- use the new doOutputBuffer from kdoctools and reduce the number of lower-level functions from kdoctools used by kio;
- unicodeError: force utf-8 as codec; no reason to use a local codec, browsers should be able to render it. Also, remove the last usage of fromUnicode().
Details
Compiles, the code branches previously handled directly and
now managed by doOutputBuffer still works.
Diff Detail
- Repository
- R241 KIO
- Branch
- master
- Lint
No Linters Available - Unit
No Unit Test Coverage
src/ioslaves/help/kio_help.cpp | ||
---|---|---|
136 ↗ | (On Diff #11864) | While at it: this change makes the method name quite strange. Rename to sendError ? |
138 ↗ | (On Diff #11864) | Here the call to data() is not followed by a data(QByteArray()) .... |
344 ↗ | (On Diff #11864) | ... while here the call to data() is followed by data(empty bytearray), as per the kio SlaveBase docu. I suggest making it consistent (the best solution depends on what the other calls to unicodeError() look like) |
src/ioslaves/help/kio_help.cpp | ||
---|---|---|
136 ↗ | (On Diff #11864) | Probably historical reasons from when UTF-8 was not "da" codec, and I'm not sure I want to dig into the history |
138 ↗ | (On Diff #11864) | Right, I simply followed the old behavior, but it's easy to fix. |
344 ↗ | (On Diff #11864) | Other calls are inside get() and the corresponding usage of data does not include data(QByteArray()), so I think I would change emitFile to always add that line at the end. |
For the record, I'm trying to reduce the number of functions used because the next step (which I hope to send for review before 5.32, even if the time is short) is finally exporting a proper public .so from KDocTools.
I consider this an implicit acceptance for the parent review https://phabricator.kde.org/D4813 then :)
Thank you! And lesson learned: change the message here too, as arcanist did not update it.