Conversation
16ad3af to
fa21d64
Compare
|
|
||
| ### Changes related to Groups | ||
|
|
||
| - [#4659](https://github.com/oscar-system/Oscar.jl/pull/4659) Implement `transitivity` and `rank_action` for $G$-sets |
There was a problem hiding this comment.
This won't be in 1.3.0 unless it gets backported, so maybe the script uses master instead of the release-1.3 branch?
There was a problem hiding this comment.
Indeed. This is because I am currently running this script from the branch behind this PR, which also targets master...
Ideally we would have completed this PR 1-2 weeks ago and merged it before branching, but that ship is sailed...
So now we should get this merged ASAP and also backported ASAP, then run the script from release-1.3 ... and hope the magic @aaruni96 implemented on my behest knows how to handle that case 😅 . Otherwise we'll have to run it manually with the appropriate arguments.
There was a problem hiding this comment.
Actually the script may not be able to deal with this even in GAP, now that I think about it: there usually the branching and release of the x.y.0 release are done on the same time and nothing else is merged in between, so there is no need to handle this.
To handle this I think we need two dates:
- the date the previous release branch (here:
release-1.2) was created - the date the current release branch (here:
release-1.3) was created
Then we want all PRs merged between the two dates, but without backport-1.2 label; and all PRs after the second date but with backport-1.3 label.
Does that sound about right?
There was a problem hiding this comment.
Wouldn't it be easier to get the PR numbers from all the commit messages in-between? This would handle everything and you don't have to check for backport labels
There was a problem hiding this comment.
perhaps but then we need a completely different implementation, and e.g. need to fetch the entire history in the GH workflow. Also it is easy to mess up a merge to remove the PR number...
Though of course what I describe is only for the x.y.0 release.
For x.y.(n+1) release, we want to compare to the previous x.y.n release. I am not sure if the backport PRs contain the necessary information? But we can get the release dates for both tags via gh easily and query for all PRs merged in between and with the "backport-x.y" label
|
For reading the changelog, I would find it way more intuitive if the pr numbers in each section increase instead of decrease. In the current setting it could happen that I first read how some function got changed, and a few lines down I read that it got introduced. This somehow feels to be the wrong way around |
fa21d64 to
f553b8b
Compare
|
I am missing a lot of PRs from the November/December period. Does the script handle PRs that were merged to master after the 1.2 branch off, that were not backported? My example I was looking for is #4339 |
f553b8b to
bee449f
Compare
Great catch! My original script supported this, but apparently this was lost at some point. Indeed, I see this in the script log: So it only takes things "after v1.2.2" into account (i.e. the date of that release). But that's not correct. There is an important ingredient missing from the release process compared to what we have in GAP and what the script expects (CC @aaruni96 @benlorenz): when branching off a new release branch, say An alternative is to just use (We use the |
|
A much simpler option for now (that GAP used in the past) is to just manually look up the value and hard code it -- that's the quickest way to get this working and we can look into a proper fix after the release is out. (I've also added a v1.4dev signed tag for experiments) |
bee449f to
cb7a0ca
Compare
|
I added the date hack and now #4339 is there |
cb7a0ca to
79dcd63
Compare
649aa02 to
7b9260e
Compare
7b9260e to
2ebf3dd
Compare
2ebf3dd to
aad7957
Compare
aad7957 to
ef7fc72
Compare
ef7fc72 to
29322b4
Compare
29322b4 to
c4f0002
Compare
c4f0002 to
521f6a1
Compare
Co-authored-by: Max Horn <max@quendi.de>
Co-authored-by: Max Horn <max@quendi.de> Co-authored-by: Lars Göttgens <lars.goettgens@rwth-aachen.de>
Co-authored-by: Lars Göttgens <lars.goettgens@rwth-aachen.de>
Co-authored-by: Lars Göttgens <lars.goettgens@rwth-aachen.de>
* Update changelog on 2025-02-28 Co-authored-by: changelog[bot] <changelog[bot]@users.noreply.github.com> Co-authored-by: Aaruni Kaushik <aaruni96@users.noreply.github.com> Co-authored-by: Max Horn <max@quendi.de> Co-authored-by: Lars Göttgens <lars.goettgens@rwth-aachen.de> (cherry picked from commit 3dc236f) Signed-off-by: Aaruni Kaushik <aaruni@edufor.me>
Backports for 1.3.0 : Add DOI of the book #4654 Add attribute for more detailed info on tunable sections #4636 Move book tests to github runners #4629 Update Nemo to 0.49 #4645 fix a NiceMonomorphism for GAP constructed in Oscar #4640 Polyhedral: fix hash for halfspace and related types #4600 Cleanup some code related to dim(I) == -inf checks #4571 Fix bug in G4-flux computation of big model #4647 Add more missing hash functions #4638 Document how changelog maintenance works #4579 Update changelog #4667
* Update changelog on 2025-02-28 Co-authored-by: changelog[bot] <changelog[bot]@users.noreply.github.com> Co-authored-by: Aaruni Kaushik <aaruni96@users.noreply.github.com> Co-authored-by: Max Horn <max@quendi.de> Co-authored-by: Lars Göttgens <lars.goettgens@rwth-aachen.de>
Automated changes by create-pull-request GitHub action