[kdelibs/KDE/4.10] /: New branch for the future kdelibs-4.10. Setting version number to 4.9.90 for now.
Git commit fed5ee75c021cf3052f1461acc26dcf1bbd0c7b2 by David Faure. Committed on 16/08/2012 at 11:01. Pushed by dfaure into branch 'KDE/4.10'. New branch for the future kdelibs-4.10. Setting version number to 4.9.90 for now. CCMAIL: release-t...@kde.org, kde-core-devel@kde.org Please fix bugs in 4.9 branch, I'll regularly git merge into 4.10, master and frameworks. M +1-1CMakeLists.txt http://commits.kde.org/kdelibs/fed5ee75c021cf3052f1461acc26dcf1bbd0c7b2 diff --git a/CMakeLists.txt b/CMakeLists.txt index fdc39c3..23af568 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -10,7 +10,7 @@ include(CPack) set (KDE_VERSION_MAJOR 4) set (KDE_VERSION_MINOR 9) -set (KDE_VERSION_RELEASE 00) +set (KDE_VERSION_RELEASE 90) set (KDE_VERSION ${KDE_VERSION_MAJOR}.${KDE_VERSION_MINOR}.${KDE_VERSION_RELEASE} ) set (KDE_VERSION_STRING ${KDE_VERSION})
Re: Review Request: Show warning when CopyJob fails to list a subdir
--- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/106052/#review17528 --- I think the right thing to do is to fix bug 206500, and then apply this patch, rather than come up with yet another feedback mechanism. Let's make warning() actually work, and let's make sure it's not emitted erroneously. I'll look into fixing bug 206500 as soon as possible (let's say beginning of next week). kdeui/jobs/kdialogjobuidelegate.cpp http://git.reviewboard.kde.org/r/106052/#comment13720 No space after the method name (this usually happens when copy/pasting from the Qt documentation :-) In fact, no space within the parenthesis either, but you couldn't guess that: existing code doesn't follow the kdelibs coding style. When writing new code, no spaces within (...). kio/kio/job.cpp http://git.reviewboard.kde.org/r/106052/#comment13721 Left - Let. How did this come in? It wasn't in my initial patch :) - David Faure On Aug. 16, 2012, 12:18 p.m., Dan Vratil wrote: --- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/106052/ --- (Updated Aug. 16, 2012, 12:18 p.m.) Review request for kdelibs and David Faure. Description --- Display a warning when CopyJob fails to enter a subdirectory and thus can't copy it's content. Diffs - kdeui/jobs/kdialogjobuidelegate.cpp fe48f87 kio/kio/copyjob.h eb88c7a kio/kio/copyjob.cpp 8dde763 kio/kio/job.cpp a7e1baf kio/kio/jobclasses.h de27f40 Diff: http://git.reviewboard.kde.org/r/106052/diff/ Testing --- Thanks, Dan Vratil
Re: Review Request: Show warning when CopyJob fails to list a subdir
On Aug. 16, 2012, 1:10 p.m., Frank Reininghaus wrote: kdeui/jobs/kdialogjobuidelegate.cpp, line 92 http://git.reviewboard.kde.org/r/106052/diff/1/?file=78076#file78076line92 I'm afraid the users suffering from https://bugs.kde.org/show_bug.cgi?id=206500 will kill us if they get a message box for every single file. Right now, they have the option to wait until the operation is completed, then close the first (and only) message box and be happy. Ambroz Bizjak wrote: How about aggregating all the errors/warnings in a single dialog box as a list? E.g. the list would have fields like message and file, and would allow the user to easily see all that went wrong and what files were involved. He could then select multiple entries and perform the same action on them, if applicable. E.g. if the message was destination already exists, replace or skip? he could select some of the entries and perform ignore, but perform replace on others. The copying could continue in the background, and the replacements would be performed only once the user confirms then. Dan Vratil wrote: But you can get multiple kinds of warnings during the operation. How would you separate them? You would either have to have a separate dialog for each kind of error or a very complex UI dialog. In general I like the idea of a box These files failed to copy (similar to the Delete Files dialog), but I think that simple 'OK' would be enough. Ideas? Yes, that's a good idea. When a file fails to copy, pop up the These files failed to copy dialog, but when another file fails, just add it to the list in the existing dialog. User can click OK which closes the dialog, and if more files fail, another one pops up. Small problem though: what happens if the user clicks OK, without noticing that another file was added to the list just before he closed the dialog? Maybe a safeguard should be added that no less than N seconds may pass from the time the last file was added to when the user clicks OK. - Ambroz --- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/106052/#review17531 --- On Aug. 16, 2012, 12:18 p.m., Dan Vratil wrote: --- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/106052/ --- (Updated Aug. 16, 2012, 12:18 p.m.) Review request for kdelibs and David Faure. Description --- Display a warning when CopyJob fails to enter a subdirectory and thus can't copy it's content. Diffs - kdeui/jobs/kdialogjobuidelegate.cpp fe48f87 kio/kio/copyjob.h eb88c7a kio/kio/copyjob.cpp 8dde763 kio/kio/job.cpp a7e1baf kio/kio/jobclasses.h de27f40 Diff: http://git.reviewboard.kde.org/r/106052/diff/ Testing --- Thanks, Dan Vratil
