[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Automake 1.4j: Release Candidate [PATCH]

From: Richard Boulton
Subject: Re: Automake 1.4j: Release Candidate [PATCH]
Date: Tue, 31 Jul 2001 22:33:13 +0100
User-agent: Mutt/1.2.5i

On Tue, Jul 31, 2001 at 11:06:01AM -0600, Tom Tromey wrote:
> Lars> Here's one.  The following doesn't generate any object
> Lars> dependencies.  $(libtest_a_OBJECTS) is always empty.
> Thanks.  I turned this into a test case.  I'll fix it asap.

For what it's worth, and since I've looked at the dependency code fairly
recently, I took a look at this, and I think its the fault of the
@parent_conds argument in &variable_conditions_sub().  This actually looks
like outdated logic to me - duplicate or impossible conditionals are
removed later by &variable_conditions_reduce().

Removing the @parent_conds machinery and simply recursing through the
conditional variables and taking all the possible conditional combinations
makes sense to me, and works correctly as far as I can tell.  I attach a
patch to do this.  All tests behave as expected after this patch, except
for cond13.test which passes. :)

This may make things slightly less efficient, but the whole thing needs
replacing with something which doesn't do a combinatorial explosion as
the number of conditionals increases, anyway.  (As previously discussed on
the list.)


Attachment: diffs
Description: Text document

reply via email to

[Prev in Thread] Current Thread [Next in Thread]