http://permalink.gmane.org/gmane.comp.ide.gps.devel hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://permalink.gmane.org/gmane.comp.ide.gps.devel/196 <pre>For GNAT GPL 2012 on OS X... When using a remote, either Solaris or another OS X box, when I try to debug I get an error about /dev/ttys003 not found and the debugged program dies. In looking at the remote debug session log I see that GPS is sending "tty /dev/ttys003" to gdb. This doesn't exist on the remote system and is actually on OS X where GPS is running. </pre> Steve Kappel 2013-04-14T14:52:05 http://permalink.gmane.org/gmane.comp.ide.gps.devel/195 <pre> I don't think it really is more efficient in real cases where there are two or three directories in Project_Path... but OK. You could improve the patch further by getting rid of the Found variable and directly saying if not Visited.Contains (Project_Path (J)) then ... </pre> Ludovic Brenta 2012-04-10T12:07:19 http://permalink.gmane.org/gmane.comp.ide.gps.devel/194 <pre> Thanks. I agree there is a bug there, although I chose a different approach than you patch (using a Set is more efficient than looping as we did before -- this was pre-Ada2005 code). The following patch was applied, which I believe should work. Emmanuel --- a/prj_editor/src/creation_wizard-dependencies.adb +++ b/prj_editor/src/creation_wizard-dependencies.adb &lt; at &gt;&lt; at &gt; -15,6 +15,7 &lt; at &gt;&lt; at &gt; -- of the license. -- ------------------------------------------------------------------------------ +with Ada.Containers.Hashed_Sets; with Glib; use Glib; with Gtk.Box; use Gtk.Box; with Gtk.Button; use Gtk.Button; &lt; at &gt;&lt; at &gt; -58,6 +59,9 &lt; at &gt;&lt; at &gt; package body Creation_Wizard.Dependencies is package Wizard_Page_Handlers is new Gtk.Handlers.User_Callback (Gtk_Widget_Record, Project_Wizard_Page); + package File_Sets is + new Ada.Containers.Hashed_Sets (Virtual_File, Full_Name_Hash, + </pre> Emmanuel Briot 2012-04-10T12:00:53 http://permalink.gmane.org/gmane.comp.ide.gps.devel/193 <pre>While investigating http://bugs.debian.org/666958, I discovered a bug in prj_editor/src/creation_wizard-dependencies.adb; the fix is: --- a/prj_editor/src/creation_wizard-dependencies.adb +++ b/prj_editor/src/creation_wizard-dependencies.adb &lt; at &gt;&lt; at &gt; -278,8 +278,9 &lt; at &gt;&lt; at &gt; -- Make sure the path isn't duplicated Found := False; - for K in Project_Path'Range loop + for K in Project_Path'First .. J - 1 loop if Project_Path (J) = Project_Path (K) then + -- We've already processed the same path before. Abort. Found := True; exit; end if; If you look in the sources, you will see that, without the patch, the condition "Project_Path (J) = Project_Path (K)" is always true at least once (i.e. when J = K) and therefore Found always becomes True and the rest of the loop is never executed. Also, there is a risk that a path that appears twice (or more) in the Project_Path array is *never* processed. The patch carefull</pre> Ludovic Brenta 2012-04-09T13:45:02 http://permalink.gmane.org/gmane.comp.ide.gps.devel/192 <pre>  Hello Felix,   [...]     Recent versions of GPS pass the contents of the variable GPS_PYTHONHOME to its Python interpreter. (We did use "PYTHONHOME" before, but found out some cases in which this interfered too much with the users' environment).     You are probably not far from succeeding with the recent version. Don't hesitate to post problems here - it's possible that someone has solved them before.   Regards,   Nicolas   _______________________________________________ gps-devel mailing list gps-devel&lt; at &gt;lists.adacore.com http://lists.adacore.com/mailman/listinfo/gps-devel </pre> Nicolas Setton 2011-12-02T23:36:17 http://permalink.gmane.org/gmane.comp.ide.gps.devel/191 <pre>Hi everyone, I'm trying to compile GPS and link it against GTK-OSX (http://live.gnome.org/GTK%2B/OSX/Building), which provides a more native interface on MacOSX. So far, I succeeded in compiling the binary, but when I try to start it, it always fails to initialize Python (I linked against the Python framework that comes with MacOSX Lion). My first attempt looked like this: $ ~/gtk/inst/bin/gps Gtk-WARNING **: Locale not supported by C library. Using the fallback 'C' locale. ImportError: No module named site I guess the Gtk-WARNING can be ignored for now. The problem here is that Python does not know the directory where its modules are located. So I did: $ export PYTHONPATH=/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7 $ ~/gtk/inst/bin/gps Gtk-WARNING **: Locale not supported by C library. Using the fallback 'C' locale. Traceback (most recent call last): File "/System/Library/Frameworks/Python.framework/</pre> Felix Krause 2011-12-02T22:21:33 http://permalink.gmane.org/gmane.comp.ide.gps.devel/190 <pre> I forgot to send you the current aclocal.m4, if you want to test on FreeBSD. I have used it successfully on our own FreeBSD machines, but that's not a system we use a lot, so there might be differences on your own machine. Eagerly waiting for the next patches :-) Emmanuel _______________________________________________ gps-devel mailing list gps-devel&lt; at &gt;lists.adacore.com http://lists.adacore.com/mailman/listinfo/gps-devel </pre> Emmanuel Briot 2011-10-11T12:49:31 http://permalink.gmane.org/gmane.comp.ide.gps.devel/189 <pre> Thanks Emmanuel, I'll consider this a "5.0.1" issue only then. I have a couple of other patches on a different topic that I'll send along soon. Regards, John </pre> John Marino 2011-10-11T12:40:20 http://permalink.gmane.org/gmane.comp.ide.gps.devel/188 <pre> Thank you for the patch. However, it should not be needed in the current development version of GPS, because configure will now first try to use "python-config" to find out the list of libraries and flags to use. regards Emmanuel </pre> Emmanuel Briot 2011-10-11T12:07:48 http://permalink.gmane.org/gmane.comp.ide.gps.devel/187 <pre>Hi, I discovered python support wasn't getting compiled into GPS 5.0.1 on an AMD64 FreeBSD machine even though it was explicitly requested. The log said the configure test failed due to use of "-ldl" which isn't available on FreeBSD. Looking at the configure script, it was obvious that targets such as x86_64-unknown-freebsd8.2 were hitting a way-too-general x86_64-*-* case before it could reach the *-freebsd* case. The quick solution is to put the freebsd case in front of the x86_64 case. At the same time, missing support for NetBSD, OpenBSD, and DragonFly can be added. The proposed patches are attached. Regards, John --- gnatlib/aclocal.m4.orig2010-07-19 10:57:13.000000000 +0200 +++ gnatlib/aclocal.m42011-10-11 03:05:28.000000000 +0200 &lt; at &gt;&lt; at &gt; -452,6 +452,18 &lt; at &gt;&lt; at &gt; *-darwin* ) PYTHON_LIBS="-ldl -lm ${PYTHON_LIBS}" ;; + *-freebsd* ) + PYTHON_LIBS="-lm -lutil ${PYTHON_LIBS}" + ;; + *-dragonfly* ) + PYTHON_LIBS="-lm -l</pre> John Marino 2011-10-11T11:49:43 http://permalink.gmane.org/gmane.comp.ide.gps.devel/186 <pre> c.l.a is indeed your best bet. You can access newsgroups via any web browser these days. Arno </pre> Arnaud Charlet 2010-09-28T08:25:36 http://permalink.gmane.org/gmane.comp.ide.gps.devel/185 <pre>_______________________________________________ gps-devel mailing list gps-devel&lt; at &gt;lists.adacore.com http://lists.adacore.com/mailman/listinfo/gps-devel </pre> Francesco Piraneo G. 2010-09-28T08:23:15 http://permalink.gmane.org/gmane.comp.ide.gps.devel/182 <pre>Thank you for your message. I will forward it to our legal department, since this list is for technical discussions. Happy New Year! </pre> Arnaud Charlet 2010-01-03T10:46:04 http://permalink.gmane.org/gmane.comp.ide.gps.devel/181 <pre>_______________________________________________ gps-devel mailing list gps-devel&lt; at &gt;lists.adacore.com http://lists.adacore.com/mailman/listinfo/gps-devel </pre> Ludovic Brenta 2010-01-02T21:30:07 http://permalink.gmane.org/gmane.comp.ide.gps.devel/180 <pre>First, I am sorry if I bothered the community with a well known problem. I didn't find the FAQ on GPS, didn't find how to search in adacore lists, didn't find a hit through .oogle... Second, indeed reconfiguring gnome solves the problem. This is for sure the explanation. Still, I am puzzled: I discovered the problem one year ago and I didn't report it because I had no clue. Recently I have found the workaround of clicking twice the Edit menu button, and I thought this was a GPS bug (so I reported it in the devel list). Now I don't understand how your explanation can cope with this workaround, even if both work all right! Regards Arnaud Charlet wrote: </pre> Pascal MALAISE 2009-12-01T20:38:10 http://permalink.gmane.org/gmane.comp.ide.gps.devel/179 <pre> This is actually a FAQ, you'll find the answer in the archives of this list, since this has been answered a few times already. Here it is again (from the svn version of the GPS doc also): &lt;&lt; &lt; at &gt;item Using the space key brings the smart completion window under Ubuntu This is specific to the way GNOME is configured on Ubuntu distributions. To address this incompatibility, close GPS, then go to the GNOME menu &lt; at &gt;i{System-&gt;Preferences-&gt;Keyboard} (or launch &lt; at &gt;command{gnome-keyboard-properties}). Select the &lt; at &gt;i{Layout} tab, click on &lt; at &gt;i{Layout Options}. Then click twice on &lt; at &gt;i{Using space key to input non-breakable space character} and then select &lt; at &gt;i{Usual space at any level} and then close the dialogs. Arno </pre> Arnaud Charlet 2009-11-30T20:48:49 http://permalink.gmane.org/gmane.comp.ide.gps.devel/178 <pre>Thank you for the proposal but this is not the explanation, for the following reasons: - I have checked, no definition of "space" - I have just defined a shortcut as "space" (which would replace any definition of it) then removed it. The problem still appears - Please consider the workaround: open and close the "Edit" by clicking twice on the Edit button of the top bar. No more no less, 100% reproducible. This would not solve a shortcut definition. This rather sounds like a non-initialized variable or so. I am ready to help you to investigate further on, but I have no idea any more (I used all ideas in investigating this simple workaround). Regards Emmanuel Briot wrote: </pre> Pascal MALAISE 2009-11-30T19:27:17 http://permalink.gmane.org/gmane.comp.ide.gps.devel/177 <pre> You have most likely defined space as a shortcut for some action. Open the Key Shortcuts dialog (/Edit menu), and then search for space in the second column to reset that shortcut. Emmanuel </pre> Emmanuel Briot 2009-11-30T09:47:44 http://permalink.gmane.org/gmane.comp.ide.gps.devel/176 <pre>Hi, on Ubuntu 8.04 and 9.10 I have the following bug with GPS 4.3.1 (20090114) hosted on i686-pc-linux-gnu GNAT GPL 2009 (20090519): - I start "gps --debug=&lt;exe&gt;" - I click in the start/run ("&gt;") button to start execution - The run/start popup window appears, I clock in OK - The program breaks in the main, so far so good - In the debugger console, (after the prompt (gdb)) I type "p " (p and the space key) - gps switches to the messages window with the error: [2009-11-28 09:08:52] Invalid context for this action Occures with all commands typed in the console window, just after typeing &lt;space&gt;. Note that I haven't open any menu. Workaround: Before (or even after) the problem occurs, I open and close the "Edit" menu by clicking twice on the "Edit" button. This works only with the Edit menu buDoes not work with any ottton, not with the other menu buttons of GPS. </pre> Pascal MALAISE 2009-11-28T08:23:59 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.comp.ide.gps.devel