Repository navigation
Updated the 'Installing Astropy in develop mode' section ... - #1212
Conversation
…y develop in the first instance, added a link to the instructions on how to run the tests, which required adding a label to install.rst), added some additional links to GitHub pages, fixed some typos, applied other minor edits for clarity / consistency
|
I've just noticed the " Error — The Travis CI build could not complete due to an error" message, but I'm not sure why ... I ran "python setup.py build_sphinx" to check that the docs built correctly and they did :( How can I fix this? |
|
@adrn tells me that the Travis build failure looks unrelated to this pull request ... |
|
Thanks for your work on this - I'm restarting the Travis build. |
There was a problem hiding this comment.
Maybe this should give the full command, i.e:
... by running::
$ python setup.py develop -u
There was a problem hiding this comment.
It'd be fine as either python setup.py or ./setup.py. Also for consistency with other parts of the docs I agree this should be a full pre-formatted block with a dollar-sign indicating that it's a shell command:
Develop mode can be easily disabled again by running::
$ python setup.py develop -u
ditto for other examples on this page.
There was a problem hiding this comment.
In general, I don't like ./setup.py since most guides tell users to do python setup.py so I'd prefer if we stick to the latter for consistency (even though I know both will work).
|
Aside from the two comments above, this looks good to me - thanks! I think @eteq and @iguananaut will also want to review this. |
|
@ehogan - just for info, once you make new commits to your branch, you can just push the changes to the same remote branch and the commits will appear here. |
|
@astrofrog - ah, ok, I wasn't sure how that worked, thanks for the info! You mentioned two comments but I can only see one (giving the full command) ... am I missing something? |
|
@ehogan - ah you are right, there was only one comment - ignore me! |
|
IIRC, the reason we didn't encourage usage of |
|
@mdboom - I think the instructions should be there with a caveat that it doesn't work on Python 3, and we can always say it's not the main recommended method. |
|
I still want to strongly encourage it--it greatly cuts down on iteration time. I would almost refuse to do development without it. I don't think most of our contributors (@astrofrog aside) are working regularly in Python 3, though I agree there should be a warning that it won't work there (although I might be able to fix that, it's probably not worth the effort). In #1202 there seems to be a grown consensus toward creating a unified codebase sooner rather than later, at which point this would become a non-issue. |
|
I would say that it is the main recommended method for developers. Outside of Python 3 it makes development vastly easier. |
|
I guess this is a good motivation to try and phase out the use of 2to3 ;) |
|
Agreed--I think that it's great you're doing most of your development in Python 3 and I wouldn't want to make Python 3 users second-class citizens. |
|
A little OT, but do you think it would be feasible to add some hook into |
|
@taldcroft - good idea - let's discuss that in #1219. |
|
I can add a warning to say it doesn't work in Python 3 ... is there something that does work in Python 3 that I could mention instead? |
|
@ehogan - I think in that case one just has to use the |
|
@iguananaut - since this is pretty much ready apart from the comment about Python 3, how about merging this and then rebasing #1126 to add the comment about |
|
Yeah, I'll do that. |
Updated the 'Installing Astropy in develop mode' section ...
Updated the 'Installing Astropy in develop mode' section ... Conflicts: docs/development/workflow/development_workflow.rst
... (use setup.py develop in the first instance, added a link to the instructions on how to run the tests, which required adding a label to install.rst), added some additional links to GitHub pages, fixed some typos, applied other minor edits for clarity / consistency.
This is my first pull request for astropy (I'm at the Scipy2013 sprints!), so let me know if I need to change anything :)