cmw/rp-60972-make-hapipy-to-be-python2/3-compatible #6
Conversation
… python2/3 compatible
…or error.py run again.
|
I think that once we release the portal code to master that uses this branch, we would want to merge this to master as well. As you note, doing that doesn't really change anything except make it easier for future revisions. |
|
Which branch from the upstream repo did you use as the base for this branch? I want to just look at our changes (which should be small) |
I build my branch on top of the master branch of this CbitLabs repo. If you want to see our customized changes, you could check the commit history of the master branch. https://github.com/CBitLabs/hapipy/commits/master @pjsg |
|
The upstream master should have been merged in first (from https://github.com/HubSpot/hapipy) |
We stop rebasing our master branch (CbitLabs) on top of the master branch(HupSpot) since June 19, 2015 and we just keep doing our own changes and version bumps while there are new but less important changes in the upstream at the sam time. Do you think it is necessary to accept upstream's updates again now? If so, I would rebase our master branch first and backmerge master branch for this working branch. |
|
Deployed codes to iso1 instance. In BitSight Technologies Corporate company page, I made an EVA request and succeeded in receiving 2 follow up emails from eva@bitsight.com. Feel free to check the list here. https://app.hubspot.com/contacts/277648/contact/139430878/ |
… handle __str__ in HapiError class smoothly
| raise HapiTimeout(None, request, traceback.format_exc()) | ||
|
|
||
| encoding = [i[1] for i in result.getheaders() if i[0] == 'content-encoding'] | ||
| encoding = [i[1] for i in result.getheaders() if i[0].lower() == 'content-encoding'] |
There was a problem hiding this comment.
('content-encoding', 'gzip') in python 2, ('Content-Encoding', 'gzip') in python 3. Convert it to lowercase for python2/3 compatibility here.
| return self.__unicode__().encode('ascii', 'replace') | ||
|
|
||
|
|
||
| def __unicode__(self): |
There was a problem hiding this comment.
There is no builtin __unicode__ function in python3. @python_2_unicode_compatible is for python2 compatibility. It has no effect in python3.
But it will encode unicode string with utf8 as returned value of function __str__ in python2, which has the same behavior as before. Link for reference: http://python-future.org/compatible_idioms.html#custom-str-methods
| def test_unicode_error(self): | ||
|
|
||
| result = MockResult() | ||
| result.body = u'A HapiException with unicode \u8131 \xe2\x80\xa2\t' |
There was a problem hiding this comment.
Now this test works with adding u as prefix here.
| # Note the following line is missing the 'u' modifier on the string, | ||
| # this is intentional to simulate poorly formatted input that should | ||
| # still be handled without an exception | ||
| request['data'] = "A HapiException with unicode \\u8131 \xe2\x80\xa2" |
ruichen-wang
deleted the
cmw/RP-60972-make-hapipy-to-be-python3-compatible
branch
Feel free to leave comments for the changes in my branch here.
I have to add a test_credentials.json to make tests run.
And I am only focusing on the clients we are using due to some clients are for legacy APIs which may already get dropped by HubSpot. (Weird errors keep popping up during the test.)
Those unused and potential legacy wrapper clients just get deleted in my branch.
All left tests passed under python 2.7 and 3.7.
Now the portal will be using the last commit of my working branch in CbitLabs. I do not think this branch actually needs to be merged to the master branch, since we may want to save a copy of the original full Hapipy dependency for potential later reference right?