| shashwatdixit | |
| mwolff | |
| kfunk |
| KDevelop |
Compiler decides for "auto" type result to use QStringBuilder, though the
expression passed to it includes a temporary QString result from a
QStringList::join() result. Which then is dropped after the assignement, due
to getting out of scope.
BUG: 394055
| Automatic diff as part of commit; lint not applicable. |
| Automatic diff as part of commit; unit tests not applicable. |
| plugins/clang/codecompletion/completionhelper.cpp | ||
|---|---|---|
| 81 | Any proposal for a comment to prevent someone "fixing" this into "auto" again? | |
This does fix https://bugs.kde.org/show_bug.cgi?id=394055, doesn't it? If so please reference it in the commit message.
| plugins/clang/codecompletion/completionhelper.cpp | ||
|---|---|---|
| 81 | Not sure if needed, one needs to pay attention to that when using auto combined with concatenating strings after all. At least clang-tidy will not /automatically/ transform this using auto. | |
No idea, will give it a check and if so, add the reference, okay. Thanks for review.
| plugins/clang/codecompletion/completionhelper.cpp | ||
|---|---|---|
| 81 | Many people might not be aware of that trap with QStringBuilder and auto. Or know what QStringBuilder is at all, as it hides away usually. And I was more thinking of "human" fixers who are in for code consistency or modern looking code (/me tries to look innocent) ;) Okay, so being unsure I will leave that comment away for now, too much generalizing from ego perhaps :) | |
| plugins/clang/codecompletion/completionhelper.cpp | ||
|---|---|---|
| 81 | Another possibility might be to keep the auto and force the QString conversion on the right hand side. Slightly more verbose, but protects you from breaking this again. | |