Suggestions on good first bugs

Robert O'Callahan robert at ocallahan.org
Thu Mar 5 14:25:05 PST 2015


On Fri, Mar 6, 2015 at 8:53 AM, Daniel Näslund <dannas at dannas.name> wrote:

> I'm looking for good first bugs. Something that I can finish in a few
> evenings that helps me sort of the structure of the code base.
>

Great!


> The TODO wiki page links to
> https://github.com/mozilla/rr/labels/goodfirstbug.
>
> #1153 Remove constraint on srcdir/objdir...
>     I was bitten by this when I first tried to build. I'll make an
>     attempt (though I despise every build system just as much as any
>     other developer).
>
> #1054 Implement remaining syscalls
>     Rocallahan says [3]
>     "I think we shouldn't clutter up rr with support for syscalls that
> no-one
>     ever uses in practice (e.g. because they're obsolete). So I think
>     anytime someone submits a patch supporting a syscall, we require some
>     evidence that it's actually needed (e.g., someone says they hit it
>     recording some program)."
>
>     Any syscalls that we know are needed?
>

One way to proceed here is to try recording different applications and just
fix the bugs that come up. Some of those bugs will be additional syscalls
that need to be supported.

Other suggestions? ===
>
> rocallahan said in [2]:
>
>     "Even just installing an unrelated package could cause problems.
>     Probably we should always just copy /etc/ld.so.cache into the trace so
>     that you can do Unrelated package changes without invalidating traces."
>
>     Is that something that is feasible? Needed?
>

I think it would definitely be worth doing this if it allows unrelated
packages to be installed/removed without breaking replayability of existing
traces. It's really easy to do; see should_copy_mmap_region.

cgjones said in [4]
>
>     "To narrow things down, you can check the "SigCgt" (et al.) entries in
>     /proc//status to see what linux thinks the signal setup is. It would be
>     super great to add another function for rr's "check mode" that ensures
>     rr's view of signal setup is the same as linux's. We do this for the
>     memory map already."
>
>     Is that something that can be done as a first-time project?
>


Sure.

>
> Any other suggestions on first good bugs?
>

I think trying to record and replay different kinds of applications would
be a great way to improve rr and learn about the code.

Rob
-- 
oIo otoeololo oyooouo otohoaoto oaonoyooonoeo owohooo oioso oaonogoroyo
owoiotoho oao oboroootohoeoro oooro osoiosotoeoro owoiololo oboeo
osouobojoeocoto otooo ojouodogomoeonoto.o oAogoaoiono,o oaonoyooonoeo
owohooo
osoaoyoso otooo oao oboroootohoeoro oooro osoiosotoeoro,o o‘oRoaocoao,o’o
oioso
oaonosowoeoroaoboloeo otooo otohoeo ocooouoroto.o oAonodo oaonoyooonoeo
owohooo
osoaoyoso,o o‘oYooouo ofooooolo!o’o owoiololo oboeo oiono odoaonogoeoro
ooofo
otohoeo ofoioroeo ooofo ohoeololo.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/rr-dev/attachments/20150306/2231f745/attachment.html>


More information about the rr-dev mailing list