Michael, Here's a better solution. Tell me if this is closer to what I should be doing.
I copied and unpacked cremona-20071124.p4 from my 2.9 fresh build to a separate location. From in there I did (after "hg incoming") and "hg pull" from some version of my own source tree which had the changes I had committed in it. There was one small conflict (in the top level Makefile) which I fixed. So now I have a copy of cremona-20071124.p4 which does have all the changes (bug fixes etc) which I want to be in the next Sage distribution. Now the question is: how to get this version to you. I see from "hg log" that there (starting from the head) lots of changesets made by me, going back to this one: changeset: 40:6c7dd2a4c889 user: [EMAIL PROTECTED] date: Wed Dec 05 19:34:53 2007 +0100 summary: gcc 4.3 build fixes So if I make a bundle from that as base, you should be able to handle it ok, right? If this works, I know what to do in future: I make sure that my changes apply OK to the current Sage version of the code, managing any conflicts which arise myself, and then submit the appropriate bundle from there, instead of making the bundles in my won tree. I guess the next step (in my education) is to use some kind of "sage -clone" to make a copy of the cremona*spkg source (if that works). John On 19/12/2007, John Cremona <[EMAIL PROTECTED]> wrote: > Sorry, I explained badly: I had not tried to merge with the Sage > source tree at all: it was just that trac displayed an error message > when I tried to upload the bundle. I later saw that, despite the > error messages, 3 versions of the bundle were attached to the trac. > So it seems that trac itself is trying to judge compatibility of a > bundle with some version of the source, but which? I don't know. > > Of course I can send you a tarball of everything as it now is. But I > thought that with hg that would not be necessary. > > If there could be a definitive version of my source code held on > sagemath which I could push to (when I make changes) or pull from (if > anyone else does) then this would be easier for everyone. The effect > would then be the same as having a central definitive repository, as I > am used to from CVS etc. > > Tell me if you still want that tarball. > > John > > On 19/12/2007, mabshoff > <[EMAIL PROTECTED]> wrote: > > > > > > > > On Dec 19, 12:37 pm, "John Cremona" <[EMAIL PROTECTED]> wrote: > > > > Hi John, > > > > > I have changed the mwrank source code so that non-minimal models are > > > now handled properly. All the work is done on a minimal model, but > > > points are mapped back to the input model if that is different. > > > (Non-integral models are still not supported.) > > > > > > Note that it is not correct to say that this was a Sage-wrapping > > > problem! The stand-alone mwrank binary used to manage this properly by > > > converting to a minimal model before calling the two_descent class, > > > and the latter just assumed it was being given a minimal model. The > > > Sage wrapping went straight into the two_descent. Now, the two_descent > > > class handles this itself (so that the mwrank.cc program has less to > > > do). > > > > Excellent. > > > > > I will attach a bundle which corrects this; it should be possible to > > > apply this to the cremona*.spkg in Sage 2.9. > > > > > > No, that failed, since it did not recognise the parent. Help! > > > > > > I tried three different things for the --base in creating the bundle, > > > after looking carefully at the changeset numbers in the current > > > cremona*spkg, but trac would not allow me to upload any of them, > > > saying that it did not recognise the parent. > > > > > > So far my experience with hg has been that it is tolerable when I am > > > working on my own, but impossible to actually share changes in > > > practice. I have read the manual. I know that it all works > > > perfectly for everyone else but me. I don't want to spend the rest of > > > my sage-devel life emailing patches to mabshoff. > > > > Nothing wrong with that, I don't mind doing the integration. > > > > > What am I doing wrong?!! > > > > I assume you might have tried to unbundle against the wrong repo > > since there are two in that spkg, on in the root directory and on in > > src. You need to apply against the one in src. That works for me: > > > > [EMAIL PROTECTED]:/tmp/Work-mabshoff/sage-2.9.1.alpha1/spkg/standard/ > > cremona-20071124.p5/src$ hg unbundle ~/jec-20071219.hg > > adding changesets > > adding manifests > > adding file changes > > added 16 changesets with 28 changes to 131 files (+1 heads) > > (run 'hg heads' to see heads, 'hg merge' to merge) > > [EMAIL PROTECTED]:/tmp/Work-mabshoff/sage-2.9.1.alpha1/spkg/standard/ > > cremona-20071124.p5/src$ hg heads > > changeset: 60:38b80d9b4edb > > tag: tip > > user: John Cremona <[EMAIL PROTECTED]> > > date: Wed Dec 19 11:11:18 2007 +0000 > > summary: two_descent class now works with nonminimal curves > > > > changeset: 44:d371669a38e4 > > user: [EMAIL PROTECTED] > > date: Tue Dec 18 13:38:30 2007 -0800 > > summary: Apply John Cremona's initilization fix for qcurves/ > > sieve_search.cc > > > > [EMAIL PROTECTED]:/tmp/Work-mabshoff/sage-2.9.1.alpha1/spkg/standard/ > > cremona-20071124.p5/src$ hg merge 38b80d9b4edb > > merging Makefile > > conflicts detected in /tmp/Work-mabshoff/sage-2.9.1.alpha1/spkg/ > > standard/cremona-20071124.p5/src/Makefile > > /tmp/Work-mabshoff/sage-2.9.1.alpha1/spkg/standard/cremona-20071124.p5/ > > src/Makefile contains conflict markers. > > Keep this version? [y/n] > > y > > merging procs/interface.h > > merging qcurves/reader.h > > conflicts detected in /tmp/Work-mabshoff/sage-2.9.1.alpha1/spkg/ > > standard/cremona-20071124.p5/src/qcurves/reader.h > > /tmp/Work-mabshoff/sage-2.9.1.alpha1/spkg/standard/cremona-20071124.p5/ > > src/qcurves/reader.h contains conflict markers. > > Keep this version? [y/n] > > y > > merging qrank/getcurve.cc > > 21 files updated, 4 files merged, 0 files removed, 0 files unresolved > > (branch merge, don't forget to commit) > > > > But as you can see above I have two merge conflicts. So can you send > > me a tarball of the complete repo of the code off-list (or post a link > > so I can download it) and I will drop that vanilla repo into the spkg > > and so some testing to make sure everything is still working as > > expected. That way we will also avoid merge conflicts down the road. > > > > > John > > > -- > > > John Cremona > > > > Cheers, > > > > Michael > > > > > jec-20071219.hg > > > 59KDownload > > > > > > > > > -- > John Cremona > -- John Cremona --~--~---------~--~----~------------~-------~--~----~ 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/ -~----------~----~----~----~------~----~------~--~---
