Closed ejulio closed 4 years ago
A meta-question about this PR. https://github.com/scrapy/scrapy/pull/3881 is going to be merged soon (hopefully); it makes quite a serious change to the itemloaders, adding dataclass support. What's the plan of pulling changes to this repo? Should it be done because this PR is merged, or is there a way to do it afterwards (maybe with rebases)?
@kmike probably, the best approach, would be to merge those changes to master
in this repo and then merge this PR.
That could help keeping the history sane and avoid some merges
hey @ejulio! I think docs may need some files around them, to be buildable.
@kmike, regarding the missing docs files, I just copied the text ones. As probably the ones for building are quite standard, so not sure about keeping the history of those. But I can add them as well
@ejulio keeping history is not important for other doc files, but there should be away to build docs (a sphinx project); docs should be built by readthedocs.org eventually.
@kmike I'm working on it in another PR https://github.com/ejulio/itemloaders/pull/4
:exclamation: No coverage uploaded for pull request base (
master@81f54fb
). Click here to learn what that means. The diff coverage isn/a
.
@@ Coverage Diff @@
## master #1 +/- ##
=========================================
Coverage ? 98.00%
=========================================
Files ? 4
Lines ? 251
Branches ? 0
=========================================
Hits ? 246
Misses ? 5
Partials ? 0
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 81f54fb...81f54fb. Read the comment docs.
Hey @ejulio! The PR looks good to me; I think we can update master to match Scrapy, merge master to this branch (ensuring that changes are preserved), and then merge this branch to master.
Draft while removing dependencies from scrapy