Re: gfortran-4.3 the default gfortran compiler, blas and lapack uploaded to unstable
Matthias Klose writes: > I did have a look at Camm's packages on p.d.o., fixed some obvious > things and uploaded those to unstable; Camm did change all package > names and the soname of all libraries, such that these packages do not > conflict with the existing packages in unstable, therefore the > decision to upload those to unstable. Thanks for this! > - use the unversioned compiler again for all of blas, lapack, atlas. > > - don't use -ffloat-store in the blas build (currently ix86 only) > > - update lapack to 3.1 About Lapack, I have this to say; if you recall, you uploaded a lapack I prepared in October to experimental (version 3.1); why not try that in experimental with the new Blas? Please note that the current Blas in unstable which you uploaded yesterday is the new upstream version, which _borrows_ some code from the new lapack, so I am wondering if it would be a good idea to build the old Lapack against the new Blas. Also, the new upstream Lapack has several of Camm's old patches applied upstream, so I am wondering what strategy can be adopted to counter this, as I really don't know which patches to keep, which not to. Finally, another very important TODO: Update Atlas to 3.8.0: https://sourceforge.net/project/showfiles.php?group_id=23725 Thanks. Kumar -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
How to handle ICEs on buildds
Dear Debian GCC Maintainers, With reference to the following build log: http://buildd.debian.org/fetch.cgi?pkg=lam&arch=alpha&ver=7.1.2-1.2&stamp=1202605929&file=log&as=raw The build fails due to an internal compiler error. Now, as many of the interested people do not have access to the machine of the relevant architecture, we can't file the bugs ourselves. If we do file a bug, the gcc maintainers usually ask us to report it upstream, which is not possible in my case. Therefore, I was wondering whether you could outline the procedure you would recommend to handle these cases. Many thanks! Kumar P.S. 1. Please CC me, as I am not in the list. 2. The above seems to be related to PR33410, which Martin Michlmayr has been following up. Could you confirm, please, Martin? -- Kumar Appaiah, 458, Jamuna Hostel, Indian Institute of Technology Madras, Chennai - 600 036 signature.asc Description: Digital signature
Bug#442036: Instance of this bug for lam on alpha
Dear GCC maintainers (and Martin in particular), I have reason to believe that lam, which was uploaded to unstable to use gfortran after this bug was found, is also affected by this bug, from [1]. Could you confirm that this is the case? Thank you. Kumar [1]: http://buildd.debian.org/fetch.cgi?pkg=lam&arch=alpha&ver=7.1.2-1.2&stamp=1202605929&file=log&as=raw -- Kumar Appaiah, 458, Jamuna Hostel, Indian Institute of Technology Madras, Chennai - 600 036 signature.asc Description: Digital signature
Bug#442036: Instance of this bug for lam on alpha
On Tue, Feb 19, 2008 at 02:10:04PM +0100, Martin Michlmayr wrote: > * Kumar Appaiah <[EMAIL PROTECTED]> [2008-02-19 17:41]: > > I have reason to believe that lam, which was uploaded to unstable to > > use gfortran after this bug was found, is also affected by this bug, > > from [1]. Could you confirm that this is the case? > > Definitely looks like it. The good news is that this bug has been > fixed upstream, so all you need is for doko to upload a new gcc-4.2 > package. That is good to know. What I wanted to confirm is that the bug will be fixed in the gcc version used for that build as well, which happens to be 4.3-20080202-1. Therefore, I guess I can request a give back only after an upload of gcc-4.3 is made with this fix. Is this correct? Thanks! Kumar -- Kumar Appaiah, 458, Jamuna Hostel, Indian Institute of Technology Madras, Chennai - 600 036 signature.asc Description: Digital signature
Help with gcc regression on ia64
(Cross-post) Dear Debian GCC and IA-64 users, I would request your help in investigating lam FTBFSing on ia64 due to an ICE. The log is here: http://buildd.debian.org/fetch.cgi?pkg=lam&arch=ia64&ver=7.1.2-1.4&stamp=1207431083&file=log&as=raw From my initial guess, this should have been fixed in the Debian bug #447426, but has shown its ugly face again! :-( This bug is significant since it blocks the migration of this package and several other packages to testing. So, I'd appreciate help with it. Thanks in advance. Kumar -- Kumar Appaiah, 458, Jamuna Hostel, Indian Institute of Technology Madras, Chennai - 600 036 signature.asc Description: Digital signature
Bug#441481: Forwarded Upstream
forwarded 441481 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33382 tags 441481 +upstream thankyou Dear Debian GCC team, FYI, the bug is confirmed upstream as well. So, I am tagging it appropriately. Thanks. Kumar -- Kumar Appaiah, 458, Jamuna Hostel, Indian Institute of Technology Madras, Chennai - 600 036 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
