Thank you very much Paul,
I just pulled from Gnulib and built Gnuastro (both the most recent
commits at this moment) and shown below. But the problem remained.
Gnulib: 90e79512d8b385801218d6e9c4d88ff77186560b
Gnuastro: c70f16c8f13ccdb8ad606a33a526846180748284
Just for the record, I am using GCC 10.2.0 and Glibc 2.32 and this is
the last working Gnulib commit that I have checked:
Gnulib (working): e2739ba6310893be93d01a23cbfed8d8dfb08966 (Nov 11)
Since 'stdio.h' is a pretty basic library in Gnulib, and it is loading
everything else properly, it may take some time for me to check (because
I have dig into the low-level aspects of Gnulib which I am not too
familiar with). Any tips/pointers would be appreciated.
I guess as a start, its easier for me to find the approximate commit
that it is introduced in, I'll send you the results when I find it.
Cheers,
Mohammad
On 11/27/20 4:49 AM, Paul Eggert wrote:
On 11/25/20 5:09 AM, Mohammad Akhlaghi wrote:
I just pulled the recent changes from Gnulib's 'master' branch and
bootstrapped Gnuastro (which uses Gnulib) on my Arch GNU/Linux
operating system.
But I got many warnings related to '_GL_ATTRIBUTE_SPEC_PRINTF_SYSTEM'
in 'error.h' and 'argp.h'.
I'm not seeing this with Fedora 33 x86-64 (see attached). I used
gnuastro commit c70f16c8f13ccdb8ad606a33a526846180748284 (Wed Nov 25
21:32:37 2020 +0000) and Gnulib commit
90e79512d8b385801218d6e9c4d88ff77186560b (Thu Nov 26 22:13:20 2020 +0100).
I see that Bruno recently fixed some problems in this area in Gnulib,
and possibly this is contributing to the problem, whatever it is.
I suggest bootstrapping from scratch with the latest Gnulib, and (if
that doesn't work) investigating why bootstrapped/lib/stdio.h's
definition of _GL_ATTRIBUTE_SPEC_PRINTF_SYSTEM isn't working for you.
