On Thu, 2015-10-15 at 00:29 -0400, Carlos O'Donell wrote:
> The upstream glibc community has been working hard to rectify these
> issues, and Red Hat in particular spent a lot of time making the
> netgroup caching as bullet-proof as possible. If you have real bugs,
> please file them upstream and come talk to the community.

Sorry, didn't mean te offend anyone. I'm sure nscd has seen a lot of
improvements since the last time I had a look at it. It was very
complex though. Then again caching is one of the two hard things in
computer science ;)

> > Not all NSS interfaces are consistent in terms of memory
> > management, return codes, etc. so there could be something
> > wrong.
> 
> They should be consistent. Again, if you find something wrong, please
> file a bug. The upstream glibc community is radically different now
> and certainly more accepting of discussions on these issues.

At least the _nss_xxx_initgroups_dyn() function is expected to grow the
passed groupsp buffer itself instead of returning NSS_STATUS_TRYAGAIN+
ERANGE like most other functions. The Solaris NSS implementation
(although far from perfect and having many nasty other "features") uses
an initialiser that returns a backend pointer that is passed to all
functions to allow memory and other resource management in the NSS
module itself. Also having a single return code and avoiding messing
with errno would also be nicer.

However, I'm not filing a bug about this because I think keeping a
stable interface at this point is still better than improving the
existing interface.

All your efforts on nscd and libc are very much appreciated!

Anyway, I'm willing to test some patches to see if we can improve the
nscd version in sid (2.19-22) and/or jessie (2.19-18+deb8u1) but I'm
having a hard time reproducing the bug I originally filed.

Thanks,

-- 
-- arthur - adej...@debian.org - http://people.debian.org/~adejong --

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to