Uh oh!
There was an error while loading. Please reload this page.
Moving gcloud.bigtable.happybase into third_party/ - #1777
Conversation
theacodes
commented
May 5, 2016
The only alternative I can think of is symlinking, and I'm not sure how I feel about that. |
theacodes
commented
May 5, 2016
Another alternative is just making this a separate, optional package |
theacodes
commented
May 5, 2016
dhermes
commented
May 6, 2016
Thanks for taking this on. Both test failures seem to indicate that the imports fail. I'm also curious about the contents of the |
theacodes
commented
May 6, 2016
Yeah this is weird. I I'm starting to think that moving this into a separate library is the best way to go. |
dhermes
commented
May 6, 2016
I'm fine with |
tseaver
commented
May 6, 2016
FWIW, the code this PR moves into I'm not sure that moving it does anything to help the legalities, compared to just having the header comments mention the separate license (which could just be tacked onto the current |
theacodes
commented
May 6, 2016
This is what our open-source laywers have told us to do. It's best to On Fri, May 6, 2016, 4:09 PM Tres Seaver notifications@github.com wrote:
|
dhermes
commented
May 9, 2016
@jonparrott Why the close? |
theacodes
commented
May 9, 2016
Because this approach won't work. |
dhermes
commented
May 9, 2016
OK |
* chore: use faster query_and_wait API in _read_gbq_colab * try to fix unit tests * more unit test fixes * more test fixes * fix mypy * fix metrics counter in read_gbq with allow_large_results=False * use managedarrowtable * Update bigframes/session/loader.py * split out a few special case return values for read_gbq_query * support slice node for repr * fix failing system test * move slice into semiexecutor and out of readlocalnode * unit test for local executor * split method instead of using reloads * fix reference to _start_query * use limit rewrite for slice support * do not use numpy for offsets
Towards #1762
This technically works, but some things to note:
python setup.py developand by extensionpip install -edo not respect thepackage_dirargument tosetup, as such,gcloud.bigtable.happybasewill not be available if it's installed that way.Don't merge until we get confirmation from OSPO that this is the correct thing to do.