This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [patch] New test+use texinfo @click - @HAVE_MAKEINFO_CLICK@ [Re: doc build failure (Re: [patch 04/12] entryval#3: Virtual tail call frames)]
- From: Pedro Alves <pedro at codesourcery dot com>
- To: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- Cc: Andreas Schwab <schwab at linux-m68k dot org>, gdb-patches at sourceware dot org, Ulrich Weigand <uweigand at de dot ibm dot com>, eliz at gnu dot org
- Date: Wed, 12 Oct 2011 16:26:24 +0100
- Subject: Re: [patch] New test+use texinfo @click - @HAVE_MAKEINFO_CLICK@ [Re: doc build failure (Re: [patch 04/12] entryval#3: Virtual tail call frames)]
- References: <20111011233417.GA17487@host1.jankratochvil.net>
On Wednesday 12 October 2011 00:34:17, Jan Kratochvil wrote:
> On Mon, 10 Oct 2011 23:56:34 +0200, Pedro Alves wrote:
> > On Monday 10 October 2011 19:47:26, Jan Kratochvil wrote:
> > > + if makeinfo conftest.texinfo >&5 2>&5; then
> >
> > Shouldn't this be $(MAKEINFO) instead (like in Makefile.in)?
> > Though I'm not sure how to do that here. Maybe
> > AC_CHECK_PROGS(MAKEINFO, makeinfo) ?
>
> Yes, it should, although it is more complicated.
>
> There is IMO a bug in toplevel configure it does not pass down MAKEINFOFLAGS
> and includes "--split-size=5000000" directly into MAKEINFO. Without the
> special handling of --split-size=5000000 it would be used during first
> compilation but no longer used after automatic re-run of configure in gdb/
> (although only if one deletes gdb/config.cache).
I see. We should get that fixed someday.
> Therefore I added --split-size=5000000 there explicitly, which means it is
> duplicated during the dirst compilation but it does not matter.
>
> --split-size=5000000 was not used in gdb/doc/ before at all but I think it was
> a bug, even gcc uses --no-split.
No opinion on that.
> On Tue, 11 Oct 2011 00:18:54 +0200, Andreas Schwab wrote:
> > You should not modify MAKEINFO. If you say
> > "make info MAKEINFO=/usr/local/bin/makeinfo" you lose the flags.
>
> True, toplevel configure is already violating that but fixed it in gdb/ .
Thanks. Looks good to me.
--
Pedro Alves