mirror of
https://we.phorge.it/source/arcanist.git
synced 2025-02-16 16:58:38 +01:00
1fcf7bac4e
Summary:
@Afaque_Hussain has done a bunch of utf8 work here; combined with PhutilEditDistanceMatrix we can now do utf8 diffs correctly, in a general way, without a significant performance impact.
Use PhutilEditDistanceMatrix and `phutil_utf8v_combined()` to compute accurate diffs for all (or, at least, most) UTF8 text.
The only thing this doesn't handle completely correctly is lines beginning with combining characters. This is messy/expensive to handle and will probably never actually happen, so I'm punting for now. Nothing should actually break.
The utf8 stuff will be slow, but we only pay for it when we need it.
Test Plan:
Ran unit tests. I changed a few unit tests to use a non-combining character (snowman) for clarity, and some results are different now (since we get combining characters right).
{F44064}
Reviewers: btrahan, Afaque_Hussain
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2379
Differential Revision: https://secure.phabricator.com/D6019
|
||
|---|---|---|
| .. | ||
| configuration | ||
| difference | ||
| differential | ||
| docs | ||
| events/constant | ||
| exception | ||
| extensions | ||
| hgdaemon | ||
| infrastructure/testing | ||
| lint | ||
| parser | ||
| repository | ||
| unit | ||
| workflow | ||
| workingcopyidentity | ||
| __phutil_library_init__.php | ||
| __phutil_library_map__.php | ||