Re: [sage-devel] Re: Broken optional packages...
Hi Jean-Pierre, On Wed, Feb 05, 2014 at 06:00:58AM -0800, Jean-Pierre Flori wrote: > > See also > https://groups.google.com/d/msg/sage-devel/zlyuxWIcERQ/cLPRt2nufr4J > > You surely spotted that, but every compiled cython file is present twice in > the sage tree. Actually, when running "make lib-clean", it calls "make clean" from the src/ directory that removes the whole src/build directory, so i did not notice those directories. But, perhaps this behaviour contradicts Francois' statement that "src/build/cython_debug is useful". Also, should we add the removal of src/doc/output/doctrees and src/doc/output/inventory somewhere in one of the Makefiles, and if yes, in which target ? Ciao, Thierry -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.
Re: [sage-devel] Re: Broken optional packages...
On Wednesday, February 5, 2014 1:54:56 PM UTC, Thierry (sage-googlesucks@xxx) wrote: > > - It is still possible to use/develop Sage without any problem after > such a strip ? > - Are some Sage debugging tool that need unstripped binaries ? > Clearly segfaults are going to be a lot less useful with stripped binaries, so I would say developing gets a lot harder IF you run into a problem that needs debugging. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.
Re: [sage-devel] Re: Broken optional packages...
On Wednesday, February 5, 2014 2:54:56 PM UTC+1, Thierry (sage-googlesucks@xxx) wrote: > > On Tue, Feb 04, 2014 at 12:57:04PM -0800, William Stein wrote: > > It would be cool if there were a make target that did all of this..., > > e.g., "make cache-clean" since it is not at clear from the name that > > "upstream" is a cache directory that can be safely deleted, especially > > given that in the previous version of sage it (=spkg/standard/) had > > much the same content, and deleting it would be devastating. > > +1 ! > > Could someone tell what other directories/files can be removed without > any loss. I guess the following are OK, but i am not sure: > - src/doc/output/doctrees > - src/doc/output/inventory > > Also, the Makefile has a 'micro_release' target that strips all > binaries. > - To what extend does this affect the build ? > - It is still possible to use/develop Sage without any problem after > such a strip ? > - Are some Sage debugging tool that need unstripped binaries ? > > > Now, from a non-developper point of view (i mean someone that will not > touch the sources, but still want to be able do do 'sage -i' to install > optional packages), what more can be removed ? > > - .git > - sage/local/libexec/git-core > - sage/local/share/git* > - what else ? > > See also https://groups.google.com/d/msg/sage-devel/zlyuxWIcERQ/cLPRt2nufr4J You surely spotted that, but every compiled cython file is present twice in the sage tree. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.
Re: [sage-devel] Re: Broken optional packages...
On Tue, Feb 04, 2014 at 12:57:04PM -0800, William Stein wrote: > It would be cool if there were a make target that did all of this..., > e.g., "make cache-clean" since it is not at clear from the name that > "upstream" is a cache directory that can be safely deleted, especially > given that in the previous version of sage it (=spkg/standard/) had > much the same content, and deleting it would be devastating. +1 ! Could someone tell what other directories/files can be removed without any loss. I guess the following are OK, but i am not sure: - src/doc/output/doctrees - src/doc/output/inventory Also, the Makefile has a 'micro_release' target that strips all binaries. - To what extend does this affect the build ? - It is still possible to use/develop Sage without any problem after such a strip ? - Are some Sage debugging tool that need unstripped binaries ? Now, from a non-developper point of view (i mean someone that will not touch the sources, but still want to be able do do 'sage -i' to install optional packages), what more can be removed ? - .git - sage/local/libexec/git-core - sage/local/share/git* - what else ? Doing all that, i was able to build Sage in a 284M squashfs file, which is not bad if we want to have live CD back (or smaller live USB). Ciao, Thierry -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.
[sage-devel] Re: Broken optional packages...
