Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Jun 17 21:07
    GitLab | Slava Ostroukh pushed 3 commits to kwant
  • Jun 17 21:07
    GitLab | Slava Ostroukh pushed to kwant
  • Jun 17 21:07
    Slava Ostroukh merged merge request #387 Transpose array-like in herm_conj in kwant
  • Jun 17 20:59
    GitLab | Slava Ostroukh pushed 3 commits to kwant
  • Jun 17 20:54
    Slava Ostroukh commented on merge request #393 check for input length in Monatomic.n_closest in kwant
  • Jun 17 20:46
    Anton Akhmerov closed issue #409 Monatomic.closest doesn't check input dimensionality in kwant
  • Jun 17 20:42
    GitLab | Anton Akhmerov pushed 1 commits to kwant
  • Jun 17 20:42
    Anton Akhmerov merged merge request #393 check for input length in Monatomic.n_closest in kwant
  • Jun 17 20:31
    GitLab | Slava Ostroukh pushed 3 commits to kwant
  • Jun 17 20:29
    GitLab | Anton Akhmerov pushed 5 commits to kwant
  • Jun 17 20:26
    Anton Akhmerov opened merge request #393 check for input length in Monatomic.n_closest in kwant
  • Jun 17 20:25
    GitLab | Anton Akhmerov pushed 1 commits to kwant
  • Jun 17 20:23
    GitLab | Slava Ostroukh pushed 4 commits to kwant
  • Jun 17 20:23
    GitLab | Slava Ostroukh pushed to kwant
  • Jun 17 20:23
    Slava Ostroukh merged merge request #392 Merge CI fixes from master in kwant
  • Jun 17 20:22
    GitLab | Slava Ostroukh pushed 1 commits to kwant
  • Jun 17 20:02
    Anton Akhmerov approveded merge request #392 Merge CI fixes from master in kwant
  • Jun 17 19:57
    GitLab | Slava Ostroukh pushed 5 commits to kwant
  • Jun 17 19:52
    Slava Ostroukh opened merge request #392 Merge CI fixes from master in kwant
  • Jun 17 19:44
    GitLab | Slava Ostroukh pushed 1 commits to kwant
kwant-bot
@kwant-bot
cwg:
But anyway, for changes that consist of multiple commits it's IMHO useful to merge them with an explicit merge commit thus providing an "envelope".
cwg:
I have no statistics on this, but I prefer our approach: fix problems on stable in stable, and merge this into master. Since we've been doing this consistently (except when we forgot), we could simply continue.
slavoutich:
I agree, the difference is not too important. I will rearrange changes on top of stable.

cwg:
Yeah, it's not meant to happen. But then also our INSTALL.rst says:

Kwant tarballs include the Cython generated files, and cythonization is disabled when building not from git. If ever necessary, this default can be overridden by giving the --cython option to setup.py.

anton-akhmerov:
We don't even specify the cython version 🤔
slavoutich:
I still think that "fix in master, cherry-pick to stable" approach is better if we would have to maintain several stable branches.
cwg:
But I'm not going to fight over this, I only think that since this is a recommendation of the Cython authors, and it has been working all the time (except now), that's not enough of a reason to change it.
kwant-bot
@kwant-bot
anton-akhmerov:
Alright, I don't mind keeping it.
cwg:
What I'm trying to say: if you're really motivated, go and remove it. :-)
cwg:
But with Slava's change I think we can actually release 1.4.3 and this will solve the problem as well.
anton-akhmerov:
I'm good. I've removed a few hundred other lines
cwg:
BTW, I've been looking at the vectorization stuff, and this is what I think currently: It's still not clear to me why TermOperator is suppsed to correspond to a "Density". Does the TermOperator have an API to disable the summing like kwant.operator.Density?
anton-akhmerov:
BTW @slavoutich thanks for review, I'll incorporate it shortly
cwg:
(Unfortunately I have to leave really soon, I was hoping to start this discussion earlier, but there were always other things.)
kwant-bot
@kwant-bot
slavoutich:
API is far from being complete for new operators :) Currently operators can be only evaluated. Summing or not is a question of another method, that actually get a wave function, may be not act(ket) -> another_ket but something like expectation_value.
kwant-bot
@kwant-bot
slavoutich:
@cwg I changed the MR to point to stable, but f8331485 needs to be applied before to make PyFlakes happy. Is everything OK now?
kwant-bot
@kwant-bot
bowy:
you can write this for an if and statement?
kwant-bot
@kwant-bot
slavoutich:
Yeah, seems like if a == b == c is a shorthand of if a == b and b == c
kwant-bot
@kwant-bot
anton-akhmerov:
It's mainly meant for writing 1 < 3 < 4, which is handy math notation. However it also works for 1 < 5 > 4, which is more unusual.
kwant-bot
@kwant-bot
anton-akhmerov:
Oops, overlooked the discussion; I'll take it over in the follow-up work.
kwant-bot
@kwant-bot
slavoutich:
@anton-akhmerov Any specific reason for a job failure or just random stuff?
https://gitlab.kwant-project.org/kwant/kwant/-/jobs/210079
anton-akhmerov:
SSH config changed in december due to migrations
slavoutich:
Can you pin the relevant commit?
slavoutich:
This one?
anton-akhmerov:
Yeah, that's what I get as well
kwant-bot
@kwant-bot
slavoutich:
No, there are two subsequent commits after the last stable -> master merge
slavoutich:
Another stone into the discussion how it is better to proceed: even since we are trying to carefully fix in stable and merge to master, we anyway end up with cherry-picking stuff from master into stable.
kwant-bot
@kwant-bot
slavoutich:
By the way, I liked Sphinx branch naming conventions: in our case the default branch would be 1.x and stable branch would be 1.4.x. Naming starts to have more sense IMO.
kwant-bot
@kwant-bot
slavoutich:
I wonder why tests virtually freeze during the test_gauge phase in CI...
kwant-bot
@kwant-bot
slavoutich:
Any idea why this failure happens?
https://gitlab.kwant-project.org/kwant/kwant/-/jobs/210111
kwant-bot
@kwant-bot
anton-akhmerov:
Did you turn on coverage reports?
anton-akhmerov:
It's a tight cython loop, and coverage increases runtime by a lot.
slavoutich:
Nope, there are too many changes in master, that are not in stable
slavoutich:
And I think that the idea of cherry-picking them is also not perfect. Mind if I just copy the .gitlab-ci.yml from master as a single commit and that's it?
kwant-bot
@kwant-bot
slavoutich:
Something like this instead of twenty cherry-picks.
kwant-bot
@kwant-bot
anton-akhmerov:
I don't mind if it merges
kwant-bot
@kwant-bot
slavoutich:
Should I merge or wait for Christoph?
https://gitlab.kwant-project.org/kwant/kwant/-/merge_requests/392
kwant-bot
@kwant-bot
anton-akhmerov:
That one's good I think
kwant-bot
@kwant-bot
anton-akhmerov:
Mergefest? Mergefest!
slavoutich:
Merge, merge, merge!!!
kwant-bot
@kwant-bot
anton-akhmerov:
I'm going to YOLO it and merge my bugfix without review.
kwant-bot
@kwant-bot
anton-akhmerov:
Uh, oh, I was faster; now gotta rebase your MR
kwant-bot
@kwant-bot
slavoutich:
Don't do it, even for 1-string changes
anton-akhmerov:
my bad. sorry
kwant-bot
@kwant-bot
anton-akhmerov:
Shall we also merge stable → master for a good measure?
slavoutich:
I think that yes: fixes that are not in master is a sin
anton-akhmerov:
👍