Review Request: [PATCH] Make KIO::Scheduler correctly re-use ioslaves that have been put on hold...

[Andreas Hartmetz]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://svn.reviewboard.kde.org/r/6271/#review9510
-----------------------------------------------------------



/trunk/KDE/kdelibs/kio/kio/scheduler.cpp
<http://svn.reviewboard.kde.org/r/6271/#comment10511>

    (see below)
    if (cmd == CMD_GET)
        return true;
    return cmd == CMD_SPECIAL && protocol.startsWith(...)



/trunk/KDE/kdelibs/kio/kio/scheduler.cpp
<http://svn.reviewboard.kde.org/r/6271/#comment10510>

    trailing whitespace; there's another one below



/trunk/KDE/kdelibs/kio/kio/scheduler.cpp
<http://svn.reviewboard.kde.org/r/6271/#comment10509>

    Better move everything into a mayReturnContent(KIO::Job *) method and avoid the long function name.



/trunk/KDE/kdelibs/kio/kio/scheduler.cpp
<http://svn.reviewboard.kde.org/r/6271/#comment10513>

    This is the core of the patch, right...
    I don't know if the slave-on-hold mechanism was meant to be used between KParts in the same app or only between applications. In the latter case more changes in KIO will probably be necessary. In the former case, I screwed up something around this line in the scheduler rewrite.



/trunk/KDE/kdelibs/kio/kio/scheduler.cpp
<http://svn.reviewboard.kde.org/r/6271/#comment10512>

    coding style: braces even for one-liners


- Andreas


On 2011-01-04 09:55:29, Dawit Alemayehu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://svn.reviewboard.kde.org/r/6271/
> -----------------------------------------------------------
> 
> (Updated 2011-01-04 09:55:29)
> 
> 
> Review request for kdelibs.
> 
> 
> Summary
> -------
> 
> The patch attempts to fix the following two oustanding problems in KIO::Scheduler that have been around for a while (see the possibly related bug reports in the Bugs section above):
> 
> #1. Set the m_checkOnHold flag to true every time Scheduler::publishSlaveOnHold is invoked. Right now that flag is only set to true when an instance of KIO::Scheduler is created. This results in the flag never being true after the first ioslave has been put on hold and reused unless the programmer explicitly calls KIO::Scheduler::checkSlaveOnHold which is not documented at all. See the description about putting ioslaves on hold in KIO::get's API documentation.
> 
> #2. Modify SchedulerPrivate::doJob to correctly set the m_checkOnHold flag for http requests when a job's command is CMD_SPECIAL.  That is necessary because HTTP_POST, which is handled as a special command, can return content just like a get request.
> 
> 
> This addresses bugs 123121 and 148307.
>     https://bugs.kde.org/show_bug.cgi?id=123121
>     https://bugs.kde.org/show_bug.cgi?id=148307
> 
> 
> Diffs
> -----
> 
>   /trunk/KDE/kdelibs/kio/kio/scheduler.cpp 1211439 
> 
> Diff: http://svn.reviewboard.kde.org/r/6271/diff
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Dawit
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-core-devel/attachments/20110104/887891d5/attachment.htm>