Julien Puydt <julien.pu...@laposte.net> writes: > Le mercredi 16 mai, William Stein a écrit: >> I'm against shipping the exact upstream tarballs with Sage, as >> explained above (e.g., what if they contain opaque Windows binaries?). > > My impression is that the current situations occur (sorted in very fast > decreasing frequency) : > (1) upstream tarball is ready to ship (tarballs obtained from 'make > dist' with autotools generally belong to that category) ; > (2) upstream tarball is in fact obtained by taking a snapshot from a > repository (there something more might be needed, like running some > kind of autogen.sh, removing .hg/.svn/.git/.whatever) ; > (3) upstream tarball is a mess for various reasons. > > So there is definitely a need for some script to prepare the shipped > tarball. Let me call it spkg-get-upstream-tarball for the rest. > > For (1), that script could just be a one-liner wget (or is that a too > big dependency?).
Python has the ability to download files with the urllib2 module. We don't need wget for this. > For (2), that script would be a few lines to checkout and remove the > cruft. Perhaps sage could even have a generic spkg-get-upstream-from-vs > script which would be called something like : > spkg-get-upstream-from-vs --git git://url/to/upstream/foo > that would take care of doing the checkout and removing the obvious > directories ; that way each spkg-get-upstream-tarball would just call > that generic script, then do additional actions. IMO it's not really necessary to have a script which automates this. We don't upgrade versions of stuff very often, and I think this can easily be done manually. > For (3), that script would do everything by hand. It doesn't make sense to me to have a script that would "do everything by hand". Why not just... do it by hand? > How many (3) are there anyway? Good question. -Keshav ---- Join us in #sagemath on irc.freenode.net ! -- 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