My idea was actually the second one, so nothing has to be changed in current sage packages.I don't see this as so painfull (as the
Debian is currently doing something similar for debian packages (actually for each Debian package there are 3 sources files: a .dsc file, with description and checksum, .diff.gz (the differencies as a patch to pristine sources) and .orig.tar.gz (the pristine sources) I think that this good be a good model to follow. But yes, perhaps is just having tar to report if the opeation of unpacking was sucessfull or not. Pablo On 10/21/07, William Stein <[EMAIL PROTECTED]> wrote: > > On 10/21/07, Pablo De Napoli <[EMAIL PROTECTED]> wrote: > > > > I'm currently working on ticket #329 > > > > My idea is adding to each .spkg file a .spkg.md5 file with the md5checksum > > This should prevent file corruption. > > Are you literally "adding to each .spkg file". If so, > make sure this is completely automatic. I.e., whenever anybody does > sage -pkg directory-version > the md5 file is created inside the resulting spkg. What are you > going to create the md5 hash of, by the way, given that the spkg > doesn't exist when you create the md5 hash to add to the spkg? > The alternative is that we have to have separate files > directory-version.spkg > and > directory-version.spkg.md5 > and then whenever anybody ever wants to trade spkgs, they have > to copy around, get, etc. 2 separate files. That would be painful > in practice. > > Just out of curiosity, shouldn't tar report if the file it is > unpacking is somehow corrupt? Why do we need md5 hashes at all > if the whole point is to determine whether or not a download of > a .tar.bz2 file (an spkg) was corrupted or not? Should we be > able to get that information from tar during the extract process, > or at least change how we make the tarball so that information > is available. > > I really don't want to have to keep track of twice as many files > if it isn't absolutely necessary. > > > > > > I've already reimplemented the md5sum standard utility (from the > > coreutils package) in python (using the md5 module), so that we > > don't need to add an extra dependency to sage. > > > > I still have to modify the logic of the scripts (sage-download-package, > > etc.) > > so that they do the right thing. > > > > Pablo > > > > > > > > > > On 10/20/07, Timothy Clemans <[EMAIL PROTECTED]> wrote: > > > > > > Hi, > > > > > > I think I have done "sage -upgrade" a few times when William was in > > > the process of uploading a new release. I think it would be helpful if > > > Sage would check a file on sagemath that gave the latest release that > > > had been completed uploaded. Another possibility might be that William > > > would upload the files to directories that Sage doesn't look in and > > > then move them over to the release directories after they have been > > > completely uploaded. > > > > > > Timothy > > > > > > > > > > > > > > > > > > > > -- > William Stein > Associate Professor of Mathematics > University of Washington > http://wstein.org > > > > --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~----------~----~----~----~------~----~------~--~---