Hi, Another thing I do when building Sage is delete the large number of cached files that get downloaded as part of installing standard or optional packages. This used to involve deleting spkg/optional/*. Now what you do is: rm -rf upstream from SAGE_ROOT. For example, in my install that upstream directory was "wasting" 3.4GB. Also, to delete left-over files from failed package installs: rm -rf local/var/tmp/sage/build/ Thanks to Andrew Ohana for clarifying this. It would be cool if there were a make target that did all of this..., e.g., "make cache-clean" since it is not at clear from the name that "upstream" is a cache directory that can be safely deleted, especially given that in the previous version of sage it (=spkg/standard/) had much the same content, and deleting it would be devastating. -- William On Sat, Feb 1, 2014 at 10:36 PM, William Stein wrote: > Hi, > > I just wanted to raise awareness (and apologize) that the > re-organization of the layout of Sage broke so many optional spkg's. > In case people run into similar problems, here's what I'm doing to > workaround this mess for sage-6.2 in SageMathCloud. > > 1. These databases are broken: > > cunningham_tables > database_jones_numfield > database_kohel > database_odlyzko_zeta > database_sloane_oeis > database_stein_watkins_mini.p0 > > One workaround is to do this: > > # in SAGE_ROOT > ln -s local/share data > > but the SAGE_DATA variable is gone as well... so you have to add this > *BACK* to src/bin/sage-env (after SAGE_SHARE): > > # this is to be compatible with optional packages that haven't > # been updated to not use SAGE_DATA > export SAGE_DATA="$SAGE_SHARE" > > (Hopefully, this can be added back or the packages fixed for sage-6.3.) > > 2. Another broken package is > > p_group_cohomology > > which doesn't build: > > pGroupCohomology/resolution.pxd:21:0: 'interrupt.pxi' not found > > I don't know a fix. > > 3. There's an optional package called pybtex which doesn't work. > > The package TOPCOM-0.17.4 fails to build: > > ./spkg-install: line 11: cd: > /usr/local/sage/sage-6.1/spkg/standard/: No such file or directory > > I don't know a fix. > > This is all on Ubuntu 12.04 with pretty much anything imaginable installed. > > Trac http://trac.sagemath.org/ticket/14962 is supposed to be about > packages broken by the git reorganization, but it's closed. > > -- William > > -- > William Stein > Professor of Mathematics > University of Washington > http://wstein.org -- William Stein Professor of Mathematics University of Washington http://wstein.org -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.
Re: [sage-devel] Re: Broken optional packages...
> On 2014-02-03, anne1.schill...@gmail.com wrote: >> make: *** [build] Error 1 >> :sage anne$ more /Applications/sage/logs/pkgs/cchache.log >> Attempting to download package cchache > > Do you mean ccache instead of cchache? Well, this is what the computer says. Perhaps it was installed with a typo? How can I get rid of it, since now my make does not complete compilations any longer. Best, Anne -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.
[sage-devel] Re: Broken optional packages...
Hi Anne, On 2014-02-03, anne1.schill...@gmail.com wrote: > make: *** [build] Error 1 >:sage anne$ more /Applications/sage/logs/pkgs/cchache.log > Attempting to download package cchache Do you mean ccache instead of cchache? Best regards, Simon -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.
[sage-devel] Re: Broken optional packages...
Hi William, Zitat von William Stein : 2. Another broken package is p_group_cohomology which doesn't build: pGroupCohomology/resolution.pxd:21:0: 'interrupt.pxi' not found I don't know a fix. I do, and it ''is'' fixed: See http://sage.math.washington.edu/home/SimonKing/Cohomology/p_group_cohomology-2.1.4.p1.spkg and trac ticket #15666. As stated in comment 4, I would prefer to have this reviewed and fixed soon, so that it builds and works in recent Sage versions. Other items on the to do list, such as transition to the new spkg layout, are less urgent from my perspective, and is nothing that I'd have the capacity to do right now. Best regards, Simon This message was sent through https://webmail.uni-jena.de -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.
[sage-devel] Re: Broken optional packages...
On Sat, Feb 1, 2014 at 10:36 PM, William Stein wrote: > Hi, > > I just wanted to raise awareness (and apologize) that the > re-organization of the layout of Sage broke so many optional spkg's. > In case people run into similar problems, here's what I'm doing to > workaround this mess for sage-6.2 in SageMathCloud. > > 1. These databases are broken: > > cunningham_tables > database_jones_numfield > database_kohel > database_odlyzko_zeta > database_sloane_oeis > database_stein_watkins_mini.p0 > As per Volker's post ( https://groups.google.com/forum/#!topic/sage-devel/onWFAVz-LDA) these packages need updated copyright information. The copyright for the sloane database is provided in that thread, as is permission to use Andrew Odlyzko's database, but for the others copyright information is still needed. One workaround is to do this: > > # in SAGE_ROOT > ln -s local/share data > > but the SAGE_DATA variable is gone as well... so you have to add this > *BACK* to src/bin/sage-env (after SAGE_SHARE): > > # this is to be compatible with optional packages that haven't > # been updated to not use SAGE_DATA > export SAGE_DATA="$SAGE_SHARE" > > (Hopefully, this can be added back or the packages fixed for sage-6.3.) > > 2. Another broken package is > > p_group_cohomology > > which doesn't build: > > pGroupCohomology/resolution.pxd:21:0: 'interrupt.pxi' not found > > I don't know a fix. > > 3. There's an optional package called pybtex which doesn't work. > > The package TOPCOM-0.17.4 fails to build: > > ./spkg-install: line 11: cd: > /usr/local/sage/sage-6.1/spkg/standard/: No such file or directory > Using this sage -i topcom works just fine for me. If you are doing sage -i TOPCOM you are probably getting the old spkg that doesn't work with the new layout (this should really be fixed so that the packages are case-insensitive). I don't know a fix. > > This is all on Ubuntu 12.04 with pretty much anything imaginable installed. > > Trac http://trac.sagemath.org/ticket/14962 is supposed to be about > packages broken by the git reorganization, but it's closed. > > -- William > > -- > William Stein > Professor of Mathematics > University of Washington > http://wstein.org > -- Andrew -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.