Re: Trapping application "crashes"?
Hi, ext Jayesh Salvi wrote: > If you can modify application's code (at least the main() entry point) you > can register signal handlers for SIGSEGV, SIGABRT (most common causes of > program crashes) etc. Do "man signal" for more info. If the program is in > other languages then there should be corresponding ways to register these > signal handlers. > > If you don't have access to source code then only way that I can suggest is > running from debugger. > > If you end up registering your signal handlers - do only little in those > signal handlers. It's not a good idea to keep executing once you have > corrupted memory - you might run into more ugly errors. In general trapping FATAL signals causes more troubles than it's worth. On ARM you cannot get working backtraces (unless you compile all code on the target with -fno-omit-frame-pointer), so it's pretty pointless to catch SIGSEGV & SIGABRT. At worst you make it impossible to find the real problem. E.g. if C-library aborts your program due to memory corruption, no function requiring memory allocation, such as backtrace(), printf() etc works properly, they crash instead (maybe causing recursion and eventually really "fun to debug" backtrace in the core dump). - Eero ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: Trapping application "crashes"?
On Tue, Sep 04, 2007 at 11:01:40PM -0700, ext Jayesh Salvi wrote: > If you end up registering your signal handlers - do only little in those > signal handlers. It's not a good idea to keep executing once you have > corrupted memory - you might run into more ugly errors. Indeed, man signal (at least from glibc) explicitly lists the only calls that are safe to run from signal handlers. Cheers, Daniel signature.asc Description: Digital signature ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: Trapping application "crashes"?
If you can modify application's code (at least the main() entry point) you can register signal handlers for SIGSEGV, SIGABRT (most common causes of program crashes) etc. Do "man signal" for more info. If the program is in other languages then there should be corresponding ways to register these signal handlers. If you don't have access to source code then only way that I can suggest is running from debugger. If you end up registering your signal handlers - do only little in those signal handlers. It's not a good idea to keep executing once you have corrupted memory - you might run into more ugly errors. HTH Jayesh On 9/4/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: > > Is it possible for an application to trap abnormal termination? In other > words, can a callback be registered that will be invoked when the > application is about to be shut down because of an error that the > application itself has caused (e.g. memory allocation errors or other > problems that would cause the OS to terminate the application). > > > David Hazel > ___ > maemo-developers mailing list > maemo-developers@maemo.org > https://lists.maemo.org/mailman/listinfo/maemo-developers > ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: mpd
Jeffrey Barish <[EMAIL PROTECTED]> wrote: > I find that uncommenting any of the audio output sections seems to do the > trick. I have the ALSA section uncommented now. But you weren't actually able to play anything right? I got a segfault during the playing of a file with alsa, not on the startup. Running mpd manually and then starting it playing, I see: decoder PID: 1895 decoder (or child) got SIGTERM copyMpdTagToOB: !acceptMetadata || !tag main process got SIGCHLD SIGCHLD caused by player process player process died from signal: 11 So I'd be interested to know if you can actually get it to output sound via alsa (assuming you get it to index files below). > mpd is running now (according to ps), but I still can't get it to index any > sound files. I stuck a symbolic link in /var/lib/mpd/music to the > directory containing my soundfiles (ogg). I run mpd --create-db. The only > output I get is: > > current locale is "C" > setting filesystem charset to ISO-8859-1 How did you get this output? Did you try running mpd manually with: mpd --no-daemon --stdout --verbose --create-db > The files in /var/log/mpd are empty. Everything is owned by user mpd and > mpd has permission to read and write. I must have missed an important > step. Any idea? Are you sure that mpd has permission to read the directory containing your soundfiles? I'd try just copying over a file or two to /var/lib/mpd/music to test. I also haven't tested with ogg files so try copying over a mp3 file also to test. Also, as you manually ran adduser, you may have also missed out on some step in the installation script. Perhaps you could try installing the adduser package and reinstalling the package. ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: mpd
Austin Che wrote: > Creating the mpd user manually should let you run it. I would edit > /etc/mpd.conf and uncomment the libao section. I was getting > segfaults with alsa output while ao output worked fine. I find that uncommenting any of the audio output sections seems to do the trick. I have the ALSA section uncommented now. mpd is running now (according to ps), but I still can't get it to index any sound files. I stuck a symbolic link in /var/lib/mpd/music to the directory containing my soundfiles (ogg). I run mpd --create-db. The only output I get is: current locale is "C" setting filesystem charset to ISO-8859-1 The files in /var/log/mpd are empty. Everything is owned by user mpd and mpd has permission to read and write. I must have missed an important step. Any idea? -- Jeffrey Barish ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: N800 audio connector jack
