Review Request 128514: KIconLoader: reduce number of lookups when doing fallbacks

[Christoph Feck]

> On July 25, 2016, 11:46 a.m., Christoph Feck wrote:
> > src/kiconloader.cpp, line 1037
> > <https://git.reviewboard.kde.org/r/128514/diff/1/?file=472255#file472255line1037>
> >
> >     Could you please also move the preceeding debug comment?

If you want to keep the debug comment at the beginning of the loop, could you please also add a debug comment before actually calling "iconPathByName"? The currentName is modified inbetween, and the debug code was there to see which icon name is _actually_ being looked up.


- Christoph


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/128514/#review97827
-----------------------------------------------------------


On July 30, 2016, 2:08 p.m., David Faure wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://git.reviewboard.kde.org/r/128514/
> -----------------------------------------------------------
> 
> (Updated July 30, 2016, 2:08 p.m.)
> 
> 
> Review request for KDE Frameworks and Christoph Feck.
> 
> 
> Repository: kiconthemes
> 
> 
> Description
> -------
> 
> In the first pass, we already checked whether the icon exists in all icon themes.
> So in the second pass, there's no point in starting with that again, we can go
> ahead and start doing fallbacks right away.
> 
> (The comment about the two passes was removed in b84858c, but some of it still
> applies.)
> 
> Tested with: strace -e file ./kiconengine_unittest testUnknownIconNotCached |& wc -l
> Before: 6341
> After: 4589
> 
> 
> Diffs
> -----
> 
>   autotests/kiconloader_benchmark.cpp ded5e0221f8126139a019a8310e8f120672ac3b4 
>   src/kiconloader.cpp 951775d77b7ec0e2d8099ee192ca25321dab4544 
> 
> Diff: https://git.reviewboard.kde.org/r/128514/diff/
> 
> 
> Testing
> -------
> 
> strace as mentionned above.
> 
> All unittests still pass (incl fallback lookups, didn't check how complete those were, though)
> 
> 
> Thanks,
> 
> David Faure
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-frameworks-devel/attachments/20160731/54b8e8b3/attachment.html>