[sage-devel] Re: Options for sage -pkg
On 16 Okt., 14:12, Keshav Kini keshav.k...@gmail.com wrote: (not sure what trac #352, error in matrix creation options, has to do with anything). Apparently Google felt it does... ;-) -leif -- 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
[sage-devel] Re: Options for sage -pkg
On 17 Okt., 17:58, leif not.rea...@online.de wrote: On 16 Okt., 14:12, Keshav Kini keshav.k...@gmail.com wrote: (not sure what trac #352, error in matrix creation options, has to do with anything). Apparently Google felt it does... ;-) FWIW, they weren't smart enough to also manipulate the URL of the ticket reference I gave in my first post. B) -leif -- 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
Re: [sage-devel] Re: Options for sage -pkg
the smaller the file size, the quicker md5sum on it is done. -- 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
Re: [sage-devel] Re: Options for sage -pkg
I think what Leif meant was that most compression formats (including gzip and bz2) implement good enough error detection to obviate md5 hashing. See this post http://trac.sagemath.org/sage_trac/ticket/329#comment:31 on trac #329 (not sure what trac #352, error in matrix creation options, has to do with anything). -- 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
[sage-devel] Re: Options for sage -pkg
On 4 Okt., 12:38, Alexander Dreyer alexander.dre...@itwm.fraunhofer.de wrote: Hi everybody, I've had added the ability to use command line switches together with sage -pkg here:http://trac.sagemath.org/sage_trac/ticket/11676 For instance, you can write sage -pkg -nocompress instead of sage -pkg_nc, which is the current way to do it. We recently had some discussion regarding compression on #352 [1]. IMHO it doesn't make sense to not compress every spkg, i.e., having all spkgs compressed by bzip2 or gzip is advantageous for two reason: * We save space even if some files in the tar are already compressed (at least with gzip; with bzip2 the overhead -- if any -- is very low, although gzip is also faster). * We get integrity checking / checksums without further changes and without additional files. (We may have to support gzip in sage-spkg, but the required changes are trivial.) The drawback is that the need to decompress *every* spkg adds to the CPU time (and the memory usage), but we have only a few spkgs that currently are indeed plain tar files, and we would have almost the same overhead if we computed and checked e.g. md5sums by ourselves, to be stored separately, which complicates things. -leif [1] http://trac.sagemath.org/sage_trac/ticket/329 -- 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
[sage-devel] Re: Options for sage -pkg
I agree that we should just compress every spkg we distribute, but while developing the spkg its nice to be able to turn off compression as it can take quite a bit of time. -- 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
[sage-devel] Re: Options for sage -pkg
On 4 Okt., 15:56, Volker Braun vbraun.n...@gmail.com wrote: I agree that we should just compress every spkg we distribute, but while developing the spkg its nice to be able to turn off compression as it can take quite a bit of time. The change proposed on #352 was to use 'gzip -1' for uncompressed spkgs, which for me at least operates instantly, i.e. without noticable speed penalty. (I also tried the Fortran spkg, which is the only uncompressed standard spkg we currently have, and it's about 33 MB, i.e. amongst the largest spkgs we have. The difference between 'tar xOf ... /dev/ null' and 'tar xOzf ... /dev/null' is in tenth of seconds on an already fully loaded machine. Creation / compressing is similarly fast.) -leif -- 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
Re: [sage-devel] Re: Options for sage -pkg
On 2011-10-04 15:44, leif wrote: * We get integrity checking / checksums without further changes and without additional files. (We may have to support gzip in sage-spkg, but the required changes are trivial.) I really do not understand how md5sums are related to compression. -- 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
