Restructure the tree in a sane way
linux/dev. But only if using a sane RCS, so leave it
as-is for now. Also, a bunch of (header) files from there may probably be
Remove dead files from the GNU Mach source tree
For exported files (via
make install), the plan is to first stick some
#error This file is scheduled for removal. Write to <email@example.com> if you
have a reason to have it kept available. into them, and then actually remove
them after some months.
For some of the internal header files (containing function prototypes and the
like), it might actually be useful to use them. (And then get rid of a bunch
extern ... statements in other files.)
This following list was assembled by putting such a
#error ... line into each
gnumach-1-branch's header files (exported and internal; save the
linux/ ones (only internal) for simplicity), and then trying to build GNU
Mach until this would succeed again (by removing offending
#error ...s), and
afterwards using the set of exported files for building a cross toolchain
(again still removing offending
#error ...s). A very crude and imprecise
So, additionally to the list given below, there may actually be a bunch of
further files (also exported ones) that serve no real value, but are being
#included through one way or another.
Currently used, but copyright violation? Rewrite?
Copyright violation? Currently unused, but could be used in principle (or be rewritten, to avoid the copyright oddity).
Copyright violation. Unused. Remove.
Might be usable for SMP? Remove otherwise.
Was probably once exported, but is no longer.
Currently not exported, but was probably once meant to be.
Remove dead functions, variables, etc. from source files
Rewrite ugly code
IRC, freenode, #hurd, 2012-09-06
<mcsim> hello. Why size parameter of rpc device_read has type "mach_msg_type_number_t *"? Why not just "vm_size_t *"? <mcsim> this parameter has name data_count <braunr> that's one of the reasons mach is confusing <braunr> i can't really tell you why, it's messy :/
Code Analysis, Coverity
IRC, freenode, #hurd, 2014-02-03
<teythoon> braunr: at the end of kmem_cache_error the comment says 'never reached'. i do not believe that to be true... <braunr> teythoon: yes, this is a mistake <braunr> those kmem_error calls were panics in the original version <braunr> the comment was left over <braunr> or rather, the panic equivalent not added
IRC, freenode, #hurd, 2014-02-04
<teythoon> braunr: but the code relies uppon the fact that kmem_cache_error does not return <teythoon> the fact that it doesn't upsets static analysis tools <teythoon> http://darnassus.sceen.net/~teythoon/qa/gnumach/scan-build/2014-01-17/report-DfjC43.html#Path9 <braunr> 21:20 < braunr> or rather, the panic equivalent not added <braunr> teythoon: you're right <braunr> it needs to be fixed <teythoon> i setup a coverity project for gnu mach <teythoon> it's interesting <teythoon> very enterprisy >,< <braunr> teythoon: heh <teythoon> braunr: i assume kmem_error is supposed to panic <teythoon> if so, i'd propose to use panic instead of printf in that macro definition <braunr> yes <braunr> teythoon: again, that's how it's done in the original implementation <teythoon> and fix kmem_cache_error to use kmem_warn near the top of the function <braunr> teythoon: i suggest you use the x15 variant as a reference (https://git.sceen.net/rbraun/x15.git/plain/kern/kmem.c) <teythoon> right
IRC, OFTC, #debian-hurd, 2014-02-05
<pere> ah, cool. <URL: https://scan.coverity.com/projects/1307 > is now in place. :) <teythoon> pere: it was you who made me create the coverity project <teythoon> and it was most fruitful already <pere> glad to hear that. :) <pere> while the mach thread made me discover scan-build, which seem promising but have way to few rules. <pere> teythoon: btw, could I get access to the mach coverity status? <teythoon> pere: sure <pere> I've used up all my 'join project' slots, so I suspect you have to invite me. try firstname.lastname@example.org <teythoon> pere: ok <teythoon> pere: i'm fuzzy about the roles <teythoon> what kind of invite shall i send you ? <pere> not sure either. <teythoon> ok <pere> contributor/* would be fine for me. want to check the details, not just the overview. <teythoon> oh * pere looked up the role description <teythoon> do you have a link ? <teythoon> i can always change the role afaics <pere> go to the member list, found it there. <pere> (the link is project specific, it seem. <teythoon> right, found it <pere> gah, it gave me a new user. :( <pere> I assumed I would be allowed to accept the invite using my normal user. <pere> can you try again with email@example.com, which hopefully will hook the invite to my github authenticated user? <teythoon> sure <pere> were you allowed to invite that address? <pere> ah, I got it in my dashboard without having to do anything more. <pere> heh, the email I got claimed the user would not be created unless I provided a password, but given the member list for mach, I fail to see how that could be true.
IRC, OFTC, #debian-hurd, 2014-02-06
<teythoon> pere: your coverity invitations are still saying 'not accepted', did neither work ? <pere> teythoon: one worked just fine. I got access. the other I did not accept. <teythoon> pere: good :)
IRC, OFTC, #debian-hurd, 2014-02-25
<pere> I am quite impressed with the defect density of Mach, <URL: https://scan.coverity.com/projects/1307 >. :) <pere> 12 outstanding issues (but 54 dismissed was surprisingly high :). <youpi> pere: 54 dismissed because that's a pattern generated by mig <youpi> repeated dozens of times