I don't what was actually taking the time: the waiting was for
"reading source" and "writing output".
Luckily my edits worked first time so I did not have to do the cycle
again; but and work on docstrings will be rather tedious unless
someone can work out why this is happening.
John
I think the source may be #7683
http://trac.sagemath.org/sage_trac/ticket/7683
which adds a config value autodoc_builtin_argspec to Sphinx. This lets
us set
autodoc_builtin_argspec = sage.misc.sageinspect.sage_getargspec
to detect the arguments of methods and functions defined in Cython files.
Sphinx compares the current set of config values to a pickled set to
decide whether to reread all documents on the grounds
[config changed]
But autodoc_builtin_argspec is a function, which, I think, Sphinx
pickles as None. Moreover, sage_getargspec gets a different identity
id(sage_getargspec)
[nearly] each time Sphinx is run. This mismatch may trigger the full
rebuild.
A possible workaround: In
SAGE_LOCAL/lib/python2.6/site-packages/Sphinx-0.6.3-py2.6.egg/sphinx/environment.py
import inspect and insert, e.g.,
if inspect.isfunction(config[key]):
continue
around line 474.
John
2009/12/28 Pat LeSmithe <qed...@gmail.com>:
> --
> To post to this group, send an email to sage-...@googlegroups.com
> To unsubscribe from this group, send an email to sage-devel+...@googlegroups.com
> For more options, visit this group at http://groups.google.com/group/sage-devel
> URL: http://www.sagemath.org
>
Without making a clone, anytime I edit, then run sage -b, a resulting
build of the HTML reference manual appears to begin all over again.
As John mentioned, this really discourages one from getting the
documentation fine-tuned on a contribution.
Thanks in advance to anyone who can get this back to its previous
behavior!
Rob
On Tue, Dec 29, 2009 at 7:46 AM, John Cremona <john.c...@gmail.com> wrote:
> I really hope we can fix this. It takes 10 minutes to make a clone on
> my laptop now!
See ticket #7796 [1] for an updated Sphinx spkg. The fix is
essentially what Pat suggested.
[1] http://trac.sagemath.org/sage_trac/ticket/7796
--
Regards
Minh Van Nguyen