Installation failure on openSuSE 11.3
Florent Hivert
I tried to install from scratch the new 4.6.rc_0 release and I'm running into
trouble:
- first of all there is the well know readline problem (see #9530 and #9523)
so I replaced readline-6.0.p2.spkg.patch by the one on #9530. It seem that
this perfectly worked for me and I was ready to put a positive review on
#9530. However,
- I got a very similar problem than Minh on #10104:
It seems that except for the doc, sage compiled properly. My install.log says:
[...]
real 299m31.649s
user 281m3.546s
sys 7m46.051s
To install gap, gp, singular, etc., scripts
in a standard bin directory, start sage and
type something like
sage: install_scripts('/usr/local/bin')
at the Sage command prompt.
To build the documentation, run
make doc
Sage build/upgrade complete!
however, sage refuse to start:
popcorn-*age/sage-4.6.rc0 $ ./sage
----------------------------------------------------------------------
| Sage Version 4.6.rc0, Release Date: 2010-10-21 |
| Type notebook() for the GUI, and license() for information. |
----------------------------------------------------------------------
**********************************************************************
* *
* Warning: this is a prerelease version, and it may be unstable. *
* *
**********************************************************************
---------------------------------------------------------------------------
ImportError Traceback (most recent call last)
/home/data/Sage-Install/sage-4.6.rc0/local/bin/<string> in <module>()
/home/florent/src/Sage/sage-4.6.rc0/local/lib/python2.6/site-packages/sage/misc/preparser_ipython.py in <module>()
6 ###########################################################################
7
----> 8 import sage.misc.interpreter
9
10 import preparser
/home/florent/src/Sage/sage-4.6.rc0/local/lib/python2.6/site-packages/sage/misc/interpreter.py in <module>()
100
101 import os
--> 102 import log
103
104 import remote_file
/home/florent/src/Sage/sage-4.6.rc0/local/lib/python2.6/site-packages/sage/misc/log.py in <module>()
63
64 import interpreter
---> 65 import latex
66 import misc
67
/home/florent/src/Sage/sage-4.6.rc0/local/lib/python2.6/site-packages/sage/misc/latex.py in <module>()
38 import random
39
---> 40 from misc import tmp_dir, graphics_filename
41 import sage_eval
42 from sage.misc.misc import SAGE_DOC
/home/florent/src/Sage/sage-4.6.rc0/local/lib/python2.6/site-packages/sage/misc/misc.py in <module>()
1930 return "0"*(size-len(str(s))) + str(s)
1931
-> 1932 import sage.server.support
1933
1934 def embedded():
/home/florent/src/Sage/sage-4.6.rc0/local/lib/python2.6/site-packages/sage/server/support.py in <module>()
17 import sage.misc.pager
18
---> 19 import sage.misc.sagedoc as sagedoc
20 import sage.misc.sageinspect as sageinspect
21
/home/florent/src/Sage/sage-4.6.rc0/local/lib/python2.6/site-packages/sage/misc/sagedoc.py in <module>()
33 from sage.misc.viewer import browser
34 from sage.misc.misc import SAGE_DOC, tmp_dir
---> 35 from sagenb.misc.sphinxify import sphinxify
36 import sage.version
37
ImportError: No module named sagenb.misc.sphinxify
WARNING: Failure executing code: 'import sage.misc.preparser_ipython; sage.misc.preparser_ipython.magma_colon_equals=True'
---------------------------------------------------------------------------
ImportError Traceback (most recent call last)
/home/florent/src/Sage/sage-4.6.rc0/local/lib/python2.6/site-packages/IPython/ipmaker.pyc in force_import(modname)
64 reload(sys.modules[modname])
65 else:
---> 66 __import__(modname)
67
68
/home/data/Sage-Install/sage-4.6.rc0/local/bin/ipy_profile_sage.py in <module>()
1 import os
2 if 'SAGE_CLEAN' not in os.environ:
----> 3 import sage.misc.misc
4 from sage.misc.interpreter import preparser, _ip
5 preparser(True)
/home/florent/src/Sage/sage-4.6.rc0/local/lib/python2.6/site-packages/sage/misc/misc.py in <module>()
1930 return "0"*(size-len(str(s))) + str(s)
1931
-> 1932 import sage.server.support
1933
1934 def embedded():
/home/florent/src/Sage/sage-4.6.rc0/local/lib/python2.6/site-packages/sage/server/support.py in <module>()
17 import sage.misc.pager
18
---> 19 import sage.misc.sagedoc as sagedoc
20 import sage.misc.sageinspect as sageinspect
21
/home/florent/src/Sage/sage-4.6.rc0/local/lib/python2.6/site-packages/sage/misc/sagedoc.py in <module>()
33 from sage.misc.viewer import browser
34 from sage.misc.misc import SAGE_DOC, tmp_dir
---> 35 from sagenb.misc.sphinxify import sphinxify
36 import sage.version
37
ImportError: No module named sagenb.misc.sphinxify
Error importing ipy_profile_sage - perhaps you should run %upgrade?
