I would like to push updates for m4ri 20130416, m4rie 20130416, and ntl 6.0.0 to Rawhide. Each of those updates involves an soname bump. This will require rebuilds of the following packages:
eclib flint latte-integrale linbox Macaulay2 polybori sagemath Singular
I have built all of these in mock (although the sagemath build is still going; that one takes awhile!). This also incidentally fixes the Macaulay2 failure to build due to a bug in latte-integrale, Rex. Sorry that has taken so long, but I just got the patch from upstream to fix the problem this morning.
I can handle all of the rebuilds unless any of the maintainers want to do it themselves.
If the Singular maintainer is interested, I have worked out how to update it to 3-1-6, and also enable the polymake interface at the same time. Let me know if you would like me to do that as part of the rebuild. We should also consider making a flint-compat or flint1 package, so we can upgrade the flint package to version 2, unless sagemath can be adapted to flint 2.x without too much trouble. Singular wants version 2.
If the other maintainers involved will let me know how they would like to handle this, I can either coordinate the rebuilds or do them myself.
Regards, -- Jerry James http://www.jamezone.org/
2013/5/3 Jerry James loganjerry@gmail.com:
Hi,
Sorry for the delay responding.
I would like to push updates for m4ri 20130416, m4rie 20130416, and ntl 6.0.0 to Rawhide. Each of those updates involves an soname bump. This will require rebuilds of the following packages:
eclib flint latte-integrale linbox Macaulay2 polybori sagemath Singular
I have built all of these in mock (although the sagemath build is still going; that one takes awhile!). This also incidentally fixes the Macaulay2 failure to build due to a bug in latte-integrale, Rex. Sorry that has taken so long, but I just got the patch from upstream to fix the problem this morning.
I started working on updating to sagemath 5.9 that was just released. But if the 5.8 build finished, and most likely did, as most of the time is spent building documentation, it should be ok to update.
I can handle all of the rebuilds unless any of the maintainers want to do it themselves.
If the Singular maintainer is interested, I have worked out how to update it to 3-1-6, and also enable the polymake interface at the same time. Let me know if you would like me to do that as part of the rebuild. We should also consider making a flint-compat or flint1 package, so we can upgrade the flint package to version 2, unless sagemath can be adapted to flint 2.x without too much trouble. Singular wants version 2.
The Singular abi/api is somewhat volatile, so, I prefer to keep at the version used by sagemath. Testing/updating after the m4ri and m4rie updates should be a better plan.
If the other maintainers involved will let me know how they would like to handle this, I can either coordinate the rebuilds or do them myself.
Regards,
Jerry James http://www.jamezone.org/
Paulo
On Mon, May 6, 2013 at 8:18 AM, Paulo César Pereira de Andrade paulo.cesar.pereira.de.andrade@gmail.com wrote:
I started working on updating to sagemath 5.9 that was just released. But if the 5.8 build finished, and most likely did, as most of the time is spent building documentation, it should be ok to update.
No, the build failed because of the new version of NTL. Sagemath's ntl_wrap.{h,cpp} assume that many of the fundamental types (ZZ, ZZ_p, ZZX, etc.) are structs. In NTL 6.0.0, they are classes, not structs. I've got a patch to adapt sagemath to this, but didn't have time to test it over the weekend. I've just started a test build.
If the build succeeds, what would you like me to do? I can send you the patch, and you can work it into the 5.9 update, or I can do a build of 5.8 with the patch.
The Singular abi/api is somewhat volatile, so, I prefer to keep at the version used by sagemath. Testing/updating after the m4ri and m4rie updates should be a better plan.
OK, that makes sense.
Rex, are you okay with me going forward with the rebuilds, or would you like to handle your own? -- Jerry James http://www.jamezone.org/
2013/5/6 Jerry James loganjerry@gmail.com:
On Mon, May 6, 2013 at 8:18 AM, Paulo César Pereira de Andrade paulo.cesar.pereira.de.andrade@gmail.com wrote:
I started working on updating to sagemath 5.9 that was just released. But if the 5.8 build finished, and most likely did, as most of the time is spent building documentation, it should be ok to update.
No, the build failed because of the new version of NTL. Sagemath's ntl_wrap.{h,cpp} assume that many of the fundamental types (ZZ, ZZ_p,
Ok. This will not be one of the easy updates, as sagemath 5.9 besides not changing much "build requires" from sagemath 5.8, had a lot of refactoring on the key portions of the rpm package build.
ZZX, etc.) are structs. In NTL 6.0.0, they are classes, not structs. I've got a patch to adapt sagemath to this, but didn't have time to test it over the weekend. I've just started a test build.
If it works for sagemath 5.8, updating for sagemath 5.9 should be trivial.
If the build succeeds, what would you like me to do? I can send you the patch, and you can work it into the 5.9 update, or I can do a build of 5.8 with the patch.
This is fine, feel free to rebuild sagemath 5.8 in rawhide if you think it is required to avoid breakage for some time/days. If everything goes fine, I will add your patch to the sagemath 5.9 package.
The Singular abi/api is somewhat volatile, so, I prefer to keep at the version used by sagemath. Testing/updating after the m4ri and m4rie updates should be a better plan.
OK, that makes sense.
Rex, are you okay with me going forward with the rebuilds, or would you like to handle your own? -- Jerry James http://www.jamezone.org/
Paulo
On Mon, May 6, 2013 at 12:25 PM, Paulo César Pereira de Andrade paulo.cesar.pereira.de.andrade@gmail.com wrote:
This is fine, feel free to rebuild sagemath 5.8 in rawhide if you think it is required to avoid breakage for some time/days. If everything goes fine, I will add your patch to the sagemath 5.9 package.
Something went wrong with the i386 build. I only tried the x86_64 build in mock, due to the length of time it takes to build. I will do an i386 mock build overnight, so I can fix the problem in the morning. Sorry about that.
Macaulay2 is still building, but everything else on the list is now done in Rawhide. -- Jerry James http://www.jamezone.org/
-
Jerry James -
Paulo César Pereira de Andrade