Resolve symbolic links in $HOME/.sage

[jdemeyer]

Similarly to SAGE_ROOT, it might be good to resolve symbolic links in $HOME/.sage when assigning DOT_SAGE. This can be useful for example when $HOME/.sage is a symbolic link to a non-existent directory (the directory would be created later in sage-sage).

Probably we should use whatever solution comes from #5852.

Depends on #11924

CC: @nexttime

Component: scripts

Author: Jeroen Demeyer

Reviewer: John Palmieri

Merged: sage-4.8.alpha6

Issue created by migration from https://trac.sagemath.org/ticket/11704

[jdemeyer]

Dependencies: #5852

[jdemeyer]

Changed dependencies from #5852 to none

[jhpalmieri]

comment:9

Note also #11924: currently if DOTSAGE doesn't end with a slash, then there are unintended consequences. (#11924 should be an easy review.)

[jdemeyer]

Dependencies: #11924

[jhpalmieri]

comment:11

I wonder if we should put the resolvelinks script in a separate file in the root repo, and then call it from the top-level sage script (as in #5852) or from sage-env (for this ticket). As such, this seems related to #11073.

[jdemeyer]

comment:12

Replying to @jhpalmieri:

I wonder if we should put the resolvelinks script in a separate file in the root repo, and then call it from the top-level sage script (as in #5852) or from sage-env (for this ticket).

We cannot put it in a seperate file because $SAGE_ROOT/sage would need to know SAGE_ROOT in order to find this new separate file. So, in any case, $SAGE_ROOT/sage needs the resolvelinks script. So, we would have to source $SAGE_ROOT/sage from sage-env which doesn't sound like a very good idea (it could be done though).

[jhpalmieri]

comment:13

Well, what's the purpose of canonicalizing SAGE_ROOT in the sage script? It seems useful to have canonical names for paths, and it's important for some doctests (etc.), but do we actually need to do this for the script to function properly right at the start? If not, we could detect a preliminary value for SAGE_ROOT first, find the resolvelinks script, run it if it exists (e.g., check after seeing if sage-sage exists and before exporting SAGE_ROOT), and then find and export the permanent value for SAGE_ROOT.

[jdemeyer]

comment:14

Replying to @jhpalmieri:

Well, what's the purpose of canonicalizing SAGE_ROOT in the sage script?

It's precisely for the reason I said: we need to be able to determine a usable value of $SAGE_ROOT before we can do anything with Sage.

It's true that it doesn't matter that it's canonical, but we still need to dereference symbolic links in $0 (the script filename) to find $SAGE_ROOT. Since #5852, we support an installation where somebody makes a symbolic link from /usr/local/bin/sage to /usr/local/sage/sage-4.8/sage or whatever.

I agree the code duplication is not so nice, but I still think the patch can be reviewed as-is.

[jhpalmieri]

Reviewer: John Palmieri

[jdemeyer]

comment:16

Amazing, this causes

**********************************************************************
File "/mnt/usb1/scratch/jdemeyer/merger/sage-4.8.alpha4-remove-base/devel/sage-main/sage/schemes/elliptic_curves/ell_rational_field.py", line 526:
    sage: E.conductor(algorithm="gp")
Exception raised:
    Traceback (most recent call last):
      File "/mnt/usb1/scratch/jdemeyer/merger/sage-4.8.alpha4-remove-base/local/bin/ncadoctest.py", line 1231, in run_one_test
        self.run_one_example(test, example, filename, compileflags)
      File "/mnt/usb1/scratch/jdemeyer/merger/sage-4.8.alpha4-remove-base/local/bin/sagedoctest.py", line 38, in run_one_example
        OrigDocTestRunner.run_one_example(self, test, example, filename, compileflags)
      File "/mnt/usb1/scratch/jdemeyer/merger/sage-4.8.alpha4-remove-base/local/bin/ncadoctest.py", line 1172, in run_one_example
        compileflags, 1) in test.globs
      File "<doctest __main__.example_12[5]>", line 1, in <module>
        E.conductor(algorithm="gp")###line 526:
    sage: E.conductor(algorithm="gp")
      File "/mnt/usb1/scratch/jdemeyer/merger/sage-4.8.alpha4-remove-base/local/lib/python/site-packages/sage/schemes/elliptic_curves/ell_rational_field.py", line 552, in conductor
        self.__conductor_gp = Integer(gp.eval('ellglobalred(ellinit(%s,0))[1]'%list(self.a_invariants())))
      File "integer.pyx", line 681, in sage.rings.integer.Integer.__init__ (sage/rings/integer.c:6786)
    TypeError: unable to convert x (=read("/mnt/usb1/scratch/jdemeyer/merger/sage-4.8.alpha4-remove-base/home/.sage/temp/sage.math.washington.edu/9425//interface//tmp11939") ) to an integer
**********************************************************************

[jdemeyer]

comment:17

I keep positive review here, the latter problem is now #12221.

[jdemeyer]

Changed dependencies from #11924 to #11924, #12221

[jdemeyer]

Attachment: 11704.patch.gz

Patch to the SCRIPTS repository (local/bin)

[jdemeyer]

Changed dependencies from #11924, #12221 to #11924

[jdemeyer]

comment:18

I changed the patch slightly to add the slash after $DOT_SAGE. I don't see how this could cause #12221, but it seems safer to keep the slash, just like it was before. Needs review please.

[jdemeyer]

comment:20

John, can you please review this new revised patch?

[jhpalmieri]

comment:21

The only real change is to append the slash, right? It works for me in my testing.

[jdemeyer]

Merged: sage-4.8.alpha6