[prev in list] [next in list] [prev in thread] [next in thread]
List: kde-core-devel
Subject: Re: Review Request: Speedup KIconEngine::actualSize by not loading the
From: "Albert Astals Cid" <aacid () kde ! org>
Date: 2010-05-19 18:06:19
Message-ID: 20100519180619.4059.73062 () localhost
[Download RAW message or body]
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviewboard.kde.org/r/4052/#review5745
-----------------------------------------------------------
Without having a look at the code i think it is somewhat a big deal actualSize() not \
returning the actual size of the icon, even if it is in some corner cases, can we fix \
the slowlyness this by adding a fastButNotReallActualSize() method?
- Albert
On 2010-05-19 13:26:30, David Faure wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://reviewboard.kde.org/r/4052/
> -----------------------------------------------------------
>
> (Updated 2010-05-19 13:26:30)
>
>
> Review request for kdelibs.
>
>
> Summary
> -------
>
> KIconEngine::actualSize was loading the icon (from the cache, fortunately, but that \
> still takes time), just to return its size. However the size is mostly mandated by \
> the caller of KIconLoader::loadIcon...
> Looking at the (complex) KIconLoader code I think there are indeed a few cases \
> where the icon wouldn't have the requested size, but the question is whether it's a \
> big deal for the users of actualSize()... In the case of file manager views, the \
> user is KFileItemDelegate::Private::decorationSizeHint which is surely happier to \
> reserve the same space for every icon, even if a few ones are only available as \
> somewhat different sizes. I just wonder 1) how to test the case where we get a \
> different size than the requested one, and 2) if this faster (but only correct 99% \
> of the time, I suppose) actualSize() would break anything.
>
> This addresses bug 237668.
> https://bugs.kde.org/show_bug.cgi?id=237668
>
>
> Diffs
> -----
>
> trunk/KDE/kdelibs/kdeui/icons/kiconengine.cpp 1128493
>
> Diff: http://reviewboard.kde.org/r/4052/diff
>
>
> Testing
> -------
>
> konqueror /usr/bin
> (after the kio fix r1128475)
> This is much faster now than before. And to make sure actualSize() wasn't returning \
> wrong results, I added asserts in pixmap() and paint().
> Plus existing unittests and adding a rudimentary unit-test for actualSize.
>
>
> Thanks,
>
> David
>
>
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic