Re: Yet another g++ transition page
Hello Mike, On Thu, Jul 21, 2005 at 05:02:38PM -0400, Mike Furr wrote: > Done. Also, adjacent packages in the dependency graph which are part of > the same source package do not increase the "level" a package lies in > (as suggested by Jochen Voss). Very nice. I am happy to see that the depth is down to 4 now. Thank yoy very much, Jochen -- http://seehuhn.de/ signature.asc Description: Digital signature
Re: Yet another g++ transition page
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Steve Langasek wrote: > Nice. :) Wishlist request: could you have it tally the total number of > packages whose transition depends on each package in the list, and print > that number next to the source package name? Better yet, could the list be > sorted by this number? :) Done. Also, adjacent packages in the dependency graph which are part of the same source package do not increase the "level" a package lies in (as suggested by Jochen Voss). The script also generates the broken package list that I emailed earlier and this is posted there too. Now that I finally got around to finishing these changes, the pages should update regularly again. Cheers, - -Mike -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC4A1s7ZPKKRJLJvMRAp5EAKC1VcTlnGPV14/esZUtZmqDYlgIUwCfX4UH 3kk6tjKaBfs6/DEyO2yPn3I= =k2p0 -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Yet another g++ transition page
Hi Mike, On Mon, Jul 18, 2005 at 01:24:33PM -0400, Mike Furr wrote: > After starting writing a program to help me track my packages > dependencies for the g++ transition, I decided to put up the results for > the entire archive in case it was useful for others. > http://people.debian.org/~mfurr/gxx/ > The main difference between my page and several of the others is that it > uses information from Packages files for every architecture to decide > if it is ready to be nmu'd. Therefore, it shows when a package has been > uploaded, but not built on all architectures. Also, individual > dependency graphs are available for each package affected by the > transition (that my program could detect). > Of course, the results are based on Packages files, which means a lot > could have happened from the time it was last updated and when you're > viewing it. So be sure to check with the packages/buildds before doing > anything. It will be updated ~once/day. Nice. :) Wishlist request: could you have it tally the total number of packages whose transition depends on each package in the list, and print that number next to the source package name? Better yet, could the list be sorted by this number? :) That would certainly be helpful in identifying which packages should be given the highest priority. Thanks, -- Steve Langasek postmodern programmer signature.asc Description: Digital signature
Re: Yet another g++ transition page
Hi Mike, On Mon, Jul 18, 2005 at 04:44:11PM -0400, Mike Furr wrote: > Jochen Voss wrote: > > Just to let you know: ... > yep, same bug. I found another one ;-) The entry "configlet(configlet-frontends)" should be level 6 instead of 8. The tree starts out as configlet-frontends * python-configlet o python2.3-configlet + gksu but configlet-frontends, python-configlet, and python2.3-configlet are all built from the same source package. I think they only have to wait for gksu. Also, if possible, it might be useful to list for each package how many other packages from yout list depend on it. Then it will be easier to spot the real blockers in the very long level 1 list. All the best, Jochen -- http://seehuhn.de/ signature.asc Description: Digital signature
Re: Yet another g++ transition page
On Monday 18 July 2005 11:23 am, Mike Furr wrote: > Daniel Burrows wrote: > > I think you need to take virtual packages into account; apt-based stuff > > is showing up in Level 1, which AIUI is wrong. > > Well, it does the best it can. It resolves all of the information it > has available. However, for example, apt-move has a dependency on > libapt-pkg-libc6.3-5-3.3 which is not provided by any package currently > in unstable so my tool has no possible way of resolving this. If it > depended on libapt...-3.9, then it would correctly correlate this with > the "apt" package. Whoops, I see. So, just to further my comprehension here, why doesn't apt show up in the dependency tree of aptitude? Obviously libsigc has gone through the transition already, so it's not listed, but I'm quite certain that apt hasn't. Daniel -- /--- Daniel Burrows <[EMAIL PROTECTED]> --\ | Gil-Galad was an Elven king; | | of him the harpers sadly sing. | \- A duck! -- http://www.python.org / pgpbYE3zIw8F4.pgp Description: PGP signature
Re: Yet another g++ transition page
Hi Mike, On Mon, Jul 18, 2005 at 01:24:33PM -0400, Mike Furr wrote: > After starting writing a program to help me track my packages > dependencies for the g++ transition, I decided to put up the results for > the entire archive in case it was useful for others. Looks useful. Just to let you know: the link labelled "gforge" (level 8) points to the page http://people.debian.org/~mfurr/gxx/html/g/f/gforge.html which doesn't exist. I hope this helps, Jochen -- http://seehuhn.de/ signature.asc Description: Digital signature
Re: Yet another g++ transition page
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Daniel Burrows wrote: > I think you need to take virtual packages into account; apt-based stuff is > showing up in Level 1, which AIUI is wrong. Well, it does the best it can. It resolves all of the information it has available. However, for example, apt-move has a dependency on libapt-pkg-libc6.3-5-3.3 which is not provided by any package currently in unstable so my tool has no possible way of resolving this. If it depended on libapt...-3.9, then it would correctly correlate this with the "apt" package. Cheers, - -Mike -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC2/O47ZPKKRJLJvMRAj+rAJ9XtHa9YzIp9z+/fnL0nn+qgq8bDQCaApe0 FYREkk1yIWNLD3jOMOmYUEc= =7FnW -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Yet another g++ transition page
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Daniel Burrows wrote: > Whoops, I see. So, just to further my comprehension here, why doesn't apt > show up in the dependency tree of aptitude? Obviously libsigc has gone > through the transition already, so it's not listed, but I'm quite certain > that apt hasn't. Whoops, that would be a bug in that part that generates the individual pages. Jochen Voss wrote: > Just to let you know: the link labelled "gforge" (level 8) points to > the page http://people.debian.org/~mfurr/gxx/html/g/f/gforge.html > which doesn't exist. yep, same bug. It should be fixed in a few minutes... - -m -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC3BSZ7ZPKKRJLJvMRApJAAJ99N6Sjvk+Dl5BMBrQ1TPkvm/hL6QCffsj/ BFltxgZltBzYvfZBy2R/y5U= =8MhS -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Yet another g++ transition page
On Monday 18 July 2005 10:24 am, Mike Furr wrote: > Of course, the results are based on Packages files, which means a lot > could have happened from the time it was last updated and when you're > viewing it. So be sure to check with the packages/buildds before doing > anything. It will be updated ~once/day. I think you need to take virtual packages into account; apt-based stuff is showing up in Level 1, which AIUI is wrong. Daniel -- /--- Daniel Burrows <[EMAIL PROTECTED]> --\ |"Is it too late to extricate myself| | from this plot line?" | |"Yes." -- Fluble | \ The Turtle Moves! -- http://www.lspace.org ---/ pgpS1k9shxFSY.pgp Description: PGP signature