_count ()

Utility function used to check how many threads where created and joined.

Return value: how many threads were joined.

Function: int lw6sys_check_thread_count ()

Utility function used to check how many threads where created and joined. This one will compare the results of lw6sys_get_thread_create_count and lw6sys_get_thread_join_count.

Return value: 1 if both are equals, 0 if not (error...).

Function: int64_t lw6sys_get_timestamp ()

Returns a 64-bit timestamp, for general purpose. The unit is milliseconds, should return the number of milliseconds since EPOCH. Don’t use this for accurate date handling, but rather to technical stamp events.

Return value: the timestamp.

Function: int64_t lw6sys_get_uptime ()

Returns the number of milliseconds since program was started. Milliseconds are often referred to as ’ticks’.

Return value: the number of milliseconds (64-bit)

Function: int32_t lw6sys_get_cycle ()

Returns a 32-bit timestamp, which is likely to "loop" and have twice the same value during a single program execution. The idea here is just to provide a 32-bit value, not too big, for animation purposes. The idea is that with 64-bit values, numbers are too big and if the goal is just to animate a cursor or spin a sphere, one does not care if every ten hours there’s a display glitch because value became zero again. Besides, those values are often used for their "rest" in a module operation, to translate textures for instance, and having too big numbers causes floating point imprecisions. In fact those values or even only 20-bit. The function is based on lw6sys_get_uptime so it will return 0 at game startup.

Return value: the cycle value, a 20-bit integer.

Function: void lw6sys_timer_update (int64_t * timestamp, int64_t * uptime, int32_t * cycle)

timestamp: the timestamp in msec since EPOCH (output), can be NULL

uptime: the uptime in msec since startup (output), can be NULL

cycle: a 20-bit value for animation purpose.

Returns timestamp & uptime with only one system call.

Return value: none (parameters modified).

Function: void lw6sys_sleep (float seconds)

seconds: the number of seconds to wait, fractions allowed

Will sleep for the given amount of seconds. Same as lw6sys_delay only input is provided as a floating number of seconds instead of ticks.

Function: void lw6sys_delay (int msec)

msec: the number of milliseconds (ticks) to wait

Will sleep for the given amount of seconds. Provides accurate timing and has "about-millisecond" precision, since it uses usleep or select internally. Might however be interrupted in some cases, so consider function can always return quicker than specified. A common usage of this function is polling loops, where you don’t care if 2 polls are very close, but simply want to avoid polling continuously, therefore consumming 100% of the CPU for nothing.

Function: void lw6sys_idle ()

Will sleep for a minimal amount of time, just giving the OS a chance to let other threads/processes execute themselves. This can make a big difference in polling loops between a process that eats 100% CPU and a process that has a very moderate load. of ticks.

Function: void lw6sys_snooze ()

Will sleep for some time, like lw6sys_idle, except it’s a "longer" time, use this when you don’t really care about reactivity but are more concerned about saving CPU, not running uselessly the same polling code.

Function: void lw6sys_time_init ()

Global initializations required to handle time properly.

Function: char * lw6sys_date_rfc1123 (int seconds_from_now)

seconds_from_now: an offset to add to current time

Gives the date according to RFC1123, this is typically usefull for HTTP protocol.

Return value: newly allocated string.

Function: char * lw6sys_date_clf ()

Gives the date in a format which is compatible with Apache CLF Common Log Format.

Return value: newly allocated string.

Function: char * lw6sys_readable_uptime (int64_t timestamp_delta)

timestamp_delta: the duration to show, in msec

Returns a readable form of an uptime, typically 1d 12:34:06 for one day, 12 hours, 34 min, 6 sec or 7:03:45 for 7 hours, 3 minutes 45 sec.

Return value: newly allocated string

Function: extern char * lw6sys_url_http_from_ip_port (const char * ip, int port)

ip: IP address

port: IP port

Returns an http URL pointing to ip: port that is, adds a heading http:// and a trailing /, and treats port 80 as default. This is used to create public_url in net modules.

Return value: a newly allocated string, NULL on error.

Function: lw6sys_url_t * lw6sys_url_parse (const char * url)

url: the URL to parse

Parses a URL, this is not a complete RFC compliant parser, it’s only used to transform URLs into their ’canonical’ form as well as getting basic info such as on which port one should connect.

Return value: a newly allocated struct, NULL on error

Function: void lw6sys_url_free (lw6sys_url_t * url)

url: the url struct to free

Frees a URL struct and all its members.

Return value: none.

