http://permalink.gmane.org/gmane.comp.lang.gambas.user hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://permalink.gmane.org/gmane.comp.lang.gambas.user/23827 <pre> Thanks for explaining the problem and possible solutions Benoît! I have forwarded your input to wiringPi's author so lets see whats happening and if he may consider adding a "interpreter friendly" interrupt function to his library. /CJ ------------------------------------------------------------------------------ Try New Relic Now &amp; We'll Send You this Cool Shirt New Relic is the only SaaS-based application performance monitoring service that delivers powerful full stack analytics. Optimize and monitor your browser, app, &amp; servers with just a few lines of code. Try New Relic and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may </pre> CJ 2013-05-25T21:48:46 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23826 <pre>Le 25/05/2013 23:05, Antonio Sánchez a écrit : You can't at the moment, but you can handle the "Icon" event instead, and set the Icon property: that allows you to create the image preview yourself as you like. Regards, </pre> Benoît Minisini 2013-05-25T21:11:50 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23825 <pre>For our purpose, yes, it is a bit low. Where could we change this limit? 2013/5/25 Benoît Minisini &lt;gambas&lt; at &gt;users.sourceforge.net&gt; ------------------------------------------------------------------------------ Try New Relic Now &amp; We'll Send You this Cool Shirt New Relic is the only SaaS-based application performance monitoring service that delivers powerful full stack analytics. Optimize and monitor your browser, app, &amp; servers with just a few lines of code. Try New Relic and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may </pre> Antonio Sánchez 2013-05-25T21:05:16 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23824 <pre>Le 25/05/2013 22:43, Antonio Sánchez a écrit : When the file is too big (&gt; 64K), the preview is not shown. Maybe the limit is a bit low? </pre> Benoît Minisini 2013-05-25T20:53:00 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23823 <pre>We are using FileChooser control with "ShowPreview" to True, but only get some previews and we don't know why. We find no aparent reason. I attached a screenshot. ------------------------------------------------------------------------------ Try New Relic Now &amp; We'll Send You this Cool Shirt New Relic is the only SaaS-based application performance monitoring service that delivers powerful full stack analytics. Optimize and monitor your browser, app, &amp; servers with just a few lines of code. Try New Relic and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may_______________________________________________ Gambas-user mailing list Gambas-user&lt; at &gt;lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/gambas-user </pre> Antonio Sánchez 2013-05-25T20:43:00 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23822 <pre>Le 25/05/2013 20:54, CJ a écrit : The interpreter cannot be run in another thread. This is the reason why you get a crash. Two possible solutions : 1) A pure Gambas solution. But that needs writing a component in C/C++ to implement an interrupt handler that would write to a pipe that would be read by the main thread (the one that runs the Gambas interpreter). 2) Ask to the wiringPi's author to add a new function to handle the interrupt (or modify the current one) that proceeds that way: - Create an internal pipe. - Create the thread. - When an interrupt occurs, the thread writes a byte to the pipe. - The function returns the file descriptor of the other side of the pipe. That way, Gambas would just have to watch the file descriptor returned by the new wiringPiISR function to know if an interrupt has occured. Note that many other interpreters will be happy with that solution! Regards, </pre> Benoît Minisini 2013-05-25T19:35:17 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23821 <pre> The same thing. Here is the response from wiringPi's author Gordon Henderson... The interrupt handler is a standard pthread process. The wiringPiISR code creates the thread and that thread then waits for the interrupt, then calls the ISR function - so the function called has to be "thread safe" - something I possibly ought to better document... My bad! Reason to this is that I cleaned up the code before making the tar-ball below is the correct output for the attached code... root&lt; at &gt;raspi-dev:/home/Projekt/libwiringPi_irq# ./libwiringPi_irq.gambas