On Sep 4, 2007, at 7:24 PM, Jami Pekkanen wrote: > Is it possible that something special is required in the input > connection? There at least seems to be some resistors in the headset's > circuit, but I assumed they are just for the push-button. Also does > anybody have a pinout for the jack? I've worked with some systems that sensed the input impedance. You may want to try putting a small resistor in there. I believe that one of the iPaqs even tried to work as mono out/ mono in or stereo out on three pins, depending on the impedance of the output. Perhaps measuring the impedance of the headset first, then putting resistors on to your device, and then taking them off one by one to see which was necessary would work? ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: N800 audio connector jack
On Wed, Sep 05, 2007 at 02:24:52AM +0300, Jami Pekkanen wrote: > Hello, > > This isn't exactly maemo specific question, but I think this is the best > forum anyway. > > I'm trying to solder together an adapter that splits the N800's 4-pole > audio plug to two separate 3.5 mm females so that I can plug in separate > input and output devices, eg electric guitar input and amplifier output > (with maemo doing effects in between). > > So I took apart the headset that comes with the device and got the > output part working fine, but I don't seem to get any signal to the > input. In fact the device even doesn't switch the internal microphone > off when I plug the modified system in. > > I'm assuming that there's two grounds (golden wires, another without > coating), left output on a red wire, right output on green and input on > white. The guesses were right for output, but as I said, the input > doesn't get any signal at all. > > Is it possible that something special is required in the input > connection? There at least seems to be some resistors in the headset's > circuit, but I assumed they are just for the push-button. Also does > anybody have a pinout for the jack? I know of a pinout for a very similar device: http://web.it.kth.se/~maguire/iPAQ-photos/iPAQ-audio-adapter.html E -- Erik Hovland mail: [EMAIL PROTECTED] web: http://hovland.org/ PGP/GPG public key available on request ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
N800 audio connector jack
Hello, This isn't exactly maemo specific question, but I think this is the best forum anyway. I'm trying to solder together an adapter that splits the N800's 4-pole audio plug to two separate 3.5 mm females so that I can plug in separate input and output devices, eg electric guitar input and amplifier output (with maemo doing effects in between). So I took apart the headset that comes with the device and got the output part working fine, but I don't seem to get any signal to the input. In fact the device even doesn't switch the internal microphone off when I plug the modified system in. I'm assuming that there's two grounds (golden wires, another without coating), left output on a red wire, right output on green and input on white. The guesses were right for output, but as I said, the input doesn't get any signal at all. Is it possible that something special is required in the input connection? There at least seems to be some resistors in the headset's circuit, but I assumed they are just for the push-button. Also does anybody have a pinout for the jack? - Jami ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: mpd
>> So I went ahead and ported mpd and mpg321 to maemo. In >> case anyone else wants these packages as I couldn't find them >> already out there: http://austin.mit.edu/maemo > > chage: can't open shadow password fileStopped: `/usr/bin/chage -M 9 mpd' > returned error code 1. Exiting. I don't know where that is coming from as I don't think any of the mpd scripts calls chage. It does use adduser so that should be added to the dependency list. > I tried creating the mpd user by running adduser manually, but then I get > segmentation fault whenever I run mpd. Creating the mpd user manually should let you run it. I would edit /etc/mpd.conf and uncomment the libao section. I was getting segfaults with alsa output while ao output worked fine. ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: Python and GStreamer
Austin Che wrote: > So I went ahead and ported mpd and mpg321 to maemo. In > case anyone else wants these packages as I couldn't find them > already out there: http://austin.mit.edu/maemo I'm having trouble installing mpd. I installed all the packages that mpd requires. When I try to install mpd, I get: /media/mmc1 # dpkg -i mpd_0.13.0-1_armel.deb Selecting previously deselected package mpd. (Reading database ... 15122 files and directories currently installed.) Unpacking mpd (from mpd_0.13.0-1_armel.deb) ... Setting up mpd (0.13.0-1) ... chage: can't open shadow password fileStopped: `/usr/bin/chage -M 9 mpd' returned error code 1. Exiting. Removing user `mpd' ... dpkg: error processing mpd (--install): subprocess post-installation script returned error exit status 1 Errors were encountered while processing: mpd I tried shadowconfig on, but then I get "bash: not found". Installing bash would change everything. I tried creating the mpd user by running adduser manually, but then I get segmentation fault whenever I run mpd. Hints? -- Jeffrey Barish ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
