These are chat archives for astropy/astropy

22nd
Jun 2017
Kelle Cruz
@kelle
Jun 22 2017 16:15
Hello gitter people. I'm struggling with something I've struggled with before but still haven't mastered: I want to make an edit to something in astropy but my local version is out of date with upstream. I pulled from upstream, resloved some conflicts, and pushed to my branch. HOWEVER, there's still a conflict with a file (docs/index.rst, which has been under lots of development recently). Is there a best way to just overwrite my local with the upstream?
^push to my fork. the master branch on my fork.
matrixbot
@matrixbot
Jun 22 2017 16:16
bsipocz have you already made changes, or want to start from a fresh branch?
bsipocz what does git remote -v show?
Kelle Cruz
@kelle
Jun 22 2017 16:18
I havent made the changes yet. I guess I should start from a fresh branch...I just want to add the "What's New" line! Here's my git remote -v
adrn git@github.com:adrn/astropy.git (fetch)
adrn git@github.com:adrn/astropy.git (push)
mhvk git@github.com:mhvk/astropy.git (push)
mhvk git@github.com:mhvk/astropy.git (fetch)
origin git@github.com:kelle/astropy.git (fetch)
origin git@github.com:kelle/astropy.git (push)
upstream git@github.com:astropy/astropy.git (push)
upstream git@github.com:astropy/astropy.git (fetch)
I could edit directly on github but I want to get better at this...
matrixbot
@matrixbot
Jun 22 2017 16:18
bsipocz ok, then:
git fetch --all
bsipocz git checkout -b add_whatsnew upstream/master
bsipocz that will create a new branch originating at the top of the current astropy master
bsipocz and you should be good to go
bsipocz ohh, what does git status say? I assume you don't have any uncommitted changes hanging around, right?
Kelle Cruz
@kelle
Jun 22 2017 16:20
it will create a new branch which I will then push to my fork in order to initiate the PR?
matrixbot
@matrixbot
Jun 22 2017 16:21
bsipocz yes, add your commits, then git push kelle add_whatsnew
bsipocz when you go either your fork or the astropy/astropy repo, github will recognize the freshly pushed branch, and automatically offer the option to open a PR
Kelle Cruz
@kelle
Jun 22 2017 16:22
I have some files showing up in git status but they are "untracked" things I dont care about.
matrixbot
@matrixbot
Jun 22 2017 16:22
bsipocz yeah, they shouldn't cause any troubles
bsipocz if you have some modified files, they sometimes create conflicts when you change branches
Kelle Cruz
@kelle
Jun 22 2017 16:23
excellent, thanks! And I guess I should put the issue in the commit? #6231
matrixbot
@matrixbot
Jun 22 2017 16:24
bsipocz you don't need to refer to the issue in the commit message, when merging we'll have the PR number there anyway
Kelle Cruz
@kelle
Jun 22 2017 16:25
and should I always do git fetch --all? What's the workflow I should follow each time to get my local and fork up-to-date? or if I do a fresh branch each time, it doesnt matter?
matrixbot
@matrixbot
Jun 22 2017 16:26
bsipocz Well, doing a git fetch upstream should be enough, but I prefer to sync all the remotes I have, too.
bsipocz This only makes your remotes up-to-date, but then you can start up new branches off those remotes.
bsipocz e.g. we usually suggest to get rid of the master branch of the fork, so there is no confusion when it goes off sync off locally from upstream/master
Kelle Cruz
@kelle
Jun 22 2017 16:28
ahhh. Yes, I could see that being useful.
matrixbot
@matrixbot
Jun 22 2017 16:28
bsipocz (though thinking about it, probably it's not yet put back to the astropy contribution docs)
bsipocz also, while I love the command line in general, something it's nice to have a git gui to visualize all the branches out there
Kelle Cruz
@kelle
Jun 22 2017 16:31
oh, in the commit message I should add something for the ci?
matrixbot
@matrixbot
Jun 22 2017 16:32
bsipocz you can do [docs only] if it only changes rst files
Kelle Cruz
@kelle
Jun 22 2017 16:36
ok, PR made and requested you as a reviewer. Thanks!
matrixbot
@matrixbot
Jun 22 2017 16:37
bsipocz thanks!
Zé Vinícius
@mirca
Jun 22 2017 23:35
@keflavich are you around?
Adam Ginsburg
@keflavich
Jun 22 2017 23:35
yes
@mirca yes
Zé Vinícius
@mirca
Jun 22 2017 23:37
@keflavich let me know if I can do anything else in #6267
Adam Ginsburg
@keflavich
Jun 22 2017 23:39
@mirca are you going to pull out the scipy convolution tests from #6015 and move them into #6267 (into test_convolve.py)? You don't have to, I just think it's a bit cleaner
Zé Vinícius
@mirca
Jun 22 2017 23:44
@keflavich Sorry, which test are you referring to?
you added the numpy convolution test (great!), so I was suggesting also making a scipy one and pulling all scipy uses out of 6015
Zé Vinícius
@mirca
Jun 22 2017 23:49
Oh, ok, cool
Actually, this test checks for fitting astropy “convolved” models, so, it needs to be in #6015. But yeah, I’m gonna write one against scipy in #6267
Adam Ginsburg
@keflavich
Jun 22 2017 23:53
ok, thanks!