Re: Re: musl bugs found through gnulib

[Rich Felker <dalias@aerifal.cx>]

On Mon, Jun 18, 2012 at 12:49:44AM +0200, Bruno Haible wrote:
> Replacements of *printf, because of
>   checking whether printf supports infinite 'long double' arguments... no

I'm just going to change isnanl to handle the invalid bit patterns
because I'm sick of this issue...

>   checking whether printf supports the 'ls' directive... no

Fixed in git on Jun 8.

>   checking whether printf survives out-of-memory conditions... no

This test must be doing something wrong. musl's printf does not use
any dynamic memory.

> Replacement of duplocale, because of
>   checking whether duplocale(LC_GLOBAL_LOCALE) works... no

I'll look into this. Locales are not presently supported, but the
interfaces are intended to work and behave as they should for an
implementation that has exactly one possible locale.

> Replacement of fdopen, because of
>   checking whether fdopen sets errno... no

I'll look into it.

> Replacement of futimens, because of
>   checking whether futimens works... no

Odd, this is just a syscall wrapper, no?

> Replacement of getcwd, because of
>   checking whether getcwd handles long file names properly... no, but it is partly working
>   checking whether getcwd aborts when 4k < cwd_length < 16k... no

I'm not sure if it's related, but musl defines PATH_MAX and does not
attempt to support explicit pathnames longer than PATH_MAX. I'll have
to look at what the test is doing.

> Replacement of getopt, because of
>   checking whether getopt is POSIX compatible... no

This was a bug I reported to gnulib. The test is checking for GNU
semantics which are incompatible with POSIX, i.e. the test is
misnamed.

> Replacement of glob, because of
>   checking for GNU glob interface version 1... no
> (not sure this is a bug or just an incompatibility compared to glibc)

I think gnulib just wants GNU glob functionality. This is legitimate
for it to replace since we (at present, at least) don't attempt to
offer any GNU functionality in glob.

> Replacement of iconv and iconv_open, because of
>   checking whether iconv supports conversion between UTF-8 and UTF-{16,32}{BE,LE}... no

These conversions are definitely supported. The only related thing we
don't support is BOMs and stateful/endian-detected UTF-16/32.

> Replacement of mktime, because of
>   checking for working mktime... no

I need to check this. I've suspected it might have a bug.

> Replacement of perror, because of
>   checking whether perror matches strerror... no

Odd, it uses strerror..

> Replacement of popen, because of
>   checking whether popen works with closed stdin... no

Sounds like a bug; I'll check it out.

> Replacement of regex, because of
>   checking for working re_compile_pattern... no

This is a check for the GNU regex interfaces as opposed to POSIX
regcomp/regexec.

> Replacement of strtod, because of
>   checking whether strtod obeys C99... no

This was caused by a missing negative zero case for "-0x" followed by
non-hex chars. It's fixed in git.

> For each of the replacements, first look at the test program's results
> (in config.log), then look at the test program's source code (in m4/*.m4).

OK, looking...

Rich