[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: 02-define-exts.patch
From: |
Tom Tromey |
Subject: |
Re: 02-define-exts.patch |
Date: |
04 Mar 2001 23:52:46 -0700 |
Akim> If OBJEXT and EXEEXT are not set, provide a default value, and use
Akim> them unconditionally.
This is ok, but it is unfortunate.
Currently the EXEEXT support works by rewriting _PROGRAMS variables
and relying on make to pick up the second definition. This is ugly.
However given the old implementation there wasn't a better approach.
(You can see this in action by running cygwin32.test and then looking
for bin_PROGRAMS in the generated Makefile.in.)
With this patch we'll be doing this all the time instead of only in a
certain obscure situation. Bummer.
The long term fix is probably to make automake generate variable
definition text later. Then we can modify variable definitions in
place (in %contents or %conditional) and have only the modified
version printed.
Akim> +++ tests/suffix2.test Sun, 04 Mar 2001 18:12:59 +0100 akim
(am/34_suffix2.te 1.1 775)
Akim> -test "`fgrep '.c.o' Makefile.in | wc -l`" -eq 1
Akim> +grep '^\.c\.o:' Makefile.in
Akim> +grep '^\.c\.obj:' Makefile.in
This change doesn't quite work:
* It tests for a different problem. The old problem was that we could
end up with more than one .c.o rule -- with the change we no longer
check for that.
* If the first grep fails the test case won't fail.
You need an `|| exit 1'
Tom
- 02-define-exts.patch, Akim Demaille, 2001/03/04
- Re: 02-define-exts.patch,
Tom Tromey <=
- Re: 02-define-exts.patch, Akim Demaille, 2001/03/05
- Re: 02-define-exts.patch, Tom Tromey, 2001/03/05
- Re: 02-define-exts.patch, Akim Demaille, 2001/03/05
- Re: 02-define-exts.patch, Tom Tromey, 2001/03/05
- Re: 02-define-exts.patch, Akim Demaille, 2001/03/05