D25039: Fix Clazy performance issues, const &, noexcept
Friedrich W. H. Kossebau
noreply at phabricator.kde.org
Wed Oct 30 18:43:46 GMT 2019
kossebau added a comment.
(Quick drive-by comment only as I had this in a search result...)
INLINE COMMENTS
> slavebase.h:351
> */
> - bool configValue(QString key, bool defaultValue) const;
> + bool configValue(const QString &key, bool defaultValue) const;
>
Would be a proper change, but sadly also changes the signature of a API under ABI promise, so here and in all other public API places not possibe.
Can be only done when breaking the ABI with KF6.
Should there be something else done instead? IMHO no, neither adding a TODO or even already an overload of the method with the preferred signature.
For one, checking and improving all API of KF6 with clazy should be expected to be done before any first release, so this should be catched at that time. At the same time, this change is source-compatible, client-side code does not have to be changed, so there is no real gain in client code readability, and the runtime price tag of a flat QString copy here is equally small compared to having another symbol and implementation code, so less code is better here.
See also https://community.kde.org/Policies/Binary_Compatibility_Issues_With_C%2B%2B#The_Do.27s_and_Don.27ts
REPOSITORY
R241 KIO
REVISION DETAIL
https://phabricator.kde.org/D25039
To: meven, #frameworks, dfaure
Cc: kossebau, kde-frameworks-devel, LeGast00n, GB_2, michaelh, ngraham, bruns
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-frameworks-devel/attachments/20191030/2da6282e/attachment.html>
More information about the Kde-frameworks-devel
mailing list