wiringPi Init OK... IRQ setup OK... Waiting.... Waiting.... Waiting.... Waiting.... Waiting.... MMain.Main.52: #3: Stack overflow 1: MMain.Main.52 And output from gdb... ... (gdb) bt #0 THROW (code=0) at gb_error.c:465 #1 0x0004a9e4 in THROW_STACK () at gb_error.c:489 #2 0x0004a9e4 in THROW_STACK () at gb_error.c:489 Backtrace stopped: previous frame identical to this frame (corrupt stack?) (gdb) The stack overfl</pre> CJ 2013-05-25T18:54:27 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23820 <pre>Le 25/05/2013 18:32, CJ a écrit : What happens if there is no code at all inside the interrupt handler? Do you know in which context the interrupt handler is run? (signal handler context? Something else?) And I don't understand why you get an error on line #110, whereas there is only 58 lines in your MMain.module file... Regards, </pre> Benoît Minisini 2013-05-25T17:33:14 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23819 <pre> After compiling and updating to latest stable 3.4.1 version I retested the HW interrupt code I had problems with when running 3.3.4 and have the same error when testing on the RaspberryPi... Code compiles OK but when I pull GPIO18 pin low (i.e generating the interrupt) it crashes with a "Stack overflow" error message. root&lt; at &gt;raspi-dev:/home/Projekt/libwiringPi_irq# ./libwiringPi_irq.gambas wiringPi Init OK... IRQ setup OK... Waiting.... Waiting.... Waiting.... Waiting.... Waiting.... Waiting.... Waiting.... Waiting.... Waiting.... MMain.Main.110: #3: Stack overflow 1: MMain.Main.110 Segmentation fault Below is the output from gdb... ... (gdb) bt #0 THROW (code=0) at gb_error.c:465 #1 0x0004a9e4 in THROW_STACK () at gb_error.c:489 #2 0x0004a9e4 in THROW_STACK () at gb_error.c:489 Backtrace stopped: previous frame identical to this frame (corrupt stack?) (gdb) A informed by Benoît this is how to declare it in Gambas3 to use the interrupt function in wiringPi </pre> CJ 2013-05-25T16:32:51 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23818 <pre>Just a test since I've had problems posting to this list, please ignore! /CJ ------------------------------------------------------------------------------ Try New Relic Now &amp; We'll Send You this Cool Shirt New Relic is the only SaaS-based application performance monitoring service that delivers powerful full stack analytics. Optimize and monitor your browser, app, &amp; servers with just a few lines of code. Try New Relic and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may </pre> CJ 2013-05-25T16:29:07 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23817 <pre>Le 21/05/2013 16:13, Tobias Boege a écrit : Global error management routine has been implemented in revision #5678. This routine must be a static public method of the startup class, and must be named "Application_Error". Inside this method, you can use the properties of the Error class to get information about the error. If an uncatched error is raised while running Application_Error, then the program will abort with that error replacing the old one. Beware, this was not tested yet! The IDE has been modified to try to save all modified files before aborting if an error occurs. Enjoy! </pre> Benoît Minisini 2013-05-25T14:03:25 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23816 <pre>Le 25/05/2013 11:37, Tobias Boege a écrit : My solution is in r5677: I took yours and simplified it by removing what I don't understand. :-) It creates a void 'trunk_version.h' if there is no '.svn' directory. All seems to work correctly, but I'm waiting for your opinion to be sure! :-) </pre> Benoît Minisini 2013-05-25T13:33:18 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23815 <pre> This is useful, thanks In your example code you stated there is no automatic way to convert strings to hex. It is possible to do using a loop and the Hex$ function, like so: Public Function String2Hex(s As String) As String Dim ret As String = "" Dim i As Integer For i = 1 To Len(s) ret &amp;= Hex$(Asc(s, i), 2) Next Return ret End Cheers, Ian ------------------------------------------------------------------------------ Try New Relic Now &amp; We'll Send You this Cool Shirt New Relic is the only SaaS-based application performance monitoring service that delivers powerful full stack analytics. Optimize and monitor your browser, app, &amp; servers with just a few lines of code. Try New Relic and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may </pre> Ian Haywood 2013-05-25T11:07:34 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23814 <pre> Do you mean that you can import an .info file to the wiki? This would be great to kick-start the gb.openssl documentation if you used my modified .info file. Regards, Tobi ------------------------------------------------------------------------------ Try New Relic Now &amp; We'll Send You this Cool Shirt New Relic is the only SaaS-based application performance monitoring service that delivers powerful full stack analytics. Optimize and monitor your browser, app, &amp; servers with just a few lines of code. Try New Relic and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may </pre> Tobias Boege 2013-05-25T09:40:42 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23813 <pre> Good idea. Actually I thought about that, too, but rejected it because a Makefile rule that depends on .svn will not work on non-repository sources, like: make: *** No rule to make target `.svn', needed by `all'. Stop. which is from a Makefile inside my home directory - not svn versioned. What I didn't think of was this workaround: have a .PHONY target that compares timestamps of .svn to the header's instead of letting make do it. We could catch the case that .svn doesn't exist: --- trunk_version.h: TRUNK_VERSION = $(shell LC_ALL=C svn info 2&gt;/dev/null | grep Revision | egrep -wo "[0-9]+") trunk_version.h: &lt; at &gt;if test -d .svn -a .svn -nt trunk_version.h ; then \ echo -e "/* TRUNK version if available */\n#define TRUNK_VERSION\t$(TRUNK_VERSION)\n" &gt;trunk_version.h; \ fi .PHONY: trunk_version.h --- This worked in my home directory, i.e. did nothing but without complaint, and in trunk/ it created the trunk_version.h. Maybe ".svn" above should be prefixed with the path to the top-level Gambas</pre> Tobias Boege 2013-05-25T09:37:57 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23812 <pre>Le 24/05/2013 23:34, Benoît Minisini a écrit : OK, my fault. The problem should be fixed in revision #5675. Regards, </pre> Benoît Minisini 2013-05-24T22:23:45 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23811 <pre>Le 24/05/2013 22:28, Tobias Boege a écrit : I fixed the documentation (that was the old text from Gambas 2). </pre> Benoît Minisini 2013-05-24T21:39:07 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23810 <pre>Le 23/05/2013 14:09, Shane a écrit : Indeed. Apparently I did as the curl documentation said, but the curl progress callback is never called. I will investigate... </pre> Benoît Minisini 2013-05-24T21:34:10 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23809 <pre>Le 23/05/2013 14:23, Tobias Boege a écrit : You get a point there. But you must find a way to get the svn version only when it changes. For example, we can imagine adding a special makefile rule that computes the svn version and put it in its custom header file. This rule will depend on the modification time of the /trunk/.svn directory (if it is possible). What do you think? </pre> Benoît Minisini 2013-05-24T21:33:35 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23808 <pre>Le 24/05/2013 23:14, Tobias Boege a écrit : No, component help must be written on the wiki. Component help can be written on the wiki only after I have uploaded the information files on the wiki server. I will upload the information files after having taken a look at the interface of your component. Regards, </pre> Benoît Minisini 2013-05-24T21:28:34 http://permalink.gmane.org/gmane.comp.lang.gambas.user/23807 <pre>Hi, it is said in [0] that the IDE can access an .info file of a project to get documentation out of it. I tried this with gb.openssl which I documented in its source code. It did not work since it's a component and not a project. As the gambas-devels may already know, I have a bunch of scripts to extract these comments and merge them into existing .info files. This works properly meanwhile. Consequently, I'd like to see the IDE display the help comments in a *component's* .info file. Is that possible? Regards, Tobi [0] http://gambasdoc.org/help/doc/helpcomments?v3 ------------------------------------------------------------------------------ Try New Relic Now &amp; We'll Send You this Cool Shirt New Relic is the only SaaS-based application performance monitoring service that delivers powerful full stack analytics. Optimize and monitor your browser, app, &amp; servers with just a few lines of code. Try New Relic and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may </pre> Tobias Boege 2013-05-24T21:14:52 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.comp.lang.gambas.user