WARNING: Loading of ipy_profile_sage failed.
<ERROR: name 'sage_prompt' is not defined>quit
Type quit() to exit.
<ERROR: name 'sage_prompt' is not defined>quit()
I tried:
popcorn-*.6/site-packages $ ../../../../sage -ipython
Python 2.6.4 (r264:75706, Oct 23 2010, 15:28:40)
Type "copyright", "credits" or "license" for more information.
IPython 0.9.1 -- An enhanced Interactive Python.
? -> Introduction and overview of IPython's features.
%quickref -> Quick reference.
help -> Python's own help system.
object? -> Details about 'object'. ?object also works, ?? prints more.
In [1]: import docutils
In [2]: import sympy
In [3]: import sphinx
In [4]: import sagenb
---------------------------------------------------------------------------
ImportError Traceback (most recent call last)
/home/data/Sage-Install/sage-4.6.rc0/local/lib/python2.6/site-packages/<ipython console> in <module>()
ImportError: No module named sagenb
Any idea about where to look at ?
Cheers,
Florent
Mitesh Patel
> - I got a very similar problem than Minh on #10104:
> It seems that except for the doc, sage compiled properly. My install.log says:
>
> [...]
> real 299m31.649s
> user 281m3.546s
> sys 7m46.051s
> To install gap, gp, singular, etc., scripts
> in a standard bin directory, start sage and
> type something like
> sage: install_scripts('/usr/local/bin')
> at the Sage command prompt.
>
> To build the documentation, run
> make doc
>
> Sage build/upgrade complete!
>
>
> however, sage refuse to start:
Just to check the build, what is the output of
grep "An error occurred" install.log
?
Florent Hivert
> Just to check the build, what is the output of
>
> grep "An error occurred" install.log
No problem:
popcorn-*age/sage-4.6.rc0 $ grep "An error occurred" install.log
popcorn-*age/sage-4.6.rc0 $
Cheers,
Florent
Mitesh Patel
>
> popcorn-*age/sage-4.6.rc0 $ ./sage
> ImportError: No module named sagenb.misc.sphinxify
>
Do you see the following structure in devel/
$ cd SAGE_ROOT/devel
$ ls -lF
total 12
drwxr-xr-x 2 mpatel mpatel 4096 2010-10-21 15:59 old/
lrwxrwxrwx 1 mpatel mpatel 9 2010-10-21 17:00 sage -> sage-main/
drwxr-xr-x 7 mpatel mpatel 4096 2010-10-22 00:59 sage-main/
lrwxrwxrwx 1 mpatel mpatel 11 2010-10-21 16:57 sagenb -> sagenb-main/
drwxr-xr-x 8 mpatel mpatel 4096 2010-10-21 15:33 sagenb-main/
?
Also, what happens if you reinstall the sagenb package
$ cd SAGE_ROOT
$ ./sage -f spkg/standard/sagenb-0.8.7.spkg
and try starting Sage again?
leif
Did someone (successfully) build rc0 on Skynet's menas?
I just noticed it's an openSUSE 11.1 x86_64 Core2 Quad Q6600 box.
-Leif
P.S.: #9523 and #9530 (fixing the readline problems on OpenSuSE and Arch
Linux) currently need review...
http://trac.sagemath.org/sage_trac/ticket/9523 (new readline 6.1 spkg)
http://trac.sagemath.org/sage_trac/ticket/9530 (simple patch to 6.0
spkg; currently just a patch up there, no new spkg yet available)
Florent Hivert
> On 10/24/2010 03:46 AM, Florent Hivert wrote:
> > however, sage refuse to start:
> >
> > popcorn-*age/sage-4.6.rc0 $ ./sage
> > [...]
> > ImportError: No module named sagenb.misc.sphinxify
> >
> > Any idea about where to look at ?
>
> Do you see the following structure in devel/
>
> $ cd SAGE_ROOT/devel
> $ ls -lF
> total 12
> drwxr-xr-x 2 mpatel mpatel 4096 2010-10-21 15:59 old/
> lrwxrwxrwx 1 mpatel mpatel 9 2010-10-21 17:00 sage -> sage-main/
> drwxr-xr-x 7 mpatel mpatel 4096 2010-10-22 00:59 sage-main/
> lrwxrwxrwx 1 mpatel mpatel 11 2010-10-21 16:57 sagenb -> sagenb-main/
> drwxr-xr-x 8 mpatel mpatel 4096 2010-10-21 15:33 sagenb-main/
Yep:
popcorn-*ge-4.6.rc0/devel $ ls -lF
total 12
drwxr-xr-x 2 florent users 4096 oct. 23 19:43 old/
lrwxrwxrwx 1 florent users 9 oct. 23 19:43 sage -> sage-main/
drwxr-xr-x 7 florent users 4096 oct. 23 19:43 sage-main/
lrwxrwxrwx 1 florent users 11 oct. 23 19:29 sagenb -> sagenb-main/
drwxr-xr-x 8 florent users 4096 oct. 23 19:29 sagenb-main/
popcorn-*ge-4.6.rc0/devel $
> Also, what happens if you reinstall the sagenb package
>
> $ cd SAGE_ROOT
> $ ./sage -f spkg/standard/sagenb-0.8.7.spkg
>
> and try starting Sage again?
Nothing different...
I've uploaded my install.log if it helps:
http://igm.univ-mlv.fr/~hivert/install.log
Florent
Dr. David Kirkby
>>
>
> Did someone (successfully) build rc0 on Skynet's menas?
>
> I just noticed it's an openSUSE 11.1 x86_64 Core2 Quad Q6600 box.
>
>
> -Leif
Yes, it did build ok.
Florent Hivert
popcorn-*age/sage-4.6.rc0 $ ./sage -ipython
Python 2.6.4 (r264:75706, Oct 23 2010, 15:28:40)
Type "copyright", "credits" or "license" for more information.
IPython 0.9.1 -- An enhanced Interactive Python.
? -> Introduction and overview of IPython's features.
%quickref -> Quick reference.
help -> Python's own help system.
object? -> Details about 'object'. ?object also works, ?? prints more.
In [1]: import sys
In [2]: for i in sys.path: print i
...:
/home/data/Sage-Install/sage-4.6.rc0/local/bin
[...]
/home/florent/src/Sage/sage-4.6.rc0/local/lib/python2.6/site-packages/SQLAlchemy-0.5.8-py2.6.egg
/home/data/Sage-Install/sage-4.6.rc0
[...]
/home/florent/src/Sage/sage-4.6.rc0/local/lib/python2.6/site-packages/IPython/Extensions
/home/florent/.ipython
Whereas on sage-4.5.3:
/home/data/Sage-Install/sage-4.5.3/local/bin
[...]/home/florent/src/Sage/sage/local/lib/python2.6/site-packages/SQLAlchemy-0.5.8-py2.6.egg
/home/florent/src/Sage/sage/local/lib/python2.6/site-packages/sagenb-0.8.2-py2.6.egg
/home/data/Sage-Install/sage-4.5.3/local/bin
/home/florent/src/Sage/sage/local/lib/python
[...]
/home/florent/.sage/ipython
So the problem comes from that
SAGE_ROOT/local/lib/python2.6/site-packages/sagenb-0.8.2-py2.6.egg
is missing from sys.path
Any idea about the reason why this could happen ?
Cheers,
Florent
Mitesh Patel
This probably relates to #10097
http://trac.sagemath.org/sage_trac/ticket/10097
I see the following
$ cd SAGE_ROOT/local/lib/python/site-packages
$ cat sagenb.egg-link
../../../../devel/sagenb
$ cat easy-install.pth
import sys; sys.__plen = len(sys.path)
./setuptools-0.6c9-py2.6.egg
./Twisted-9.0.0-py2.6-linux-x86_64.egg
./zope.interface-3.6.1-py2.6-linux-x86_64.egg
./zope.proxy-3.6.0-py2.6-linux-x86_64.egg
./zope.exceptions-3.6.1-py2.6.egg
./zope.testing-3.9.5-py2.6.egg
./ZConfig-2.8.0-py2.6.egg
./zdaemon-2.0.4-py2.6.egg
./ZODB3-3.7.0-py2.6-linux-x86_64.egg
./Pygments-0.11.1-py2.6.egg
./Jinja2-2.1.1-py2.6-linux-x86_64.egg
./Sphinx-0.6.3-py2.6.egg
./SQLAlchemy-0.5.8-py2.6.egg
../../../../devel/sagenb
import sys; new=sys.path[sys.__plen:]; del sys.path[sys.__plen:];
p=getattr(sys,'__egginsert',0); sys.path[p:p]=new; sys.__egginsert =
p+len(new)
$
Perhaps the relative path to sagenb is somehow missing in one of these
files in your installation? If it is missing, what happens if you add it?
Florent Hivert
> $ cat easy-install.pth
> import sys; sys.__plen = len(sys.path)
> ./setuptools-0.6c9-py2.6.egg
> ./Twisted-9.0.0-py2.6-linux-x86_64.egg
> ./zope.interface-3.6.1-py2.6-linux-x86_64.egg
> ./zope.proxy-3.6.0-py2.6-linux-x86_64.egg
> ./zope.exceptions-3.6.1-py2.6.egg
> ./zope.testing-3.9.5-py2.6.egg
> ./ZConfig-2.8.0-py2.6.egg
> ./zdaemon-2.0.4-py2.6.egg
> ./ZODB3-3.7.0-py2.6-linux-x86_64.egg
> ./Pygments-0.11.1-py2.6.egg
> ./Jinja2-2.1.1-py2.6-linux-x86_64.egg
> ./Sphinx-0.6.3-py2.6.egg
> ./SQLAlchemy-0.5.8-py2.6.egg
> ../../../../devel/sagenb
> import sys; new=sys.path[sys.__plen:]; del sys.path[sys.__plen:];
> p=getattr(sys,'__egginsert',0); sys.path[p:p]=new; sys.__egginsert =
> p+len(new)
> $
>
> Perhaps the relative path to sagenb is somehow missing in one of these
> files in your installation? If it is missing, what happens if you add it?
