Fix a few undistributed packages #3447
Conversation
|
I added a couple more related commits:
|
d-torrance
force-pushed
the
undistributed-pkgs
branch
2 times, most recently
from
141bf85 to
008c9d9
Compare
|
@d-torrance OK, I'm finally going to make an attack on PR's and issues. But I am having trouble with this one, for some reason. Besides having deleted files (which are likely fine), I am having trouble with libomp not being found. That works for me now in development, but not this branch (which I though I had merged into development...). Not sure yet what the problem is... |
Made a variable local and remove a redundant documentation key.
Export two types from Core
Make several variables local
Remove empty Example lists from documentation
Pass a list of strings (not symbols) to "export".
Remove empty Example list from documentation
Fix malformed Authors option
The current versions of these packages still exist in the main packages directory, and these older versions still exist in the git history.
d-torrance
force-pushed
the
undistributed-pkgs
branch
from
008c9d9 to
01d49f3
Compare
|
@mikestillman - I've just rebased it onto |
| @@ -209,7 +209,7 @@ jobs: | |||
| make validate-html | |||
| else | |||
| # otherwise, just validate HTML docs for packages that have been updated | |||
| PACKAGES=$(git diff --stat origin/development HEAD -- ../../Macaulay2/packages/ | grep -Po "(?<=Macaulay2/packages/)[^/\.]*(?=\.m2|/)" | uniq | xargs) | |||
There was a problem hiding this comment.
What is this doing? Is it now going to try to install these packages?
There was a problem hiding this comment.
No -- for every pull request, we do a git diff to see if any packages were updated, and if so, validate the html of these packages' docs. In particular, we look for any changes to an .m2 file of M2/Macaulay2/packages or something in a subdirectory of M2/Macaulay2/packages to generate a list of packages to check.
However, undistributed-packages is a subdirectory of M2/Macaulay2/packages but it isn't a package! So we were trying to validate the html docs of a nonexistent package, leading to occasional failing GitHub builds. So now we filter this case out.
There was a problem hiding this comment.
Ah, thanks, that sounds good!
| @@ -67,7 +67,6 @@ Headline | |||
| an interface to the Risa/Asir website and system | |||
There was a problem hiding this comment.
This package I would like to flesh out, as Risa/Asir has a few very good algorithms
| time syzM = substitute(syz gens M, ring family); | ||
| time eq = compress((gens family * syzM) % G); | ||
| time G := forceGB gens family; | ||
| time syzM := substitute(syz gens M, ring family); |
There was a problem hiding this comment.
This package is perhaps superseded by GroebnerStrata.m2, but in fact there might be some things here that I want to include there. I had forgotten that this package was here.
|
Feel free to merge it. Do you tend to do merge PR's or rebase PR's? |
|
I usually rebase since that gives us a cleaner looking git history without a bunch of merge commits. I'm assuming all the tests will pass, but I'll hold off on merging until they do just in case. |
|
Sounds good! |
This fixes some of the low-hanging fruit from #3446 to get some of the undistributed packages to load.