Closed bstaletic closed 9 months ago
@DonKult I would not mind your review as well, if you are up for it.
Merging #1719 (36a9461) into master (0607eed) will increase coverage by
0.03%
. The diff coverage is98.05%
.
@DonKult I would not mind your review as well, if you are up for it.
Sadly, I can't provide much more than emotional support as I know next to nothing about the internals of utf-8 nor use/need/know any scripture that would use the finer details of word splitting. I just remembered seeing an issue about this elsewhere in Debian.
I did test this branch through (with the vim plugin) and it updates unicode, builds & runs (and tests) just fine in the previous as well as last force-push version, so as far that counts as review I can approve of the changes.
(I do apply additional changes through to remove the *.inc files & drop other third-party embeds/downloads in favor of system-provided things, which hide the Download
issue in update_unicode.py
for me for example).
Confirming that things are not broken on your side is good to know.
(I do apply additional changes through to remove the *.inc files & drop other third-party embeds/downloads in favor of system-provided things, which hide the Download issue in update_unicode.py for me for example).
I'm curious. May I see the patches?
(I do apply additional changes through to remove the *.inc files & drop other third-party embeds/downloads in favor of system-provided things, which hide the Download issue in update_unicode.py for me for example).
I'm curious. May I see the patches?
Sure, not a secret, just not usually an upstream favorite topic. I haven't pushed unfuzzied/updated version(s) yet, but you can see them either listed in Debian's Patch tracker for what is currently in the archive or in the VCS of the ycmd packaging bits for possibly unreleased things and as some of them don't make much sense without the vim-plugin context: patches vcs. Some of them I might/should clean up and official propose some day…, there is just always another dumpster fire to deal with first. At least, that is what I claim as an excuse.
Thanks for sending a PR!
Thanks for sending a PR!
This should bring ycmd to support unicode 15.1.
update_unicode.py
.Word.cpp
.What is in this pull request
Brand new stuff
Unicode 15 changed the grapheme cluster boundary rules, by adding GB9c. That thing talks about "indic conjunct break" property, from annex 44. The data itself is DerivedCoreProperties.txt.
That has lead me to adding a new data member to our gigantic
code_points
object.Bug fixes
urllib.request.urlopen
returns a response whoseread()
(and similar) returnbytes
, which breaks our trustyDownload( url )
helper. Yes, this was a very old bug.Playing hide the data
Old link for the unversioned
emoji-data.txt
is not working, so I used the versioned one. I have just found the unversioned one: https://www.unicode.org/Public/UCD/latest/ucd/emoji/emoji-data.txtThis change is![Reviewable](https://reviewable.io/review_button.svg)
Fixes #https://github.com/ycm-core/ycmd/issues/1718