Comment on trac#17234 (unpostable via Trac.sagemath.org)

[Emmanuel Charpentier]

This was intended to be a comment on Trac#17234, but I was unable to post it on theright place (see this thread).

-------------------------------------------------------------------------------------------

Wups : failure when trying to merge this branch and the current (6.6beta2) head of develop :

I checked out develop, created "mabranche" and merged in
t/17234/rich_output_and_the_ipython_notebook. The merge went cleanly, but the build failed on a file permission error. Last lines of the install.log :

running install_egg_info
Removing /usr/local/sage-6.6/local/lib/python2.7/site-packages/sage-6.6.beta1-py
2.7.egg-info
Writing /usr/local/sage-6.6/local/lib/python2.7/site-packages/sage-6.6.beta1-py2
.7.egg-info

real    26m25.799s
user    25m54.744s
sys     0m23.312s
cp /usr/local/sage-6.6/src/ext/doctest/invalid/syntax_error.tachyon /usr/local/s
age-6.6/local/share/sage/ext/doctest/invalid/syntax_error.tachyon
cp /usr/local/sage-6.6/src/ext/doctest/rich_output/example_wavefront/scene.mtl /usr/local/sage-6.6/local/share/sage/ext/doctest/rich_output/example_wavefront/scene.mtl
cp: impossible de créer le fichier standard « /usr/local/sage-6.6/local/share/sage/ext/doctest/rich_output/example_wavefront/scene.mtl »: Permission non accordée
Makefile:412: recipe for target '/usr/local/sage-6.6/local/share/sage/ext/doctest/rich_output/example_wavefront/scene.mtl' failed
make[2]: *** [/usr/local/sage-6.6/local/share/sage/ext/doctest/rich_output/example_wavefront/scene.mtl] Error 1
make[2]: Leaving directory '/usr/local/sage-6.6/build'
Makefile:437: recipe for target 'all' failed
make[1]: *** [all] Error 2
make[1]: Leaving directory '/usr/local/sage-6.6/build'

real    49m26.767s
user    47m18.212s
sys     1m5.196s
***************************************************************
Error building Sage.

The following package(s) may have failed to build:

The build directory may contain configuration files and other potentially
helpful information. WARNING: if you now run 'make' again, the build
directory will, by default, be deleted. Set the environment variable
SAGE_KEEP_BUILT_SPKGS to 'yes' to prevent this.


I can send you the full log if you need it (just tell me where), but it's too large (11 Mb and small change) to be attached.

Should I mark this branch "needs work" ?

-------------------------------------------------------------------------------------------

HTH,

--
Emmanuel Charpentier

[Volker Braun]

This is supposed to be fxed by ​a173e92. If you tried the ticket before that commit then you need to manually delete / fix permissions in SAGE_LOCAL/share/sage/ext/doctest

[Emmanuel Charpentier]

Funnier and funnier : now, on a couple of another machines, when I 'git trac checkout 17234', git dutifully creates the branch, but says that the upstream branch disapearred...

I tried to look at this via the trac.sagemath.org interface : here, the ticket is resolved/closed : do closing a ticked make its branch disappear ?

I didn't try to add anything via this interface.

Even funnier : I restarted from develop (at 6.6beta2)), recteared a new branch, checked it out and merged my current version of t/17234/rich_output_and_the_ipython_notebook. This version has :

commit a173e92ae42f3fb9fda3c70963ff29dcdffca7e6
Author: Volker Braun <vbrau...@gmail.com>
Date:   Sun Mar 1 01:18:34 2015 +0100

as its last commit (i. e. first one to be printed by 'git log'). Yes, the log contains a173e92...

With this version, I still have problems with permissions with $SAGE_ROOT/local/share/sage/ext/doctest/rich_text_output: both this directory and its subdirectory example_wavefront are setgid. resetting these permissions to 755 seems to fix this particlar problem, but others appear (I made the mistake to ask for a parallel build, thinking that it was fixed by 6.6beta2).

Repeated uses of make seem to be inefficient... I'll try to 'make distclean && make' serially and will keep you posted.

[Volker Braun]

When closing a ticket, a script replaces the branch on the trac ticket (which might change in the future) with its current head sha1 (i.e. always one fixed commit).

[Emmanuel Charpentier]

Thank you. This explains that.

FYI : I had to re-build from scratch. I don't understand why... yet !

Thanks a lot !

--
Emmanuel Charpentier