Re: ☠ Buildbot (Sourceware): binutils-gdb-autoregen - failed 'git diff ...' (failure) (master)

[Alan Modra <amodra@gmail.com>]

On Mon, Jan 08, 2024 at 10:51:46PM +0100, Mark Wielaard wrote:
> Hi Alan,
> 
> On Mon, Jan 08, 2024 at 10:54:09AM +1030, Alan Modra wrote:
> > On Sun, Jan 07, 2024 at 06:29:36PM +0100, Mark Wielaard wrote:
> > > So the autogen tester would like to move these entries back to where
> > > they were before.
> > 
> > So why are the aclocal.m4 dependencies being sorted differently?  It
> > seems to me that what is currently in libcollector/aclocal.m4:
> > 
> > m4_include([../../config/depstand.m4])
> > m4_include([../../config/lead-dot.m4])
> > m4_include([../../config/override.m4])
> > m4_include([../../config/warnings.m4])
> > m4_include([../../libtool.m4])
> > m4_include([../../ltoptions.m4])
> > m4_include([../../ltsugar.m4])
> > m4_include([../../ltversion.m4])
> > m4_include([../../lt~obsolete.m4])
> > 
> > is sorted correctly.  This is the result I get by configuring with
> > --enable-maintainer-mode, and until I see others getting a different
> > result with --enable-maintainer-mode and the "correct" versions of
> > auto tools, I'm going to say the tester is doing something wrong.
> 
> You are right. There is a difference between what
> --enable-maintainer-mode does and what autoregen.py does.
> autoregen.py doesn't fully handle the (order of directories) of
> AC_CONFIG_MACRO_DIRS in gprofng/libcollector/configure.ac.

Yes, I squinted at that one a little after I posted my previous
email, and found that swapping the order of AC_CONFIG_MACRO_DIRS args
didn't help..

> We can make this the same by using the following hack in autoregen.py

Thanks for fixing this.  

-- 
Alan Modra
Australia Development Lab, IBM