http://gambasrad.org Showing the last 30 entries from a total of 188 Poi Issue Tracker http://gambasrad.org/bugtrak/gambas-bug-tracker/33/1 Ok, this is becoming a real problem .. for some reason Ubuntu Gutsy restarts pulseaudio periodically, even after it's been killed .. I've a sneaking suspicion they mean to make pulseaudio an integral part of Ubuntu .. http://gambasrad.org/bugtrak/gambas-bug-tracker/36/6 $ sdl-config --version 1.2.11 sdl-mixer is 1.2. Just sound .. this is a standard Ubuntu gutsy install .. # pulseaudio --version pulseaudio 0.9.6 ... don't look at me, I don't even know what SDL stand for (!) http://gambasrad.org/bugtrak/gambas-bug-tracker/36/5 Do you have any other SDL related code in your project? Which version of SDL and SDL_mixer do you use? Which version of pulseaudio? According to the SDL source, the only possible arithmetic crash in Mix_FadeInMusicPos() is a division by zero that occurs if a global variable is zero. This global variable is initialized at SDL mixer initialization, i.e. when the component is loaded. This global variable can be zero only if the internal sound buffer is zero bytes, which should not be possible. So I have no idea... http://gambasrad.org/bugtrak/gambas-bug-tracker/36/4 Hi, the connection is that if I kill pulseaudio, the program doesn't crash (!) Here's my code, but complex I'm afraid .. ;-) WAIT Music.Load("/usr/local/share/sounds/raven/startup.ogg") Music.Play() WAIT http://gambasrad.org/bugtrak/gambas-bug-tracker/36/3 I don't see the relation with pulseaudio at all. But I see that there is a crash inside the SDL_mixer library. How do you call Music.Play? Can you send me your source code? http://gambasrad.org/bugtrak/gambas-bug-tracker/36/2 Ok, and looking at this I've found the problem. Somehow (!) pulseaudio has started itself up in the background (silently) which is known to cause a problem with sound .. a signal #6 however is new (!) pulseaudio comes as standard on the latest version of Ubuntu, I don't know how to disable it apart from killing it .. which seems to have no effect on anything else. Anyway .. it might be an idea to catch this as I'm guessing I'm not the only one running Ubuntu Gutsy ... http://gambasrad.org/bugtrak/gambas-bug-tracker/36/1 Here's the stack dump; Program terminated with signal 8, Arithmetic exception. #0 0xb628a7c9 in Mix_FadeInMusicPos () from /usr/lib/libSDL_mixer-1.2.so.0 (gdb) bt #0 0xb628a7c9 in Mix_FadeInMusicPos () from /usr/lib/libSDL_mixer-1.2.so.0 #1 0xb628a8fb in Mix_FadeInMusic () from /usr/lib/libSDL_mixer-1.2.so.0 #2 0xb639c87f in CMUSIC_play (_object=0x0, _param=0x808a958) at sound.c:461 #3 0x0804e3f8 in EXEC_call_native (exec=0xb639c7d9 , object=0x0, type=0, param=0x808a958) at gbx_exec.c:837 #4 0x0804ec61 in EXEC_native () at gbx_exec.c:955 #5 0x080512a4 in EXEC_loop () at gbx_exec_loop.c:1042 #6 0x0804efdc in EXEC_function_real (keep_ret_value=1 '\001') at gbx_exec.c:695 #7 0x0805da86 in call_method (object=0x8163960, desc=0x0, nparam=-1238641504) at gbx_api.c:409 #8 0x0805e0b9 in GB_Raise (object=0x8163960, event_id=19, nparam=0) at gbx_api.c:579 #9 0xb6555b77 in post_show_event (_object=0x8163960) at CWindow.cpp:2333 #10 0xb6557f26 in CWindow::eventFilter (this=0xb65bb580, o=0x8163c30, e=0xbfe35614) at CWindow.cpp:2401 #11 0xb7a94e40 in QObject::activate_filters () from /usr/lib/libqt-mt.so.3 #12 0xb7a94ebe in QObject::event () from /usr/lib/libqt-mt.so.3 #13 0xb7acc5b3 in QWidget::event () from /usr/lib/libqt-mt.so.3 #14 0xb7ba2c5a in QMainWindow::event () from /usr/lib/libqt-mt.so.3 #15 0xb7a2caf0 in QApplication::internalNotify () from /usr/lib/libqt-mt.so.3 #16 0xb7a2f4ba in QApplication::notify () from /usr/lib/libqt-mt.so.3 #17 0xb654e5fe in MyApplication::notify (this=0x810fdd0, o=0x8163c30, e=0xbfe35614) at main.cpp:336 #18 0xb79bf209 in QApplication::sendEvent () from /usr/lib/libqt-mt.so.3 #19 0xb7acf173 in QWidget::show () from /usr/lib/libqt-mt.so.3 #20 0xb7ba2bb6 in QMainWindow::show () from /usr/lib/libqt-mt.so.3 #21 0xb6557a2c in MyMainWindow::showActivate (this=0x8163c30) at CWindow.cpp:1485 #22 0xb6557ba4 in CFORM_main (_object=0x0, _param=0x808a948) at CWindow.cpp:376 #23 0x0804e3f8 in EXEC_call_native (exec=0xb6557b70 , object=0x0, type=0, param=0x808a948) at gbx_exec.c:837 #24 0x0804ec61 in EXEC_native () at gbx_exec.c:955 #25 0x08061e81 in main (argc=0, argv=0xbfe358d4) at gbx.c:345 http://gambasrad.org/bugtrak/gambas-bug-tracker/36 http://gambasrad.org/bugtrak/gambas-bug-tracker/35/5 Ok, it looks like it's fixed itself (!) I've no idea what I did .. however it does now seem to be working. Absolutely no idea why ... :( Sorry. http://gambasrad.org/bugtrak/gambas-bug-tracker/35/4 Oups... the more bad it's i've look at this algorithm... yesteday... Sometime it will be better to me to say nothing :) http://gambasrad.org/bugtrak/gambas-bug-tracker/35/3 The IDE manages the components installed in the user home directory. Actually this is the interpreter that manages that. When trying to load a component named gb.XXX, the rule is the following: 1) Look in /usr/lib/gambas2 (depends on the installation prefix) for gb.XXX.so, then for gb.XXX.gambas. 2) If nothing is found, look in ~/.local/lib/gambas2 for the same files. 3) If nothing is really found, raise an error. The IDE implements the same algorithm to locate the components then display the component list. http://gambasrad.org/bugtrak/gambas-bug-tracker/35/1 I didn't think that ... in other case all the gambas components such as gb.form, gb.form.mdi will not run... Where are you storing your components for running it ? in /usr/local/lib/gambas2/? or in gambas users component ? in the second case ... if i remember well the ide didn't manage it yet... Fabien http://gambasrad.org/bugtrak/gambas-bug-tracker/35/2 I have an old version of GridEditor on my machine, and if I compile it and install it as a component in the user directory, everything seems to work correctly. So please: * Send me the very last GridEditor you tried to install. * Check that the symbolic links are correctly created inside ~/.local/share/gambas2/info and ~/.local/lib/gambas2. http://gambasrad.org/bugtrak/gambas-bug-tracker/35 http://gambasrad.org/bugtrak/gambas-bug-tracker/19/3 sorry for being idle i've solved this problem using the ebuild from bugs.gentoo.org i've attached it http://gambasrad.org/bugtrak/gambas-bug-tracker/28/9 Changing the database schema from mediumblob to longblob fixed the problem for the camera viewer application. This is fixed for me, but I still think a "mediumblob" should be a "blob" and a "text" should be text ... http://gambasrad.org/bugtrak/gambas-bug-tracker/28/8 I'm still not sure why you can't map "tinytext" and "text" to a text field and "mediumblob" and "longblob" to a blob /raw data field .. http://gambasrad.org/bugtrak/gambas-bug-tracker/28/7 Mmm, well I actually have customers using the camera code, and it obviously works .. yet I can't view images from their databases as it falls over as described .. so something has changed (!) In MySQL, all non integers / chars (IMHO) are / should be treated as blobs. If anyone wants to extract text from a "blob" it should be via an ".astext()" type method, rather than reading an actual block of text. I'll change a couple of the crucial DB's here and we'll see if text->blob makes fixes the problem and if any of the other apps fall over .. http://gambasrad.org/bugtrak/gambas-bug-tracker/28/6 They are variants. That does not change the problem of which Gambas datatype should be associated to a given MySQL datatype. If MySQL makes no difference between Blob and Text, this is not my fault. All other database backends I know do. But I have to make the difference, as storing strings (Text) or binary data (Blob) is not done the same way in Gambas. And I looked in the svn history: the behaviour of Blob fields in gb.db.mysql did not change since its first implementation. So I don't know what I could have broke about that in the development version... http://gambasrad.org/bugtrak/gambas-bug-tracker/28/5 Ok, so you are trying to create a difference where in fact, no difference exists (?!) There are two issues with this that affect myself in particular; a. I will need to go through and change every table I use which uses text / blob types .. this may or may not break other applications that use those tables. b. You have changed the behavior of Gambas which breaks pre-existing applications. I did see the big red label saying "Database stuff will change" .. but making changes with no compatibility or upgrade path is a real killer for people who are actually "using" Gambas. I know technically I should only be "using" version #1, but #1 is Sooooo far behind, I can't. Why not simply store Text/Blob types as a variant and let the user decide .. which is what ( I think ? ) other languages do ... ?? http://gambasrad.org/bugtrak/gambas-bug-tracker/28/4 The MySQL field datatype must be "longblob", otherwise the database component assumes it is a text field. You must understand that there is no difference between blobs and text fields in MySQL. So I had to make a choice so that I can make the difference: - Tiny and medium blob/text got the String datatype. - Long blob/text got the Blob datatype. To be sure, I need the value of resultset.Fields["image"].Type. Anyway you should have got some data in the text field. But I'm not sure too, as if the data starts with a zero character, you should get a null string. http://gambasrad.org/bugtrak/gambas-bug-tracker/23/7 Which leaved scope for a GridEditor based MySQL administrator type tool. I can live with that .. ;-) http://gambasrad.org/bugtrak/gambas-bug-tracker/31/3 Nope it's a problem with the web server. Telnet on port 80 works, however Firefox and Konquerer both hand with "Waiting for reply .. " .. almost looks like the server is trying to run some sort of pipelining without terminating each transaction properly. http://gambasrad.org/bugtrak/gambas-bug-tracker/23/6 This is not a limitation, this is a feature! The database manager and the database component is not intended to manipulate MySQL databases specifically. So if there is a feature that does not exist in all database drivers (like the autoincrement on integer fields smaller than 8 bytes), then this is the maximum I can do. http://gambasrad.org/bugtrak/gambas-bug-tracker/23/5 Mmm, a warning would be good. Removing the limitation would be better (!) http://gambasrad.org/bugtrak/gambas-bug-tracker/23/4 OK, I will modify the mysql driver so that integer fields with auto_increment flag are considered as serial fields. But writing such tables in the database manager will make these fields BIGINT. Anyway, never use the database manager to modify the structure of tables not created with it! Maybe I should add a warning about that... http://gambasrad.org/bugtrak/gambas-bug-tracker/31/2 Hmm... are you sure? I can open it from my place. Maybe there's a problem with your ISP. Can you try it again using a proxy or something like that? Here, I attach a screen shoot of the site http://gambasrad.org/bugtrak/gambas-bug-tracker/31/1 Sorry, can't actually open the link you've provided. (tried a few times) If you want to try again with a working link or paste in some text describing what you are looking for ? http://gambasrad.org/bugtrak/gambas-bug-tracker/30/1 We get around this by using two screens and opening the project twice, once on each screen. I'm guessing this is not recommended practice, however it does provide the facility you're looking for. It might be quite a nice feature to have it inside the IDE, however I'm guessing there's quite a lot of heavy duty changes that are going to take precedence over this ... http://gambasrad.org/bugtrak/gambas-bug-tracker/23/3 I've just had someone else complain he's screwed his databases by using the Database Manager. If we could implement this one minor change it could save people much personal pain, AND make the Database Manager safe to use ... (!)