Re: KBzip2Filter::uncompress outputs qDebug messages
On Wednesday 15 August 2012 17:17:32 Michael Pyne wrote: On Wednesday, August 15, 2012 16:49:13 Milian Wolff wrote: The output is always: bzDecompress returned 4 KBzip2Filter::uncompress 1 Could someone shed some light on this? Is this a serious issue? If not, can we silence the output please? It is just libbz2's way of signalling that it has reached the end of the stream. I certainly don't see any issue with squelching that (and it shouldn't be using qDebug() anyways I would imagine). OK, cool. I've created a review request for that now. Cheers. -- Milian Wolff m...@milianw.de http://milianw.de signature.asc Description: This is a digitally signed message part.
Re: Call for Projects: University Collaboration in Toulouse
Hello, On Donnerstag, 16. August 2012 12:12:12 Kevin Ottens wrote: Hello, As a reminder: I need to close the list of projects real soon now. If you did send me your proposal yet please do so NOW! On behalf of the Marble Team I'd like to express our highest interest in participation! The two mentors we'd provide are Dennis and me. We have a very interesting topic going on that has a lot to offer for students with different aspects of interest and varying capabilities: OpenStreetMap Vector Rendering We had Ander successfully working on a prototype and first implementation related to this topic during GSoC 2012: You can read all about it in his blog: http://blogs.deusto.es/gsoc-deustotech/ But there's still a lot of work left with different aspects/features that could be done either isolated on a per-student basis or by the full team. Our OpenStreetMap Vector Rendering Feature Topics include: - Street Label placement for Marble OpenStreetMap vector rendering (very important aspect and has lots of render-, algorithm-, optimization- and usability-issues involved that one could work on): Currently Marble doesn't render street names in .OSM vector rendering. Of course street names should always be rendered in a visible place along the streets and shouldn't overlap - oh and they should always render quickly. - Development of a dedicated file format for storing OpenStreetMap vector tiles. Right now we are using some JS-based file format. See e.g. http://blogs.deusto.es/gsoc-deustotech/first-vector-tiles-rendering/ - Related to the former topic: Setting up a vector tile server and setting up / developing the software for it for Marble - PBF File Format support for Marble. This would be the efficient binary incarnation of the .OSM file format . - Improvements to our vector rendering pipeline (GeoGraphicsView). This is more an architecture-refactoring and optimization topic - Vector Map Styling support: Improvement of the current styling mechanism for streets and points of interests + possible evaluation of usage of mapcss + its implementation in Marble Of course other topics for Marble work are welcome as well. Possible alternatives could cover Space + Astronomy, Historic Maps, Routing, GeoCaching, Natural Earth vector tile rendering (topographic map), Time Animation KML enhancements, etc. But we are also interested in other subjects students could come up with. You can get in touch with us here: http://edu.kde.org/marble/support.php Compiling Marble is easy: http://edu.kde.org/marble/obtain.php and the same applies for getting started with Marble development: http://techbase.kde.org/Projects/Marble/MarbleCPlusPlus Best Regards, Torsten I'm especially looking at the ones who told me in private they wanted to participate but sent nothing. You know who you are. ;-) Of course if you didn't contact me yet, you still can send me a proposal, I'll gladly add it to the list. Regards. On Tuesday 07 August 2012 19:08:18 Kevin Ottens wrote: Hello all, I have very good news, if everything goes well, I will have the opportunity to setup student projects at the University in collaboration with KDE again! We had a one year hiatus, and I have to admit I miss those projects dearly, so it'd be nice to resume the effort... I'd quite some work ahead as its a reboot and not much of my previous efforts is left. Anyway, if you are part of a KDE team, and you'd like to get some help from a small team of students (not more than four or five) for a few months, here is your chance! The requirements on your side if you volunteer are the following: * Two mentors from your team (one will kind of play the customer to the students, the other one more on the technical mentoring side); * A list of features the students would have to work on (not more than a dozen, not less than five). If you are interested, please contact me in private with the info mentionned above (mentors + feature list with some details). Don't wait! I'll finalize the list of chosen projects end of next week (and then push it to the university), and I might have a few questions for you before picking your proposal or not. Regards.
