New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
upgrade gfan to latest release (0.4plus) #7820
Comments
This comment has been minimized.
This comment has been minimized.
Author: Alex Ghitza |
comment:2
An updated spkg is up at http://sage.math.washington.edu/home/ghitza/gfan-0.4plus.spkg Some fixes are needed in the Sage library because of the upgrade. I am attaching a patch that deals with the most obvious one, but there are some doctest failures left:
I can try to figure out what's happening, but Marshall is likely to be better at it. |
comment:3
Yes, I can try to work on fixing those issues. This relates a little bit to some major refactoring of the polyhedron class over at ticket #7109. If you could review that I'd greatly appreciate it. Thanks very much for working on the gfan upgrade. |
comment:4
I'm starting to think that rays doctest might have exposed a bug in gfan itself, I'll check with Anders. |
comment:5
The new output is correct; the rays for that example are not uniquely determined. So the doctests just need to be changed to match the new output. |
comment:6
It would be good if the Makefile could be changed so that the C compiler is set by CC and the C++ compiler by CXX, and the flags CFLAGS and CXXFLAGS used. The Sun compiler was not happy with the earlier version. I dont mind trying to sort the makefile out if you want. Dave |
comment:7
Replying to @sagetrac-drkirkby:
Sure! That would make it easier for me. |
comment:8
I'll do that today, by 1800 GMT today (7 hours from now). It was coincidental, but I'd just hit the problem on version 0.3 and posted something to sage-devel about this. William was willing to remove gfan, as he was no aware of anyone using it. You might want to put a comment on sage-devel about it. Dave |
comment:9
I'm a bit concerned about the number of compiler warnings here from g++, even though -Wall was not added. On the 0.3 release, the Sun compiler rejected some code, saying things were masking others, and functions expected to return something did not. I'll get the Makefile fixed so gfan 0.4plus it at least attempts a build with Sun's compiler, but I suspect Sun's will reject some of the code as being invalid. Dave |
comment:10
The 'src' directory at http://sage.math.washington.edu/home/ghitza/gfan-0.4plus.spkg has no 'makefile' yet when I download the source code from http://www.math.tu-berlin.de/~jensen/software/gfan/gfan0.4plus.tar.gz there is a makefile. I think we should keep that, then overwrite it with a patch, so there is a record of the actual source code. i.e. something like cp patches/makefile src/makefile That way, the original source is untouched. I rather start with the original code, and make changes to that, rather than someone elses makefile. Dave |
comment:11
Well, I'm 3 minutes late, but I hope you will excuse that. I've put a new spkg at http://boxen.math.washington.edu/home/kirkby/portability/gfan-0.4plus/ I used a clean source, and just edited spkg-install and made patches/Makefile. Can you check I've note undone anything you have done, and check it works on a couple of systems. I've only checked on Solaris here. Dave |
Changed author from Alex Ghitza to Alex Ghitza, David Kirkby |
comment:13
Replying to @sagetrac-drkirkby:
I'm not sure what spkg file you're looking at. I just rechecked the one I gave above and it has the original Makefile in src/ and a modified one in patches/. Anyway, it's not important. I'm looking at your spkg now. |
comment:14
Sorry about the confusion. I can only assume I confused it with the 0.3 version by mistake. I certainly downloaded your one. I downloaded the gfan-0.4plus source code, and put that in the package. |
comment:15
It looks good, thanks for this! So far I've tested it on 32-bit archlinux and 64-bit archlinux without problems. Same on 32-bit MacOSX 10.5. There are a couple of minor things in your spkg, which I can easily fix today: we normally strip the original sources of documentation and similar things. In this case, we get rid of src/doc, src/examples, and src/homepage. This gives an spkg that's 220kb instead of 680kb. There are also a few typos in Makefile and SPKG.txt. As I said, this is great, and I can make these minor fixes today. |
comment:16
It's good to hear it was mostly there. At least one can to build gfan with Sun's compiler now, even though it soon fails. Before it was impossible to get anything useful done, as gcc was hard-coded in the Makefile. Now at least one can see the error messages. Unfortunately, whilst I know C, I do not know C++, so don't have much clue about how to resolve the issues. I'll forward them upstream. Dave |
apply after the gfan-0.4plus spkg |
This comment has been minimized.
This comment has been minimized.
comment:17
Attachment: trac_7820.patch.gz OK, I have replaced the spkg with one that has the small fixes to David's version. See the ticket description for the URL. This spkg also patches I have also replaced the patch to the Sage library with one that deals with all the issues raised by the upgrade. All of this is now ready for review. |
comment:18
Replying to @aghitza:
Forgot to mention that the bug was reported upstream, and Anders Jensen quickly sent us the fix that we are now using. It will be incorporated in the next version of gfan. |
This comment has been minimized.
This comment has been minimized.
comment:20
I've tested this and I think it looks good. There are some minor conflicts with 7109 but I am willing to rebase 7109 against this if it is merged first. I might not have time until this summer to extend the gfan interface but I am interested in doing so. If anyone beats me to it I am happy to review. |
Reviewer: Marshall Hampton |
Merged: sage-4.3.2.alpha0 |
comment:21
Merged in this order:
|
The new spkg is at
http://sage.math.washington.edu/home/ghitza/gfan-0.4plus.spkg
See
http://www.math.tu-berlin.de/~jensen/software/gfan/gfan.html
Release 0.4plus has improved performance and a lot of new functionality.
This also takes care of the following issues: remove the debian dist directory (see #5903), clarify the license (see #3043), and separate the clean upstream from the patches needed to build in Sage (see #3338).
CC: @sagetrac-mhampton @williamstein
Component: packages: standard
Author: Alex Ghitza, David Kirkby
Reviewer: Marshall Hampton
Merged: sage-4.3.2.alpha0
Issue created by migration from https://trac.sagemath.org/ticket/7820
The text was updated successfully, but these errors were encountered: