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
Package cryptominisat 5 #22817
Comments
This comment has been minimized.
This comment has been minimized.
Commit: |
New commits:
|
comment:4
works on freebsd/clang, which is a good approx of OSX. |
comment:5
cryptominisat has a testsuite (see |
comment:6
On the first run I get:
Should not cmake tar ball be already on the mirror network? |
comment:7
Replying to @seblabbe:
Yes it should. |
Changed keywords from days86 to days86, thursdaysbdx |
comment:9
I now get problem compiling cmake ...
I will upload the cmake-3.8.0.log file in a moment. |
Attachment: cmake-3.8.0.log |
comment:10
Please try #22999 - this ticket upgrades cmake to 3.8.1, and we are going forward with it anyway. |
comment:11
Replying to @videlec:
Indeed, i noticed that too. However, if i understand correctly, Also, the testing requires the installation of 7 additionnal modules. Should i package each of them or rebuild a custom tarball from a |
comment:12
On top of current #22999, I get
on OSX 10.10. |
comment:14
Replying to @sagetrac-tmonteil:
This is good for me. Did you open a ticket yet? (note that the package installation went fine on ArchLinux with gcc 6.3.1) |
comment:15
Something must be done regarding this package for OSX (see 22818#comment:33) because sage: from pycryptosat import Solver
---------------------------------------------------------------------------
ImportError Traceback (most recent call last)
<ipython-input-3-280927e3230f> in <module>()
----> 1 from pycryptosat import Solver
ImportError: dlopen(/Users/slabbe/Applications/sage-git/local/lib/python2.7/site-packages/pycryptosat.so, 2): Library not loaded: libcryptominisat5.5.0.dylib
Referenced from: /Users/slabbe/Applications/sage-git/local/lib/python2.7/site-packages/pycryptosat.so
Reason: image not found even if it has successfully installed... |
comment:16
OK as mentioned on #22818 the package needs some love on OS X. I did a few QA on People should check for adverse effects on linux. But on OS X things should now work better. New commits:
|
Changed branch from u/tmonteil/package_cryptominisat_5 to u/fbissey/package_cryptominisat_5 |
comment:17
Actually I'd be particularly interested to see if there are no side effects on debian (ubuntu should be a fine proxy for debian). debian has some particular install rules for libraries that I may have exposed. |
comment:18
On Ubuntu 16.04, on top of 8.0.beta7, I get
Then, sage: from pycryptosat import Solver
sage: s = Solver()
sage: s.add_clause((1r,1r,1r,1r,-1r))
sage: s.solve()
(True, (None, False)) I will check whether it works on my OSX later (my osx being at home). |
comment:19
I had no doubt the installation would be successful. I am just wondering if the library will end in a strange place compared to normal (local/lib/x86_64/ for example). |
comment:20
Does this answer your question?
|
comment:21
Yup. All clear. |
comment:23
Replying to @seblabbe:
In fact, I think I have a problem on Ubuntu, on another run of
Maybe the right command to install is (like for cbc) :
? |
comment:24
Yes, this is because the old cython interface only compiles with cryptominisat 2. This ticket is enough to compile cryptominisat 5, but you have to apply #22818 if you want to compile Sage. |
comment:25
Replying to @sagetrac-tmonteil:
Thank you for your answer, it helped me today to make my sage to work again when I needed it. |
comment:26
Let's make this dependent on #22999. |
comment:27
See also [/22818#comment:50 this discussion]. |
Changed author from Thierry Monteil to Thierry Monteil, François Bissey |
Reviewer: Sébastien Labbé |
comment:29
Let me just remove the cyclic dependency, which might let think that none of both ticket is ready. The description is clear enough. |
Changed branch from u/fbissey/package_cryptominisat_5 to |
Changed keywords from days86, thursdaysbdx to days86, thursdaysbdx, sdl |
Our cryptominisat is currently locked at version 2.9.6, which does not build on 32bit architecture, while the current version of cryptominisat is 5.0.1. This ticket aims at updating the package.
Note that our current cython interface does not work with cryptominisat 5. However, cryptominisat now offers it own python bindings, so we have to rewrite the interface to use that instead.
Note that cryptominisat 5 relies on cmake, which needs to be updated to compile on 32bit architecture as well, see #22814.
Tarball available at : https://github.com/msoos/cryptominisat/archive/5.0.1.tar.gz to be renamed cryptominisat-5.0.1.tar.gz
WARNING Do not merge this ticket until the Sage interface has been updated, see #22818.
Depends on #22814
Depends on #22999
Component: packages: experimental
Keywords: days86, thursdaysbdx, sdl
Author: Thierry Monteil, François Bissey
Branch/Commit:
fd86b98
Reviewer: Sébastien Labbé
Issue created by migration from https://trac.sagemath.org/ticket/22817
The text was updated successfully, but these errors were encountered: