This is the mail archive of the
libc-hacker@cygnus.com
mailing list for the glibc project.
Suspended bug reports
- To: libc-hacker@cygnus.com
- Subject: Suspended bug reports
- From: Zack Weinberg <zack@rabi.columbia.edu>
- Date: Wed, 03 Feb 1999 20:20:43 -0500
These are the suspended bug reports for glibc; now we are done with
2.1, it's time to revisit them.
* Localedef bugs: 207, 454, 838.
Localedef rewrite is already planned.
* Pthreads bugs: 798, 836.
Pthreads rewrite is already planned. (Mark, Thomas: what
are the odds this will be done for 2.2?)
* Malloc and huge blocks: 723, 930.
Why are signed types used internally?
* Network header issues: 153, 247.
POSIX doesn't specify these, Single Unix only has a tiny bit
of netinet/in.h.
The best policy, IMO, would be to mimic 4.4BSD as closely as
possible. This is what everyone else does, so it maximizes
compatibility. It would break backward source compat with
some Linux-specific code; this can be handled by a feature
select switch.
* Pthreads enhancement request: 857.
Would be obsoleted by better kernel-level support for POSIX
thread semantics; should wait for the pthreads rewrite; dunno
how generally useful it would be.
* Stdio debugging-aid request: 436.
I like this idea, it is useful and it would be analogous to
other debug-related envariables we have already. I think it
didn't get implemented out of lack of interest - did we ever
discuss it?
* Using libc from nonstandard location: 269.
Is really a gcc issue. The idea I said I had last year didn't
pan out. A general fix would take a total rewrite of the gcc
driver and its spec language; possibly a good idea, not likely
to happen anytime soon.
Has anyone tried configuring+installing gcc and glibc with the
same nonstandard prefix? That might actually work. If it
doesn't, it's probably feasible to make it work.
zw