[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: even `if FALSE' *_SOURCES should be used for TAGS
From: |
Ralf Wildenhues |
Subject: |
Re: even `if FALSE' *_SOURCES should be used for TAGS |
Date: |
Tue, 12 Oct 2010 23:29:44 +0200 |
User-agent: |
Mutt/1.5.20 (2010-08-04) |
Hello, and thanks for report and analysis!
* Stefano Lattarini wrote on Tue, Oct 12, 2010 at 05:51:58PM CEST:
> On Monday 11 October 2010, Jan Kratochvil wrote:
> > `make tags' will not index sources which are listed in *_SOURCES += ... in
> > unsatisfied AM_CONDITIONAL blocks. IMO they should be.
> I agree with you on this. By the way, "make dist" already works OOTB
> in this situation, so tags (and ctags, and ID, and cscope) should too.
> > Sure one can use some EXTRA_*_SOURCES out of the `if' block but this
> > delocalizes+duplicates the information making error prone for future
> > modifications.
> Right.
>
> > OTOH AM_CONDITIONAL can be used for example if an optional sources part is
> > / is not bundled, in such case the current automake behavior is better.
> Ralf, do you agree we should implement the behaviour requested by Jan?
Well, I see the following caveat:
'make tags' shouldn't cause files to be built (or accessed) that 'make
dist' wouldn't build; that probably means conditional nodist_ files
shouldn't be tagged.
OTOH, automake.in:handle_source_transform is the only place where
@sources and @dist_sources get possibly different values, and I don't
understand why that was done; IOW, why the implementation didn't just
have $(SOURCES) = $(DIST_SOURCES) (except for the possibility of the
no-dist option, of course). I haven't dug out why that was done.
If there is no good reason, the fix would be a one-liner.
Thanks,
Ralf