Hi,
Continuing this thread, I think that building Sage shouldn't require
X11. E.g., on t2, the new R png tests fail:
File "/scratch/wstein/build/sage-4.4.1.alpha2/devel/sage/sage/interfaces/r.py",
line 993:
sage: r.png(filename='"%s"'%filename) # filename not needed in
notebook, used for doctesting
Exception raised:
...
sage: r.png(filename='"%s"'%filename) # filename not needed in
notebook, used for doctesting
File
"/scratch/wstein/build/sage-4.4.1.alpha2/local/lib/python/site-packages/sage/interfaces/r.py
", line 356, in png
raise RuntimeError, "R was not compiled with PNG support"
RuntimeError: R was not compiled with PNG support
*******************************************************************
---
Unfortunately, this really means that those tests should all be changed to be
# optional -- x11
They won't get tested by "make test". However, doing
./sage -t -only_optional=x11 devel/sage/sage/
will test them. The release manager checklist will suggest to do this check.
I've opened a ticket for this:
http://trac.sagemath.org/sage_trac/ticket/8834
I hope maybe Karl or somebody can make those tests optional.
William
On Fri, Apr 30, 2010 at 7:48 PM, kcrisman <kcris...@gmail.com> wrote:
>
>
>>
>> > Maybe we should just *always* build it with X support except on Mac?
>> > But presumably it needs some X library to connect to to do that...?
>>
>> sphg-install in R has numerous problems. Someone updated it which
>> caused major hassles on Solaris, with more fallout on Linux. They
>> obviously did not read the documentation, or notice the warnings
>> generated by unknown options passed to the configure script
>>
>> I'm aware the method that was (is) used to build X is stupid. I fixed
>> this for Solaris, but whoever implemented it was using the wrong
>> approach.
>>
>> I created
>>
>> http://trac.sagemath.org/sage_trac/ticket/8274
>>
>> to highlight some of the issues.
>
> Yes, I think that's appeared elsewhere on the list, and I for one
> appreciate your thoroughness. My fear is that the someone who did
> much of the install script is someone who is no longer working with
> Sage. My own goal with R has been to improve our support of it in the
> ways I can; unfortunately, that doesn't extend to the issues you or
> Dan mention :( I would point out that (other than the iconv issue,
> which ended up being a bonus for Sage thanks to your work) all the
> recent changes have been improvements, not regressions :)
>
> So that said, it would be great for someone who actually knows
> something about configuring and building (other than Dave, who has
> already put in far more work on it than we could reasonably expect)
> could look at that install script and compare it with the R
> documentation. But since we have relatively few active developers who
> use R through Sage on a daily basis (at least, it seems so from the
> lists), it may take a while to find someone. I'm cc:ing John Verzani,
> who has previously interacted with Sage and was quite helpful on the
> latest plotting update, in case he has any ideas of an R devel who is
> knowledgeable about such things and also supportive enough of OSS in
> general to perhaps spend some time looking at it. We can hope!
>
> - kcrisman
>
> --
> To post to this group, send an email to sage-devel@googlegroups.com
> To unsubscribe from this group, send an email to
> sage-devel+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/sage-devel
> URL: http://www.sagemath.org
>
--
William Stein
Professor of Mathematics
University of Washington
http://wstein.org
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