sbox2 mapping mode rework
Hi, I changed sbox2 default mapping style so that it by default uses files from the host and only maps the bare minimum to target buildroot. This simplified the rules dramatically and I think it means that most tools will "just work" and only special cases like apt, dpkg, gcc and qemu need special treatment. I did a quick test and managed to build debian packages pretty well, there's maybe a little glitch somewhere with respect to fakeroot and qemu and LD_PRELOAD passing, but overall things look pretty damn good considering that now the whole set of rules is just 54 lines long. I also disabled the mapping cache by default, it was causing some trouble and the new rule set is so small that I don't think any cache is needed. This stuff is currently in the git repository so pull from there to give it a shot. /lauri ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: Treeview multiple selection mode
Jeffrey Barish wrote: > I just got multiple selection to work in a test application using PyGTK, > so something is wrong in my application. Although I have multiple selection working in a test program, I still have not been able to get it to work in my application. Only the first item that I touch gets selected. I am now capturing button-press-event and button-release-event, so I can see the press when I contact the first row and the release after I sweep out a region, but none of the other rows get selected. I have confirmed that the mode is SELECTION_MULTIPLE when I receive the first selection event. I disabled the handler for the tree selection changed signal, so the problem cannot be in my handler. I can't think what else to try. Any suggestions? -- Jeffrey Barish ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: Network connection with hildon-app-mgr in the scratchbox
Marius Vollmer escribió: > ext Mario Sánchez Prada <[EMAIL PROTECTED]> writes: > >> ...and compiling hildon-application-manager without problem I can >> perfectly run it but I can't get network connection so I'm not able >> to install anything from the repositories using this program. > > Please try the 'assume-connection' setting, available in red-pill mode > in the "Tools > Settings" dialog. The AM is trying to get a > network connection up, but that doesn't work in Scratchbox. > Thanks for your suggestion, but IMHO that doesn't fix the real problem behind the scenes... I mean, why some applications (such as the osso-xterm) can use a network connection inside the scratchbox while hildon-app-mgr can't? That's what really looks strange to me, and I didn't found a solution yet, I just have to install applications from the commandline but I still think that there should be a way to "fix" that, and I'd like to find it. Regards Mario ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Trapping application "crashes"?
Is it possible for an application to trap abnormal termination? In other words, can a callback be registered that will be invoked when the application is about to be shut down because of an error that the application itself has caused (e.g. memory allocation errors or other problems that would cause the OS to terminate the application). David Hazel ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Re: Network connection with hildon-app-mgr in the scratchbox
ext Mario Sánchez Prada <[EMAIL PROTECTED]> writes: > ...and compiling hildon-application-manager without problem I can > perfectly run it but I can't get network connection so I'm not able > to install anything from the repositories using this program. Please try the 'assume-connection' setting, available in red-pill mode in the "Tools > Settings" dialog. The AM is trying to get a network connection up, but that doesn't work in Scratchbox. ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
Network connection with hildon-app-mgr in the scratchbox
Hi, After installing all this stuff: - Scratchbox Apophis r4 [1] - Maemo 4.0 Chinook alpha SDK [2] - Hildon application manager (SVN) [3] ...and compiling hildon-application-manager without problem I can perfectly run it but I can't get network connection so I'm not able to install anything from the repositories using this program. I know I can install all what I want from the scratchbox command line with 'apt-get' as usual, but the fact is that I'd like to know why this strange behavior occurs with the application manager, since it's something which doesn't happen with other installed programs from the emu, such as the osso-xterm (I can do an 'wget URL' without problems using this app). I've tried to look into the code to see if I can find what it's going wrong there but all I got is a problem resolving the name of the repositories in the apt client utility built-in with the hildon-application-manager... but I don't know why it happens either. Any idea? Thanks in advance Mario -- [1] http://scratchbox.org/download/scratchbox-apophis [2] http://maemo.org/development/sdks/maemo_4-0_chinook_alpha_sdk.html [3] svn checkout https://garage.maemo.org/svn/hildon-app-mgr ___ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers
