Do not run "make -q" when upgrading

[jdemeyer]

The script spkg/install contains the lines

# Skip the rest if nothing to do (i.e., to [re]build):
# If "make" doesn't understand the -q option (although we require
# GNU make, which supports it), it should exit with a non-zero status
# which is not a problem.
if ${MAKE:-make} -q -f standard/deps $1; then
    echo "Nothing to (re)build / all up-to-date."
    exit 0
fi

This runs make with the -q option, which (via the makefile spkg/standard/deps) in turn runs sage-spkg. This changed recently, before #12016 make -q didn't run sage-spkg.

When upgrading, the file sage-spkg is not automatically upgraded, so we can have an old sage-spkg which does not understand the -q option which causes build errors (only visible in the log files), for example in spkg/logs/patch-2.5.9.p2.log.

This is a problem introduced by #12016.

Component: build

Keywords: sd35.5

Author: Jeroen Demeyer

Reviewer: John Palmieri, Volker Braun

Merged: sage-4.8.rc0

Issue created by migration from https://trac.sagemath.org/ticket/12248

[jdemeyer]

Author: Jeroen Demeyer

[jdemeyer]

Attachment: 12248_make_q.patch.gz

[jdemeyer]

comment:5

Successfully tested building from scratch and upgrading from sage-4.5.1.

[jhpalmieri]

comment:6

Why did #12016 cause this? Is it because of all of the plus signs in spkg/standard/deps?

[jhpalmieri]

comment:7

Unless I made a mistake, I was able to upgrade from 4.4 and 4.5.3 to Sage 4.8.alpha6 plus #9958 and its prerequisites (since that's what I was looking at earlier). I was also able to upgrade after applying the patch here. The patch looks okay to me, but I'm not sure it should be a blocker.

[jdemeyer]

comment:8

Replying to @jhpalmieri:

Why did #12016 cause this? Is it because of all of the plus signs in spkg/standard/deps?

Yes. Those plus signs mark a make rule as recursive.

Replying to @jhpalmieri:

Unless I made a mistake, I was able to upgrade from 4.4 and 4.5.3 to Sage 4.8.alpha6 plus #9958 and its prerequisites (since that's what I was looking at earlier). I was also able to upgrade after applying the patch here.

Perhaps I should have been more clear. The upgrade might work without this patch, but you will see a failed build attempt for example in spkg/logs/patch-2.5.9.p2.log.

[jdemeyer]

comment:9

Replying to @jhpalmieri:

The patch looks okay to me

Meaning "positive review"?

I'm not sure it should be a blocker.

I disagree since it causes non-fatal build errors. Seeing a scary message "An error occured... please contact sage-devel..." isn't good. Besides, this might possibly lead to an actual build failure.

[jdemeyer]

Reviewer: John Palmieri

[vbraun]

comment:10

Looks good to me!

[vbraun]

Changed reviewer from John Palmieri to John Palmieri, Volker Braun

[vbraun]

Changed keywords from none to sd35.5

[jdemeyer]

Merged: sage-4.8.rc0