Re: Review LibKdeAccessibilityClient
On Thursday 16 August 2012, Frederik Gladhorn wrote: Hi, a few of us trying to improve accessibility of KDE worked on a little library, LibKdeAccessibilityClient. What it does is making the AT-SPI over DBus protocol easy to use for KDE apps. This is the accessibility framework that we share with Gnome. It enables two way communication between apps and assistive tools. The first users for it would be the magnifiers in KWin and KMag - they can follow the focus with it. Further users will be screen readers and Simon with the AT-SPI plugin to dynamically create vocabulary for the running application's actions. I'd like to ask for initial review in order to move this tiny lib to kdesupport/frameworks. It's Qt-only and actually only consists of two LibKdeAccessibilityClientConfig.cmake.in: Line 9: why is there the if(NOT LIBKDEACCESSIBILITYCLIENT_INSTALL_DIR) set(LIBKDEACCESSIBILITYCLIENT_INSTALL_DIR @CMAKE_INSTALL_PREFIX@) endif() condition ? This will be evaluated at usage-time. At cmake-time you already should know whether LIBKDEACCESSIBILITYCLIENT_INSTALL_DIR is set or not and set it to CMAKE_INSTALL_PREFIX if it is not set, but at cmake-time of the library, not at cmake-time of the using project. Line 13: set(LIBKDEACCESSIBILITYCLIENT_INCLUDE_DIR @LIBKDEACCESSIBILITYCLIENT_INCLUDE_DIR@) This hardcodes the include dir, which will be problematic under Windows, but ok. Line 15: this one should be changed, it hardcodes a lot of stuff: et(LIBKDEACCESSIBILITYCLIENT_LIBS @LIBKDEACCESSIBILITYCLIENT_LIB_DIR@/libkdeaccessibilityclient.so) Please use the target-exporting feature of cmake install(TARGETS kdeaccessibilityclient ... EXPORT MyExports ) install(EXPORTS MyExports ... ) and then install this export into the same directory as the Config.cmake file and include() it in the Config.cmake file. CMakeLists.txt: Line 29: use find_poackage(PkgConfig) instead of include(FindPkgConfig) Line 30: you rely completely on pkg_config to find atspi-2. I would prefer if you could add a simple FindATSPI2.cmake, which first uses pkg_config and then uses its results as HINTS for the calls to find_path() and find_library() (as e.g. done in FindLibXml2.cmake). Alex
Re: Review Request: Silence debug output in KBzip2Filter::uncompress
--- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/106056/#review17566 --- Ship it! I like it, and it's even easier than what my fix was going to be. - Michael Pyne On Aug. 16, 2012, 6:16 p.m., Milian Wolff wrote: --- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/106056/ --- (Updated Aug. 16, 2012, 6:16 p.m.) Review request for kdelibs and Michael Pyne. Description --- Only output warning message if the return value of bzDecompress is negative, which represents an error. See e.g. bzlib.h: #define BZ_OK0 #define BZ_RUN_OK1 #define BZ_FLUSH_OK 2 #define BZ_FINISH_OK 3 #define BZ_STREAM_END4 #define BZ_SEQUENCE_ERROR(-1) #define BZ_PARAM_ERROR (-2) #define BZ_MEM_ERROR (-3) #define BZ_DATA_ERROR(-4) #define BZ_DATA_ERROR_MAGIC (-5) #define BZ_IO_ERROR (-6) #define BZ_UNEXPECTED_EOF(-7) #define BZ_OUTBUFF_FULL (-8) #define BZ_CONFIG_ERROR (-9) As such, all negative values (or anything below BZ_OK) represents an error. Without this patch, using the KBzip2Filter would always spam the console with the following debug messages: bzDecompress returned 4 KBzip2Filter::uncompress 1 Diffs - kdecore/compression/kbzip2filter.cpp 0f55334 Diff: http://git.reviewboard.kde.org/r/106056/diff/ Testing --- Thanks, Milian Wolff