0.9.30.1 review - svn 24489
Bernhard Reutner-Fischer
rep.dot.nop at gmail.com
Sun Feb 15 23:37:26 UTC 2009
On Sat, Feb 14, 2009 at 05:16:09PM -0600, Rob Landley wrote:
>Ok, 24489 is a cut and paste of svn 24487 (because subversion branching is
>stupid). It claims to "fix fpclassify, signbit, isfinite, isnan, isinf macros
>for long double math", but doesn't say what that _means_. It's about 1200
>lines long, uses #if 0 to chop out a block at least once, and adds a lot of
>wrappers ala:
>
>+#ifdef L_cargl
>+long double cargl (long double complex x)
>+{
>+ return (long double) carg( (double complex)x );
>+}
>+#endif
>
>It also renames several functions (finite()->isfinite()), removes several
>libm_hidden_def() and weak_alias() instances, then we get into the changes in
>Makerules and bits/mathcall.h where I haven't got a clue why it's doing any of
>that.
>
>I don't exactly object to this patch as a bugfix, it probably is. It's just
>that after attempting to review it, I still have no idea what it's a bugfix
>for. Are there packages that now build against uClibc that didn't used to, or
>is it support for more hosts now, or what?
>
>A little more explanation would be nice. It's a bit big for a bugfix-only
>release, so it would be nice to know at least one specific package that needs
>it.
It fixed macro expansion. I think that the changelog was pretty clear:
$ svn log -c24487
------------------------------------------------------------------------
r24487 | aldot | 2008-12-22 12:31:12 +0100 (Mon, 22 Dec 2008) | 3 lines
- fix fpclassify, signbit, isfinite, isnan, isinf macros for long double math support
- add rule to create preprocessor output for float- and long double math wrapper
------------------------------------------------------------------------
If you're curious then revert the patch and write testcases that use these macros.
More information about the uClibc
mailing list