François Bissey wrote:
The NEWS file for make 3.81 doesn't have anything that I can really relate
to the problem. However it would be nice to check if upgrading make on your
machine solves this particular problem. Or did I miss that someone else had
troubles with 3.81?
Not in this context. At some point we had trouble with 3.81 and 'make
ptest[long]' (GAP and/or ECL IIRC), which did not happen with 3.82
(although again, GNU Make's changelog or NEWS didn't mention anything
obviously relevant).
I don't think the *warnings* Karl-Dieter gets with 3.80 cause any harm
(other than probably annoying the user). The changes in 3.81 suppress
them, but do not change the behaviour w.r.t. '[-]include' and *empty*
include file lists; they mainly deal with how 'make' behaves if the
files to include don't exist and/or cannot be remade (according to [1]
and [2] at least).
-leif
[1] http://savannah.gnu.org/bugs/?13216
[2] http://savannah.gnu.org/bugs/?29074
On Thu, 27 Mar 2014 07:39:40 kcrisman wrote:
On Thursday, March 27, 2014 9:30:29 AM UTC-4, kcrisman wrote:
On Thursday, March 27, 2014 1:56:21 AM UTC-4, Ivan Andrus wrote:
I haven’t really followed the discussion, but if the problem is because
of lines in the Makefile like
-include
then you should be able to just add a dummy file on that line so that
it’s
-include dummy.mk …
where the … is of course the expression that will sometimes be empty.
But it’s also not hard to check for the existence of a directory.
-Ivan
I think the problem is that we don't even know that this *is* the problem
(see Leif's reply in the other thread on beta4). Is there any way to
easily check the differences between make 3.80 and 3.81 on this issue? I
assume that people building with 3.81 don't see these warning messages.
I think *if* this is the problem, then Ivan's solution of having this
include line only added if there is a profile/ directory makes sense, as
the syntax probably is fairly standard.
Neither of http://www.cmcrossroads.com/article/whats-new-gnu-make-381 and
https://notendur.hi.is/jonasson/software/make-book/make-news.pdf seem to
have something that immediately shows this problem, though. The full
list of bugs does have the thing where -include with an empty or false list
continues (http://savannah.gnu.org/bugs/index.php?13216) is fixed in 3.81,
would this be enough to cause the problems in question? Unfortunately I
don't see a link to the actual diff. http://savannah.gnu.org/bugs/?29074
and http://savannah.gnu.org/bugs/?27396 seem relevant but were never
actually dealt with, and http://savannah.gnu.org/bugs/?102 is a really old
version of something related...
--
() The ASCII Ribbon Campaign
/\ Help Cure HTML E-Mail
--
You received this message because you are subscribed to the Google Groups
"sage-release" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-release+unsubscr...@googlegroups.com.
To post to this group, send email to sage-release@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-release.
For more options, visit https://groups.google.com/d/optout.
