sync scipopt/bliss with ds4dm/Bliss#10
Conversation
- to make the alternative operator representations work with some old C++ compiler - originally scipopt/scip@0495a535129, but added check for C++ version (ciso646 is removed after C++20)
- originally scipopt/scip@3326a19b652
- originally scipopt/scip@63593d7a309
- current cmake may refuse to run with older versions like 3.5
|
@AntoinePrv can you look into this? thanks |
|
@svigerske There is no-one maintaining this repository. The only reason why it was created was to more easily package SCIP at the time, but I personally don't even know what Bliss is. I am happy to help with SCIP pacakging on Conda-Forge on when I find time (SCIP is unrelated to anything I do this days). The changes should probably move from this repository to the scipopt one. |
The changes are from scipopt, and I wanted to add them here, so I could get rid of the scipopt one. What about
ds4dm/Bliss has some GitHub workflow (and more work on this in PR #3, @matbesancon). It might be nice to keep this. There were attempts to update the more original Bliss from digraphs/bliss or https://users.aalto.fi/~tjunttil/bliss/download.html, but that doesn't seem possible (digraphs/bliss#1). |
Fine with me, as I said I don't really have stake in the project so happy to merge what you need. I can also add you as maintainer if you'd like. |
|
From the scip side, we have some interest in a maintained bliss repo. Since you say that you are not too fond of having it in ds4dm, I suggest we just take it out of your hand and move it to scipopt. (The default in SCIP moved from Bliss to Nauty, so the conda-forge builds of the scip optsuite won't need bliss anymore.) I renamed scipopt/bliss to scipopt/bliss.old. If it's ok for you, can you see if you can initiate a transfer of ds4dm/bliss to the scipopt organization? |
2 participants
We would like to retire https://github.com/scipopt/bliss, see #1.
This PR merges some changes from scipopt/bliss that are not in ds4dm/Bliss yet:
-Ipath in the makefile