I added "../../../../devel/sagenb" to my
sage-4.6.rc0/local/lib/python/site-packages/easy-install.pth
and everything seem to be ok now:
popcorn-*ine/sage-4.6.rc0 $ ./sage
----------------------------------------------------------------------
| Sage Version 4.6.rc0, Release Date: 2010-10-21 |
| Type notebook() for the GUI, and license() for information. |
----------------------------------------------------------------------
**********************************************************************
* *
* Warning: this is a prerelease version, and it may be unstable. *
* *
**********************************************************************
sage: Partitions(3).list()
[[3], [2, 1], [1, 1, 1]]
I'm not sure what to do now ? Should I open a ticket ? Is there any problem
with #10097 ? This is completely deterministic: I got the same failure 4
times (depending how I workaround the readline problem)...
Cheers,
Florent
Mitesh Patel
Does anyone see a hole in the patch at
http://trac.sagemath.org/sage_trac/ticket/10097
that would lead to this problem? Maybe we should add
"../../../../devel/sagenb" outright, if the sed command yields an
unchanged easy-install.pth.new?
By the way, how do the long doctests fare?
leif
> Does anyone see a hole in the patch at
>
> http://trac.sagemath.org/sage_trac/ticket/10097
>
> that would lead to this problem?
Though there is no error checking performed, and escaping all the slashs
looks funny (you can use almost any character, e.g. 's|/foo/bar|/baz|'),
I can't see errors in the patch.
> Maybe we should add
> "../../../../devel/sagenb" outright, if the sed command yields an
> unchanged easy-install.pth.new?
>
> By the way, how do the long doctests fare?
Ok on Ubuntu, even in parallel via NFS ;-)
-Leif
leif
> Mitesh Patel wrote:
>> Does anyone see a hole in the patch at
>>
>> http://trac.sagemath.org/sage_trac/ticket/10097
>>
>> that would lead to this problem?
>
> Though there is no error checking performed, and escaping all the slashs
> looks funny (you can use almost any character, e.g. 's|/foo/bar|/baz|'),
> I can't see errors in the patch.
It *might* once again be related to a missing "#!/usr/bin/env bash" in
the created spkg-install. (Also, I think in general it should be safer
to use a raw string: spkg_install_fd.write(r""" ... """) )
-Leif
leif
P.S.: (in spkg-dist)
rm -f "$SAGE_ROOT/devel/sagenb"
removes the symbolic link, but doesn't delete the contents of its target
(which is intended I think), i.e. it should be:
rm -rf "$SAGE_ROOT/devel/sagenb"
-Leif
Mitesh Patel
http://trac.sagemath.org/sage_trac/ticket/10176
for the SageNB problem.
See
http://trac.sagemath.org/sage_trac/ticket/9530
for the Readline problem.
Apart from finishing the review of #9896
http://trac.sagemath.org/sage_trac/ticket/9896
these are the only potential obstacles I see to releasing 4.6.
Robert Bradshaw
But if it's going to hold stuff up, I'd rather not delay 4.6 for a
system we don't yet support. (That can be the first thing in 4.6.1 if
they get reviewed...)
- Robert
leif
> On Wed, Oct 27, 2010 at 2:04 AM, Mitesh Patel <qed...@gmail.com> wrote:
>> I've opened
>>
>> http://trac.sagemath.org/sage_trac/ticket/10176
>>
>> for the SageNB problem.
>>
>> See
>>
>> http://trac.sagemath.org/sage_trac/ticket/9530
>>
>> for the Readline problem.
>>
>> Apart from finishing the review of #9896
>>
>> http://trac.sagemath.org/sage_trac/ticket/9896
>>
>> these are the only potential obstacles I see to releasing 4.6.
>
> But if it's going to hold stuff up, I'd rather not delay 4.6 for a
> system we don't yet support.
Well, "of course"^TM OpenSuSE is at least listed in README.txt as an
officially supported platform, and since #9530 doesn't affect any other
system but Arch and OpenSuSE I see no reason for not merging it, but
hesitate to give it a "dry" positive review since I have no access to
any of the above systems.
It would IMHO be a pity to break Sage 4.6 on the two recent releases
(11.2 and 11.3) just because the spkg-install applies the relevant
(unmodified) patch only on 11.1. (In fact, it doesn't patch Sage's
readline but copy / make Sage use the system's one.)
> (That can be the first thing in 4.6.1 if
> they get reviewed...)
In Sage 4.6.1 we'll most probably have an upgraded readline spkg (6.1,
#9523) which also solves the problem (at least temporarily, or will also
include the patch from #9523). But I think Sage 4.6.1 won't be released
until 4 or 6 weeks...
We're currently working on #10176, which seems to be specific to the
setup, or just by luck did not happen on the Skynet machine running the
same operating system.
Hopefully someone will finally give #9896 positive review; John Palmieri
already positively reviewed all but one of the patches there.
John Cremona took a look at the last; perhaps you could also comment on
it... ;-) (A change to module_list.py; coincidentally I just mentioned
your plans regarding Cython pragmas and module_list there.)
Regards,
-Leif
Robert Bradshaw
> Robert Bradshaw wrote:
>> On Wed, Oct 27, 2010 at 2:04 AM, Mitesh Patel <qed...@gmail.com> wrote:
>>> I've opened
>>>
>>> http://trac.sagemath.org/sage_trac/ticket/10176
>>>
>>> for the SageNB problem.
>>>
>>> See
>>>
>>> http://trac.sagemath.org/sage_trac/ticket/9530
>>>
>>> for the Readline problem.
>>>
>>> Apart from finishing the review of #9896
>>>
>>> http://trac.sagemath.org/sage_trac/ticket/9896
>>>
>>> these are the only potential obstacles I see to releasing 4.6.
>>
>> But if it's going to hold stuff up, I'd rather not delay 4.6 for a
>> system we don't yet support.
>
> Well, "of course"^TM OpenSuSE is at least listed in README.txt as an
> officially supported platform, and since #9530 doesn't affect any other
> system but Arch and OpenSuSE I see no reason for not merging it, but
> hesitate to give it a "dry" positive review since I have no access to
> any of the above systems.
>
> It would IMHO be a pity to break Sage 4.6 on the two recent releases
> (11.2 and 11.3) just because the spkg-install applies the relevant
> (unmodified) patch only on 11.1. (In fact, it doesn't patch Sage's
> readline but copy / make Sage use the system's one.
If it's a simple issue that can get reviewed quickly, then I'm all for
it. I was just suggesting that if it's something that drags on, I'd
rather not hold the release up just for that.
>> (That can be the first thing in 4.6.1 if
>> they get reviewed...)
>
> In Sage 4.6.1 we'll most probably have an upgraded readline spkg (6.1,
> #9523) which also solves the problem (at least temporarily, or will also
> include the patch from #9523). But I think Sage 4.6.1 won't be released
> until 4 or 6 weeks...
>
>
> We're currently working on #10176, which seems to be specific to the
> setup, or just by luck did not happen on the Skynet machine running the
> same operating system.
>
>
> Hopefully someone will finally give #9896 positive review; John Palmieri
> already positively reviewed all but one of the patches there.
>
> John Cremona took a look at the last; perhaps you could also comment on
> it... ;-) (A change to module_list.py; coincidentally I just mentioned
> your plans regarding Cython pragmas and module_list there.)
>
>
> Regards,
> -Leif
>
> --
> You received this message because you are subscribed to the Google Groups "sage-release" group.
> To post to this group, send email to sage-r...@googlegroups.com.
> To unsubscribe from this group, send email to sage-release...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/sage-release?hl=en.
>
>
Dr. David Kirkby
>> But if it's going to hold stuff up, I'd rather not delay 4.6 for a
>> system we don't yet support.
>
> Well, "of course"^TM OpenSuSE is at least listed in README.txt as an
> officially supported platform, and since #9530 doesn't affect any other
> system but Arch and OpenSuSE I see no reason for not merging it, but
> hesitate to give it a "dry" positive review since I have no access to
> any of the above systems.
> It would IMHO be a pity to break Sage 4.6 on the two recent releases
> (11.2 and 11.3) just because the spkg-install applies the relevant
> (unmodified) patch only on 11.1. (In fact, it doesn't patch Sage's
> readline but copy / make Sage use the system's one.)
Well, openSUSE 11.2 has been out a year. We have overlooked this for a year.
It's hard to see why it should now be a blocker.
As I've noted before, README.txt is next to useless when it comes to supported
platforms.
http://wiki.sagemath.org/SupportedPlatforms
Should be better.
Also note that the only openSUSE machine we have, which is 11.1 (I think), and
runs on Itanium, not x86.
http://build.sagemath.org/sage/buildslaves/iras-1
has failed to doctest properly in the last 6 attempts at building it. It's not
clear to me that we are very successful with even openSUSE 11.1.
I'm just in the process of downloading 11.3 and will attempt to make a virtual
machine on my OpenSoalaris x86 box. If Sage builds on 11.3 and 11.2 (which I'll
download once 11.3 has completed downloading), then I'll be in a better position
to judge things.
But it looks to me like openSUSE is a troublesome platform for Sage.
>> (That can be the first thing in 4.6.1 if
>> they get reviewed...)
>
> In Sage 4.6.1 we'll most probably have an upgraded readline spkg (6.1,
> #9523) which also solves the problem (at least temporarily, or will also
> include the patch from #9523). But I think Sage 4.6.1 won't be released
> until 4 or 6 weeks...
As I note, 11.2 has been broken for a year.
http://trac.sagemath.org/sage_trac/ticket/9523
should hopefully be a better fix, though that's a little bit more risky.
The only sensible way to get 11.3 suported is for someone to get a machine
(either real or virtual) running 11.3 and available to the buildbot. I'm
personally not going to do that, though I will download these images for
personal use.
> Regards,
> -Leif
Dave
leif
> On 10/27/10 06:09 PM, leif wrote:
>
>>> But if it's going to hold stuff up, I'd rather not delay 4.6 for a
>>> system we don't yet support.
>>
>> Well, "of course"^TM OpenSuSE is at least listed in README.txt as an
>> officially supported platform, and since #9530 doesn't affect any other
>> system but Arch and OpenSuSE I see no reason for not merging it, but
>> hesitate to give it a "dry" positive review since I have no access to
>> any of the above systems.
>
>
>> It would IMHO be a pity to break Sage 4.6 on the two recent releases
>> (11.2 and 11.3) just because the spkg-install applies the relevant
>> (unmodified) patch only on 11.1. (In fact, it doesn't patch Sage's
>> readline but copy / make Sage use the system's one.)
>
> Well, openSUSE 11.2 has been out a year. We have overlooked this for a
> year. It's hard to see why it should now be a blocker.
Because people are likely to meanwhile have upgraded their OS...
(And we got reports that it fails on e.g. 11.3.)
> As I've noted before, README.txt is next to useless when it comes to
> supported platforms.
Users downloading Sage will most probably read *that*.
(Of course there are errors in it; it IIRC states we'd support MacOS X
10.6 on PPC as well, which not even Apple does!)
> http://wiki.sagemath.org/SupportedPlatforms
>
> Should be better.
>
> Also note that the only openSUSE machine we have, which is 11.1 (I
> think), and runs on Itanium, not x86.
As you confirmed some posts ago in this thread, Sage 4.6.rc0 built fine
and passed all tests on menas, which is a openSUSE 11.1 x86_64 Core2
Quad Q6600 box.
;-)
So there's no problem in supporting OpenSuSE (at least 11.1, currently).
> But it looks to me like openSUSE is a troublesome platform for Sage.
I guess just because apparently few Sage developers use it.
If you install 11.2 or 11.3, could you also take a look at #10176 [1]?
-Leif
Florent Hivert
Just to say that I'm will to help on that issue and that I'll be hanging on
IRC for a couple of hours... I don't have to whole evening but I can't free
half an hour on that problem if needed.
Florent
Florent Hivert
> Also note that the only openSUSE machine we have, which is 11.1 (I
> think), and runs on Itanium, not x86.
I'm sorry but my 11.3 machine is my laptop which takes a lot of train so I
can't connect it to the buildfarm. However, my old laptop is openSuSE 11.1 and
shows the same strange behavior. If you need it, I can connect it to the
network and open an ssh port on it for a few days. Just for the info, those
two machine are x86_64 if it's an architecture problems.
Cheers,
Florent
Dr. David Kirkby
I'm downloading 11.1, 11.2 and 11.3 now. I'll give this a go later.
We shall see how this goes.
Dr. David Kirkby
> Dr. David Kirkby wrote:
>> On 10/27/10 06:09 PM, leif wrote:
>>
>>>> But if it's going to hold stuff up, I'd rather not delay 4.6 for a
>>>> system we don't yet support.
>>>
>>> Well, "of course"^TM OpenSuSE is at least listed in README.txt as an
>>> officially supported platform, and since #9530 doesn't affect any other
>>> system but Arch and OpenSuSE I see no reason for not merging it, but
>>> hesitate to give it a "dry" positive review since I have no access to
>>> any of the above systems.
>>
>>
>>> It would IMHO be a pity to break Sage 4.6 on the two recent releases
>>> (11.2 and 11.3) just because the spkg-install applies the relevant
>>> (unmodified) patch only on 11.1. (In fact, it doesn't patch Sage's
>>> readline but copy / make Sage use the system's one.)
>>
>> Well, openSUSE 11.2 has been out a year. We have overlooked this for a
>> year. It's hard to see why it should now be a blocker.
>
> Because people are likely to meanwhile have upgraded their OS...
> (And we got reports that it fails on e.g. 11.3.)
But I doubt this is a recent change in Sage, but rather in openSUSE. So I expect
it has been broken between the time 11.2 was released (a year ago) to the time
11.3 was released (3 months ago). I suspect this was broken for 9 months or so.
>> As I've noted before, README.txt is next to useless when it comes to
>> supported platforms.
>
> Users downloading Sage will most probably read *that*.
Yes, you are probably right. It's just over-hyped.
> (Of course there are errors in it; it IIRC states we'd support MacOS X
> 10.6 on PPC as well, which not even Apple does!)
Perhaps I should add VMS on SPARC then!
>> http://wiki.sagemath.org/SupportedPlatforms
>>
>> Should be better.
>>
>> Also note that the only openSUSE machine we have, which is 11.1 (I
>> think), and runs on Itanium, not x86.
>
> As you confirmed some posts ago in this thread, Sage 4.6.rc0 built fine
> and passed all tests on menas, which is a openSUSE 11.1 x86_64 Core2
> Quad Q6600 box.
Oh, I seem to have overlooked that one. I can't seem to access any buildbot
information now, so I can't double check that later.
> ;-)
>
> So there's no problem in supporting OpenSuSE (at least 11.1, currently).
OK. http://wiki.sagemath.org/SupportedPlatforms
will indicate that.
>
>> But it looks to me like openSUSE is a troublesome platform for Sage.
>
> I guess just because apparently few Sage developers use it.
> If you install 11.2 or 11.3, could you also take a look at #10176 [1]?
>
>
>
> -Leif
Sure. I've got 11.3 installed, but I find it quite frustrating just now. gcc is
not even installed, and installing things seems a pain compared to what I'm used
to. But no doubt a openSUSE expert would feel the same about Solaris!
Dave
Florent Hivert
> But I doubt this is a recent change in Sage, but rather in openSUSE.
> So I expect it has been broken between the time 11.2 was released (a
> year ago) to the time 11.3 was released (3 months ago). I suspect
> this was broken for 9 months or so.
I'm not sure what you are speaking about: there are two different issues:
- The realine problem: #8978, #9530, #9523 which clearly affect all the
openSuSE though there is currently a workaround for 11.1.
In my opinion this problem is settled at least temporarily. I'm ok to testify
that either the fix at #9530 or the new readline 6.1 spkg at #9523 indeed fix
the problem for both openSuSE 11.1 and 11.3 (at least x86_64)
architecture. Depending on how the new spkg is tested on the various other
architecture, I think one can go for on the other solution for the 4.6
release.
- The new problem #10176 which affect both of my machine which are openSuSE
11.1 and 11.3 and which I haven't seen in sage 4.5.3. I'm very doubtful it
is not due to sage.
What is puzzling me is that the install of the spkg write:
running build_ext
Creating /home/data/Sage-Install/readline/sage-4.6.rc0/local/lib/python2.6/site-packages/sagenb.egg-link (link to .)
Removing sagenb 0.8.7-patched from easy-install.pth file
Adding sagenb 0.8.7-patched to easy-install.pth file
But nothing happens...
So far it seems that nobody manage to reproduce it. So just for checking can
someone confirm the md5 hash (I didn't manage to find it on the web):
popcorn-~/src/Sage $ md5sum sage-4.6.rc0.tar
0d01025c3332cb3db41b5a9410970163 sage-4.6.rc0.tar
Cheers,
Florent
Dr. David Kirkby
> Hi David,
>
>> But I doubt this is a recent change in Sage, but rather in openSUSE.
>> So I expect it has been broken between the time 11.2 was released (a
>> year ago) to the time 11.3 was released (3 months ago). I suspect
>> this was broken for 9 months or so.
>
> I'm not sure what you are speaking about: there are two different issues:
>
> - The realine problem: #8978, #9530, #9523 which clearly affect all the
> openSuSE though there is currently a workaround for 11.1.
That's what I was talking about. It seems the readline issue would have broken
openSUSE since the day 11.2 was released.
> In my opinion this problem is settled at least temporarily. I'm ok to testify
> that either the fix at #9530 or the new readline 6.1 spkg at #9523 indeed fix
> the problem for both openSuSE 11.1 and 11.3 (at least x86_64)
> architecture. Depending on how the new spkg is tested on the various other
> architecture, I think one can go for on the other solution for the 4.6
> release.
It would help if you could test it on other systems if you have them. Do you
have an account on sage.math, bsd.math, or skynet? If so, testing on some other
platforms would be helpful. Ultimately, the updated readline will not be
integrated until it's been checked carefully.
> - The new problem #10176 which affect both of my machine which are openSuSE
> 11.1 and 11.3 and which I haven't seen in sage 4.5.3. I'm very doubtful it
> is not due to sage.
I was not talking about that one.
> So far it seems that nobody manage to reproduce it. So just for checking can
> someone confirm the md5 hash (I didn't manage to find it on the web):
> popcorn-~/src/Sage $ md5sum sage-4.6.rc0.tar
> 0d01025c3332cb3db41b5a9410970163 sage-4.6.rc0.tar
That is the correct checksum.
Start here
http://sage.math.washington.edu/home/release/
then poke around and you will find the checksum
http://sage.math.washington.edu/home/release/sage-4.6.rc0/sage-4.6.rc0.tar.md5
I've downloaded openSUSE 11.3, but it has no gcc. I am trying to install gcc and
m4, but it has decided it needs to download 490 MB of dependencies. It seems to
be updating KDE and all manner of things I did not really want it to do. I'll
leave this running over night - by which time I should also have the 11.1 and
11.2 iso images.
> Cheers,
>
> Florent
>
Dave
leif
> On 10/27/10 07:49 PM, leif wrote:
>>> As I've noted before, README.txt is next to useless when it comes to
>>> supported platforms.
>>
>> Users downloading Sage will most probably read *that*.
>
> Yes, you are probably right. It's just over-hyped.
>
>> (Of course there are errors in it; it IIRC states we'd support MacOS X
>> 10.6 on PPC as well, which not even Apple does!)
>
> Perhaps I should add VMS on SPARC then!
Windows 7 (or Vista) on PDP-11?
Who does still support the DEC Alpha I really liked?
It was a really *major* step forward in processor design. (Also SPARC of
course, or the Berkeley RISC it was derived from.)
-Leif
Dr. David Kirkby
> Dr. David Kirkby:
>> On 10/27/10 07:49 PM, leif wrote:
>>>> As I've noted before, README.txt is next to useless when it comes to
>>>> supported platforms.
>>>
>>> Users downloading Sage will most probably read *that*.
>>
>> Yes, you are probably right. It's just over-hyped.
>>
>>> (Of course there are errors in it; it IIRC states we'd support MacOS X
>>> 10.6 on PPC as well, which not even Apple does!)
>>
>> Perhaps I should add VMS on SPARC then!
>
> Windows 7 (or Vista) on PDP-11?
I've not used Windows 7, but I think Vista needs a supercomputer to run it well.
I bought a brand new laptop costing the US equivalent of $3500 a few years ago.
It came with Vista Business. It was too sluggish IMHO, despite the dual core 2.0
GHz chip.
> Who does still support the DEC Alpha I really liked?
Don't ask, otherwise I might fire up my Dec Alpha!! I don't suppose you will be
too surprised to learn I have one of them too! It runs tru64.
At the time of its introduction, the Alpha had amazing floating point
performance - a real step forward.
Sitting around here somewhere I have
* Dec Alpha with tru64
* HP C3600 with PA-RISC processor HP-UX 11.11B
* IBM RS/6000 with PPC running AIX
* SGI Octane with IRIX.
* Numerous SPARCs - Netra T1, Ultra 80, Blade 1000, Blade 2000 running various
release of Solaris.
and of course my Sun Ultra 27, running OpenSolaris which blows the rest away
performance wise - it's pretty quick considering it's only a normal PC with a
single socket CPU.
> It was a really *major* step forward in processor design. (Also SPARC of
> course, or the Berkeley RISC it was derived from.)
Yes, SPARC seems to have lost its way, though Oracle apparently claimed they are
going to double SPARC performance every two years. According to Wikipedia, the
total assets for Oracle are $60 billion, so I guess they have the resources.
(That's more than Intel, though of course Intel's R+D budget is spent almost
exclusively on CPUs, where as Oracles is probably mainly on its database).
> -Leif
>
Dave
Florent Hivert
I finally found what caused the sagenb installation problem. I had the
SAGE_PATH environment variable set to ".". I don't yet fully understand why
but deleting this variable completely solve the problem and putting it back
recreate the problem. Just for the check, can someone test the following on a
completely different architecture:
This is for sage4.6rc0, everything is from SAGE_ROOT
1 - remove the line:
../../../../devel/sagenb
from the file
local/lib/python2.6/site-packages/easy-install.pth
2 - check that sage is broken with error
ImportError: No module named sagenb.misc.sphinxify
Error importing ipy_profile_sage - perhaps you should run %upgrade?
WARNING: Loading of ipy_profile_sage failed.
<ERROR: name 'sage_prompt' is not defined>
3 - sage -f ./sage -f spkg/standard/sagenb-0.8.7.spkg
4 - check that the problem is fixed: sage start normally.
5 - export the environment variable SAGE_PATH with value . ie depending on
your shell:
export SAGE_PATH=.
6 - remove the line as in 1 again
7 - reinstall sagenb as in 3
8 - check that sage is still broken as in 2
9 - removing SAGE_PATH and reinstalling sagenb should fix the problem
I'm investigating this further.
By the way, is there somewhere a tar of a compiled version for
sage.sagemath.org or boxen ?
Cheers,
Florent
Florent Hivert
Sorry for asking ! Found it...
Florent
Florent Hivert
I confirm that sage.sagemath.org shows the same behavior. Let me recall that
from the doc:
# SAGE_PATH - a colon-separated list of directories which Sage searches when trying to locate Python libraries.
I'm not sure at this point what is the right decision:
- Forbid . in SAGE_PATH
- make sure to remove . it from SAGE_PATH when installing spkgs...
Cheers,
Florent
Mitesh Patel
leif
I've implemented the latter, see patch at ticket below (needs review).
> This is
>
> http://trac.sagemath.org/sage_trac/ticket/10192
>
-Leif