#17044: fix pexpect interfaces with a system-wide sage install
-------------------------+-------------------------------------------------
Reporter: was | Owner:
Type: | Status: needs_review
defect | Milestone: sage-6.4
Priority: major | Resolution:
Component: | Merged in:
interfaces | Reviewers:
Keywords: | Work issues:
Authors: | Commit:
Frédéric Chapoton | 33e6ec3d8d907d3cc04cd90a60db29d414c64a6b
Report Upstream: N/A | Stopgaps:
Branch: |
u/chapoton/17044 |
Dependencies: |
-------------------------+-------------------------------------------------
Comment (by was):
Replying to [comment:2 jdemeyer]:
> I would like to understand the problem better. So William, could you
please
> 1) say which version of Sage this refers to
At least any Sage-6.x.
> 2) describe the actual input and error you are getting
Install Sage system-wide, then try to do anything with *some* of the
pexpect-based interfaces as a normal user who doesn't own the sage
install, especially ones that use optional software. E.g.,
sage: scilab('2+3')
> 3) if any doctests fail in your setup (without this patch)
Nope. We don't do doctesting that tests permissions, since doctesting is
always run as the user that owns the Sage install.
>
> I am asking because you're certainly not the first person ever to
install Sage system-wide but it is the first time I see this problem
reported.
This problem pops up somewhat regularly on the mailing lists, and
basically gets ignored, probably because for the most part our Sage
developers usually own their own Sage installs. However, with
SageMathCloud, which has now 7000 *active users* per week all using a
system-wide Sage install, this problem comes up more. I had dome some
horrible hacks to try to get around it, e.g., pointing
SAGE_ROOT/local/share at something world writeable, but even this fails
miserable if *two* regular users use the same interface -- one creates the
user directory as their own, and the other then can't write to it.
I think much of this boils down to a stupid coding mistake (surely made by
me) involving `=""` versus `=None`.
> So the problem is probably more subtle than you think.
It probably is. I wouldn't be surprised at all if this small patch is not
the whole story, and there are similar bugs still lurking, since any
classes that derive from the base Expect class (in expect.py) could cause
their own related trouble...
--
Ticket URL: <http://trac.sagemath.org/ticket/17044#comment:3>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica,
and MATLAB
--
You received this message because you are subscribed to the Google Groups
"sage-trac" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sage-trac.
For more options, visit https://groups.google.com/d/optout.
