Re: [Geany-devel] Proposed patch to fix issues with command line file loading
On Mon, 31 Jan 2011 09:06:26 +1100 Lex Trotman ele...@gmail.com wrote: On 31 January 2011 02:54, Frank Lanitz fr...@frank.uvena.de wrote: On Fri, 28 Jan 2011 11:31:21 +0300 Eugene Arshinov earshi...@gmail.com wrote: On Thu, 27 Jan 2011 10:20:26 +0100% weltall welta...@gmail.com wrote: Hi, I've noticed various issues with file loading from command line: 1) if the option load last session is not enabled project files can't be opened from command line: this happens because the code to handle them is wrapped in a prefs.load_lastsession if so the only way to load them from command line is having that option true. 2) trying to open a file (eg from double clicking in nautilus) will screw your session as command line loading of single files ignores your session opened files, except if geany was already opened. So to fix those issues I've reorganized the code in the load_startup_file function in order to: 1) load the project files ignoring the session if a project file was specified at command line 2) allow to load more files in addition to the project file (removing so the limitation imposed there artificially - possible improvement to this would be adding a flag to open_cl_files as argument in order to avoid that pointer toying) 3) the open_cl_files is called after loading the session so the specifically selected files are opened last and so are automatically selected (what you would expect when opening a file from nautilus) 4) if the option to load last session is not enabled and there isn't a project file being loaded from command line and a project is not being loaded we just try to load files from command line if any Stefano Angeleri Just to mention, I had similar objections about a year ago when I wrote my version of session management support (the list probably knows what I'm talking about). This was discussed a little, and some patches were included in 'sm' branch. For example, load_startup_files () function in that branch differs significantly from trunk, and open_cl_files() does not have a check for `argc = 1`. Geany's SVN browse seems to be unavailable now, but I think it should be possible there to view 'sm' branch. The branch is still available. To checkout its svn co https://geany.svn.sourceforge.net/svnroot/geany/branches/sm The sf viewer seems to be really down at the moment. It is harder to find the discussion than the patches themselves, but here is a couple of links: http://lists.uvena.de/pipermail/geany-devel/2009-November/001577.html (X session management support) http://lists.uvena.de/geany-devel/2010-January/001655.html (Questions about Geany project support) I don't object against your patches, but I'm just pointing to existing discussion which may be useful. I must note that there are no plans about including those patches from 'sm' branch in trunk (because nobody would want/have time to do it, and it won't be quite easy as 'sm' branch has become rather diverse from trunk), so your way is free :) Well, not sure whether I'm right the branch wasn't ready for including to trunk and nobody did care about later on. Please correct me if I'm wrong. I guess now (post 0.20) is the time to do major additions, if a current trunk is merged with sm branch and it still works then maybe its time. I guess this is something the original author of the branch needs to do. Cheers, Frank -- http://frank.uvena.de/en/ pgpEuotRH6nr3.pgp Description: PGP signature ___ Geany-devel mailing list Geany-devel@uvena.de http://lists.uvena.de/cgi-bin/mailman/listinfo/geany-devel
Re: [Geany-devel] A few fixes for codenav, geanygdb, geanyinsertnum and geanylatex
Le 26/02/2011 13:41, Frank Lanitz a écrit : On Wed, 23 Feb 2011 19:32:07 +0100 Colomban Wendling lists@herbesfolles.org wrote: [./geanylatex/src/templates.c:47]: (error) Memory leak: template I also attach here the fixes I propose for them, if you're interested. I'm afraid I don't see why this should be a memory leak. Can you please be so kind and go into more details to prevent similar cases in future? It's a possible memory leak because you exit the function conditionally after the allocation, but don't free it in this branch: TemplateEntry *template = g_new0(TemplateEntry, 1); gchar *tmp = NULL; /* Return if its not a searched file */ if (g_str_has_suffix(file,.gtl) == FALSE) /* Leak here! */ return; /* ... */ g_free(template); So the easiest fix is to allocate the memory after the conditional exit since it isn't used before. Regards, Colomban ___ Geany-devel mailing list Geany-devel@uvena.de http://lists.uvena.de/cgi-bin/mailman/listinfo/geany-devel
Re: [Geany-devel] A few fixes for codenav, geanygdb, geanyinsertnum and geanylatex
Le 26/02/2011 14:50, Frank Lanitz a écrit : On Sat, 26 Feb 2011 14:21:11 +0100 Colomban Wendling lists@herbesfolles.org wrote: Le 26/02/2011 13:41, Frank Lanitz a écrit : On Wed, 23 Feb 2011 19:32:07 +0100 Colomban Wendling lists@herbesfolles.org wrote: [./geanylatex/src/templates.c:47]: (error) Memory leak: template I also attach here the fixes I propose for them, if you're interested. I'm afraid I don't see why this should be a memory leak. Can you please be so kind and go into more details to prevent similar cases in future? It's a possible memory leak because you exit the function conditionally after the allocation, but don't free it in this branch: TemplateEntry *template = g_new0(TemplateEntry, 1); gchar *tmp = NULL; /* Return if its not a searched file */ if (g_str_has_suffix(file,.gtl) == FALSE) /* Leak here! */ return; /* ... */ g_free(template); So the easiest fix is to allocate the memory after the conditional exit since it isn't used before. True. Ah... and now I see. You were not doing a patch against trunk of geanylatex. Its fixed already inside http://yaturl.net/2c80 Oops, I didn't though main development don't go in geany-plugins/geanylatex, sorry :/ However, I suggest you to do the allocation after the conditional exit rather than before and freeing it if taking the branch, because alloc/free is far from being a no-op ;) Cheers, Colomban ___ Geany-devel mailing list Geany-devel@uvena.de http://lists.uvena.de/cgi-bin/mailman/listinfo/geany-devel
Re: [Geany-devel] A few fixes for codenav, geanygdb, geanyinsertnum and geanylatex
On Sat, 26 Feb 2011 16:45:13 +0100 Colomban Wendling lists@herbesfolles.org wrote: Le 26/02/2011 14:50, Frank Lanitz a écrit : On Sat, 26 Feb 2011 14:21:11 +0100 Colomban Wendling lists@herbesfolles.org wrote: Le 26/02/2011 13:41, Frank Lanitz a écrit : On Wed, 23 Feb 2011 19:32:07 +0100 Colomban Wendling lists@herbesfolles.org wrote: [./geanylatex/src/templates.c:47]: (error) Memory leak: template I also attach here the fixes I propose for them, if you're interested. I'm afraid I don't see why this should be a memory leak. Can you please be so kind and go into more details to prevent similar cases in future? It's a possible memory leak because you exit the function conditionally after the allocation, but don't free it in this branch: TemplateEntry *template = g_new0(TemplateEntry, 1); gchar *tmp = NULL; /* Return if its not a searched file */ if (g_str_has_suffix(file,.gtl) == FALSE) /* Leak here! */ return; /* ... */ g_free(template); So the easiest fix is to allocate the memory after the conditional exit since it isn't used before. True. Ah... and now I see. You were not doing a patch against trunk of geanylatex. Its fixed already inside http://yaturl.net/2c80 Oops, I didn't though main development don't go in geany-plugins/geanylatex, sorry :/ However, I suggest you to do the allocation after the conditional exit rather than before and freeing it if taking the branch, because alloc/free is far from being a no-op ;) Yep, you are right. Will do this ;) Cheers, Frank -- http://frank.uvena.de/en/ pgpTvlJRMaBhJ.pgp Description: PGP signature ___ Geany-devel mailing list Geany-devel@uvena.de http://lists.uvena.de/cgi-bin/mailman/listinfo/geany-devel