Function: char * lw6sys_url_canonize (const char * url)

url: the url to check & transform

Checks if a given URL is correct and, if it is, transforms it into its canonical form. This is mostly to get rid of typesettings error, add a tailing /, transform all domain into lowercase, among other things. A canonized url passed into this function should come out exactly the same.

Return value: a newly allocated string.

Function: int lw6sys_url_is_canonized (const char * url)

url: the URL to check

Checks wether an URL is in its canonized form.

Return value: 1 if OK (canonized form), 0 if not

Function: int lw6sys_version_is_compatible (const char * version_a, const char * version_b)

version_a: 1st version to compare

version_b: 2nd version to compare

Compares two versions and tells wether they are compatible or not. Actually, it only checks that MAJOR.MINOR is the same in both cases. As a side not, it’s not case sensitive. In most LW6 relevant cases, it’s a moot issue since MAJOR.MINOR is a number, but well, just in case.

Return value: 1 if compatible, 0 if not

Function: int lw6sys_vthread_run (lw6sys_thread_callback_func_t callback_func, lw6sys_thread_callback_func_t callback_join, void * callback_data)

callback_func: the main callback, the function that will run the thread

callback_join: function which will be called when joining, at the end

callback_data: data which will be passed to the callback

This function is similar to lw6sys_thread_create, but it’s dedicated to creating a unique (one per process only) thread, which, in turn, will be able to run commands in the main thread itself. This is a hack to allow apparently spawned child threads to be actually handled by main. This is because some libraries, which LW6 uses in threads, need to be actually called in the main thread. SDL, for instance. Note that after running this you loose control on the main thread, this one will only wait for possible commands from the spawned thread, typically sent with the lw6sys_vthread_create function.

Return value: 1 on success, 0 on failure.

Function: int lw6sys_vthread_is_running ()

Returns true if lw6sys_vthread_run has been called. Note that this is not bullet proof, it will return true in a correct manner only if you call it from the vthread itself. In practise this shouldn’t be a problem, the idea is just to write portable code for the main control thread and be able to decide on the fly wether to create a thread we should prefer the lw6sys_thread_create or its equivalent the lw6sys_vthread_create function.

Return value: 1 on success, 0 on failure.

Function: int lw6sys_vthread_create (lw6sys_thread_callback_func_t callback_func, lw6sys_thread_callback_func_t callback_join, void * callback_data)

callback_func: the main callback, the function that will run the thread

callback_join: function which will be called when joining, at the end

callback_data: data which will be passed to the callback

The equivalent of lw6sys_thread_create but for the vthread infrastructure. The idea is to pretend firing a spawned thread, but in fact it’s the main thread that runs the code. This function must imperatively be called within the lw6sys_vthread_run function, else it will fail or be buggy.

Return value: 1 on success, 0 on failure.

Function: void lw6sys_vthread_join ()

The equivalent of lw6sys_thread_join but for the vthread infrastructure. The idea is to pretend firing a spawned thread, but in fact it’s the main thread that runs the code. This function must imperatively be called within the lw6sys_vthread_run function, else it will fail or be buggy.

Return value: none.

Struct: lw6sys_assoc_s

Assoc is a basic key/pair structure where key is a string. Use it for basic associations, it’s not fast when there are many keys, in that case, prefer a hash.

Member of lw6sys_assoc_s: key

Type: char *

Definition: char* lw6sys_assoc_s::key

The key, a 0 terminated standard C string.

Member of lw6sys_assoc_s: value

Type: void *

Definition: void* lw6sys_assoc_s::value

The value, pointer to arbitrary data.

Member of lw6sys_assoc_s: free_func

Type: lw6sys_free_func_t

Definition: lw6sys_free_func_t lw6sys_assoc_s::free_func

This function will be called whenever the element is deleted. You can set it to NULL in that case no callback will be called on deletion.

Member of lw6sys_assoc_s: next_item

Type: lw6sys_assoc_p

Definition: lw6sys_assoc_p lw6sys_assoc_s::next_item

Pointer on the next item, will be NULL on last element, there’s a difference between a NULL pointer and a valid assoc with only one item being EOL.

Struct: lw6sys_cache_item_s

Cache item is the object used to hold data within hash, to implement cache features. It basically stores a pointer to the actual data, and a timestamp which marks the expiration time. In practice, a cache is just an hash which contains this kind of data.

Member of lw6sys_cache_item_s: expiration_timestamp

Type: int64_t

Definition: int64_t lw6sys_cache_item_s::expiration_timestamp

Expiration time, after this time, key is considered invalid.

Member of lw6sys_cache_item_s: real_free_func

Type: lw6sys_free_func_t

Definition: lw6sys_free_func_t lw6sys_cache_item_s::real_free_func

OK, now this requires some explanation : to use standard hash / assoc function we need the cache hash to behave like a real hash. So the trick is to store within the data structure the pointer on the real free callback. This way the special cache_free callback will have a way to call the genuine free function before destroying the cache container. This duplicates the pointer, but avoids code duplication. In practice caches shouldn’t be that big anyway, so it won’t eat up all your memory anyway.

Member of lw6sys_cache_item_s: value

Type: void *

Definition: void* lw6sys_cache_item_s::value

The actual value.

Struct: lw6sys_cache_s

Cache is an object based on which works pretty much the same but adds the possiblity to give an expiration time to a key. Any key with an expiration time in the past will be removed on query and appear as non-existing to callers.

Member of lw6sys_cache_s: delay_msec

Type: int

Definition: int lw6sys_cache_s::delay_msec

Delay in milliseconds before a key expires.

Member of lw6sys_cache_s: real_free_func

Type: lw6sys_free_func_t

Definition: lw6sys_free_func_t lw6sys_cache_s::real_free_func

The real free_func to call on objects.

Member of lw6sys_cache_s: data

Type: lw6sys_hash_t *

Definition: lw6sys_hash_t* lw6sys_cache_s::data

The actual data.

Struct: lw6sys_color_8_s

Used to store colors when representing them in RGBA mode with integers ranging from 0 to 255.

Member of lw6sys_color_8_s: r

Type: u_int8_t

Definition: u_int8_t lw6sys_color_8_s::r

Red [0 ... 255].

Member of lw6sys_color_8_s: g

Type: u_int8_t

Definition: u_int8_t lw6sys_color_8_s::g

Green [0 ... 255].

Member of lw6sys_color_8_s: b

Type: u_int8_t

Definition: u_int8_t lw6sys_color_8_s::b

Blue [0 ... 255].

Member of lw6sys_color_8_s: a

Type: u_int8_t

Definition: u_int8_t lw6sys_color_8_s::a

Alpha [0 ... 255]. 255 is opaque, 0 is transparent.

Struct: lw6sys_color_f_s

Used to store colors when representing them in RGBA mode with floats ranging from 0.0f to 1.0f.

Member of lw6sys_color_f_s: r

Type: float

Definition: float lw6sys_color_f_s::r

Red [0 ... 1.0f].

Member of lw6sys_color_f_s: g

Type: float

Definition: float lw6sys_color_f_s::g

Green [0 ... 1.0f].

Member of lw6sys_color_f_s: b

Type: float

Definition: float lw6sys_color_f_s::b

Blue [0 ... 1.0f].

Member of lw6sys_color_f_s: a

Type: float

Definition: float lw6sys_color_f_s::a

Alpha [0 ... 1.0f]. 1.0f is opaque, 0.0f is transparent.

Struct: lw6sys_color_hsv_s

Used to store colors when representing them in HSV mode with floats ranging from 0.0f to 1.0f. An alpha channel has been added so this is more HSVA than HSV.

Member of lw6sys_color_hsv_s: h

Type: float

Definition: float lw6sys_color_hsv_s::h

Hue [0 ... 360.0f]. 0.0f is red, 120.0f is green, 240.0f is blue.

Member of lw6sys_color_hsv_s: s

Type: float

Definition: float lw6sys_color_hsv_s::s

Saturation [0 ... 1.0f].

Member of lw6sys_color_hsv_s: v

Type: float

Definition: float lw6sys_color_hsv_s::v

Value [0 ... 1.0f].

Member of lw6sys_color_hsv_s: a

Type: float

Definition: float lw6sys_color_hsv_s::a

Alpha [0 ... 1.0f]. 1.0f is opaque, 0.0f is transparent.

Struct: lw6sys_hash_s

Hash is a basic hash structure, relying on assoc for implementation. Actually, what it does is storing an array of assoc, the number of assoc elements is given at construction. Then when accessing a member, a quick checksum is made from the key, which enables finding out which assoc must be queried. If the hash is properly sized, then once one has found the right assoc, finding the right key is fast, since there are only a few of them in each assoc, and it avoids scanning for for all keys, which is the very purpose of the hash.

Member of lw6sys_hash_s: size

Type: int

Definition: int lw6sys_hash_s::size

Number of assoc used for this hash, passed at construction.

Member of lw6sys_hash_s: entries

Type: lw6sys_assoc_t **

Definition: lw6sys_assoc_t** lw6sys_hash_s::entries

Array of assoc holding the actual data.

Member of lw6sys_hash_s: free_func

Type: lw6sys_free_func_t

Definition: lw6sys_free_func_t lw6sys_hash_s::free_func

This function will be called whenever the element is deleted. You can set it to NULL in that case no callback will be called on deletion.

Struct: lw6sys_hexa_serializer_s

The hexa (for hexadecimal) serializer is a tool used to simplify serialization processes, you can just push/pop basic data types on it, it will concatenate the string, allocate memory, do all this dirty stuff without requiring you to plan the size of the buffer, among other things.

Member of lw6sys_hexa_serializer_s: buf

Type: char *

Definition: char* lw6sys_hexa_serializer_s::buf

Data buffer.

Member of lw6sys_hexa_serializer_s: buf_size

Type: int

Definition: int lw6sys_hexa_serializer_s::buf_size

Size of data buffer, in bytes.

Member of lw6sys_hexa_serializer_s: pos

Type: int

Definition: int lw6sys_hexa_serializer_s::pos

Current position within the buffer, this is, typically, the place where data will be appended at the next push call, or where it will be fetched from at the next pop call.

Struct: lw6sys_list_s

List is a basic list system, with a void * pointer to hold arbitrary data and a callback function for deletion. Provides basic functions to push, pop, walk, any array-like call will of course be very slow. As of current implementation, front operations are fast, but back operations are slow.

Member of lw6sys_list_s: data

Type: void *

Definition: void* lw6sys_list_s::data

Opaque pointer on element data.

Member of lw6sys_list_s: free_func

Type: lw6sys_free_func_t

Definition: lw6sys_free_func_t lw6sys_list_s::free_func

This function will be called whenever the element is deleted. You can set it to NULL in that case no callback will be called on deletion.

Member of lw6sys_list_s: next_item

Type: lw6sys_list_p

Definition: lw6sys_list_p lw6sys_list_s::next_item

Pointer on the next item, will be NULL on last element, there’s a difference between a NULL pointer and a valid list with only one item being EOL. Other way to state it: NULL and empty list are two different things.

Struct: lw6sys_module_pedigree_s

Structure used to store informations about a module. This describes the module, its author license, this is both a legal check and a technical check, to maximize the chances the code we’re running is the right one, and to trace it.

Member of lw6sys_module_pedigree_s: id

Type: char *

Definition: char* lw6sys_module_pedigree_s::id

Module id, for instance, could be "gl1".

Member of lw6sys_module_pedigree_s: category

Type: char *

Definition: char* lw6sys_module_pedigree_s::category

Module category, for instance, could be "gfx".

Member of lw6sys_module_pedigree_s: name

Type: char *

Definition: char* lw6sys_module_pedigree_s::name

Module name, readable (displayable) name.

Member of lw6sys_module_pedigree_s: readme

Type: char *

Definition: char* lw6sys_module_pedigree_s::readme

Module readme text.

Member of lw6sys_module_pedigree_s: version

Type: char *

Definition: char* lw6sys_module_pedigree_s::version

Module version.

Member of lw6sys_module_pedigree_s: copyright

Type: char *

Definition: char* lw6sys_module_pedigree_s::copyright

Module (short) copyright information.

Member of lw6sys_module_pedigree_s: license

Type: char *

Definition: char* lw6sys_module_pedigree_s::license

Module (short) license.

Member of lw6sys_module_pedigree_s: date

Type: char *

Definition: char* lw6sys_module_pedigree_s::date

Date of module compilation.

Member of lw6sys_module_pedigree_s: time

Type: char *

Definition: char* lw6sys_module_pedigree_s::time

Time of module compilation.

Struct: lw6sys_mutex_s

Mutex is our own wrapper on the pthread mutex object. Why not use the pthread mutex directly? For debugging, this allows us to place and instrument hooks if needed.

Member of lw6sys_mutex_s: id

Type: u_int32_t

Definition: u_int32_t lw6sys_mutex_s::id

The id of the object, this is non-zero and unique within one run session, incremented at each object creation.

Struct: lw6sys_progress_s

Structure used to store progress information. The idea is that is that must be usable in polling mode or in multithreaded mode, and we must be able to truncate a progress indicator into several stages. So this structure contains a range start, a range end, and its value between those two, which is meant to be written by the code executing the operation and read by the caller/rendering thread.

Member of lw6sys_progress_s: min

Type: float

Definition: float lw6sys_progress_s::min

Where the progress operation starts.

Member of lw6sys_progress_s: max

Type: float

Definition: float lw6sys_progress_s::max

Where the progress operation ends.

Member of lw6sys_progress_s: value

Type: float *

Definition: volatile float* lw6sys_progress_s::value

Somewhere between min and max.

Struct: lw6sys_spinlock_s

Spinlock is our own wrapper on a spinlock based mutex. Why not use the pthread spinlock directly? For debugging, this allows us to place and instrument hooks if needed. Additionnally, some implementations of pthread do not provide spinlock and in that case we provide our own alternative.

Member of lw6sys_spinlock_s: id

Type: u_int32_t

Definition: u_int32_t lw6sys_spinlock_s::id

The id of the object, this is non-zero and unique within one run session, incremented at each object creation.

Struct: lw6sys_thread_handler_s

Thread handler is our own wrapper on the pthread object. Why not use the pthread handler directly? Basically to store basic flags and context data (void * pointer on our thread data for instance) along with the handler. This is merely for debugging and convenience. Internally this will be casted to _lw6sys_thread_handler_t.

Member of lw6sys_thread_handler_s: id

Type: u_int32_t

Definition: u_int32_t lw6sys_thread_handler_s::id

The id of the object, this is non-zero and unique within one run session, incremented at each object creation.

Struct: lw6sys_url_s

Describes an URL, with its elements splitted, this is just to be able to use half-parsed URLs and avoid redoing this parsing everytime.

Member of lw6sys_url_s: use_ssl

Type: int

Definition: int lw6sys_url_s::use_ssl

1 if in https, 0 if in http.

Member of lw6sys_url_s: host

Type: char *

Definition: char* lw6sys_url_s::host

Host name.

Member of lw6sys_url_s: port

Type: int

Definition: int lw6sys_url_s::port

IP port.

Member of lw6sys_url_s: uri

Type: char *

Definition: char* lw6sys_url_s::uri

URI, that is, everything after the port.

Struct: lw6sys_whd_s

Contains the shape of a 3D box. There are 3 differences with its "XYZ" equivalent. First, sometimes w*h*d reads better than x,y,z. Then, xyz is signed, whd is unsigned. Finally, these are real int32 values, they are not 14-bit limited. It does not really cost any memory for it’s usually used as a single "shape" attribute for a whole map. At the same time, it’s very often used as a test value in loops, so it’s interesting to have it in a value that’s easy to optimize for the compiler (exactly one register...)

Member of lw6sys_whd_s: w

Type: u_int32_t

Definition: u_int32_t lw6sys_whd_s::w

Width.

Member of lw6sys_whd_s: h

Type: u_int32_t

Definition: u_int32_t lw6sys_whd_s::h

Height.

Member of lw6sys_whd_s: d

Type: u_int32_t

Definition: u_int32_t lw6sys_whd_s::d

Depth.

Struct: lw6sys_xyz_s

All-in 32 bit 3D position, to save memory.It’s a deliberate choice in Liquid War to handle "limited size" levels. In fact 14 bits still allows 8000x8000 maps, which are at least 100 times too slow to play now (2008). Should we follow Moore’s law we’d have at least 6 years until those are playable, and well, until then, let’s wait. The point is that storing this information (x*y) on 4 bytes might be very important in some cases, since it can reduce memory footprint on structs which are stored in numerous quantities, and therefore maximize chances that we use level 1 &amp; 2 caches and other nice things which happen when memory consumption is not too high.Point is: why use INT32 and then limit it to 14 bits instead of using an INT16 or short in the first place? Answer: it’s easier to handle INT32 all the time in the rest of the code. Compiler and CPU might even handle that better than short. Then, and only when data will be read/written in the struct will it be truncated. Typical example is: we want to multiplicate y by w (which is a width). Result is beyond INT16/short scope but we want to handle it! Casting everything to INT32/int is a pain. With this int y:14 trick, we use y as a "full-featured" INT32/int and well, when it will be read/written we’ll loose values over 8191, but we simply do not care.

Member of lw6sys_xyz_s: x

Type: int32_t

Definition: int32_t lw6sys_xyz_s::x

X position, from -8192 to +8191.

Member of lw6sys_xyz_s: y

Type: int32_t

Definition: int32_t lw6sys_xyz_s::y

Y position, from -8192 to +8191.

Member of lw6sys_xyz_s: z

Type: int32_t

Definition: int32_t lw6sys_xyz_s::z

Z position, from -8 to +7.


Next: , Previous: , Up: C API   [Contents][Index]