September 2006 - graphic-forms-devel@lists.common-lisp.net

[graphic-forms-devel] SLIME vs. message loops
by Jack Unrue 09 Sep '06

09 Sep '06

I need to do some research into the interaction between SLIME and Windows GUI message loops. Michael Jung was able to get a test program working if he ran SBCL from a command prompt instead of SLIME -- and a command prompt window is the mode I run the various CLs in as well, so I had not yet seen this problem with SLIME. I will post an update once I have investigated this further. -- Jack Unrue

1 1

[graphic-forms-devel] API change: select-span, selection-span --> selected-span
by Jack Unrue 06 Sep '06

06 Sep '06

I've got another little API cleanup issue to report. The existing generic functions SELECT-SPAN and SELECTION-SPAN in the GFW package are going to be collapsed to SELECTED-SPAN which will have a corresponding SETF function defined as well. This affects existing code that uses edit controls. -- Jack Unrue

1 0

Re: [graphic-forms-devel] graphic-forms
by Michael Jung 04 Sep '06

04 Sep '06

At 19:15 04.09.2006, you wrote: >If it were me at this point, I would start sprinkling debug statements >around to try to find out more info. I would do this with Alastair's >hello.lisp since it is such a small amount of code to work with. >Check that REGISTERCLASSEX is not returning zero, and that >CREATEWINDOW is returning a non-null foreign pointer. Check >that his WINDOW-PROCEDURE function is getting called, >and print out the argument values, so that you can see what >events get processed. > >If those debug statements indicate everything is succeeding, >but you still don't see any window, then I would suggest installing >a Spy utility (I use Winspector Spy) so that you can check whether >the window is getting created but not being made visible. Alastair's >window classname is "hello". Winspector Spy says that window class is registered and window is created. At the moment it looks as if it get stuck in the message loop Michael Jung

2 1

Re: [graphic-forms-devel] graphic-forms
by Michael Jung 04 Sep '06

04 Sep '06

At 19:23 04.09.2006, you wrote: >Another thing to try is to restart SBCL without the callback >patch, and remove the :stdcall attribute from Alastair's >code. Win32 has a special hack that allows non stdcall >callbacks to be used as window procedures. This would >help rule out whether that patch is causing the problem. Jack, I tried this (hopefully the right way) too but it changed nothing Michael Jung

1 0

Re: [graphic-forms-devel] graphic-forms
by Michael Jung 04 Sep '06

04 Sep '06

At 19:23 04.09.2006, you wrote: >Also, you're only loading the callback patch once, right? I don't >know what happens if you execute that code multiple times; it >might be harmless or it might not. Graphic-Forms automatically >loads it via the ASD file. > >Another thing to try is to restart SBCL without the callback >patch, and remove the :stdcall attribute from Alastair's >code. Win32 has a special hack that allows non stdcall >callbacks to be used as window procedures. This would >help rule out whether that patch is causing the problem. It didn't change anything loading the patch before compiling graphic-forms or not. I will try and remove the attribute and see what happens Michael Jung

1 0

Re: [graphic-forms-devel] graphic-forms
by Michael Jung 04 Sep '06

04 Sep '06

At 19:15 04.09.2006, you wrote: >On 9/4/06, Michael Jung <jung(a)mijn-design.de> wrote: >> >>Hello Jack, >> >>At 17:55 04.09.2006, you wrote: >> >> >OK, that's helpful info. Check your taskbar and see if in fact a window >> >did get created but at the bottom of the desktop Z-order instead of >> >the top. I notice this happening no matter whether I use SBCL or >> >CLISP, and it has something to do with the console window. Whereas, >> >if I run an application delivered as an executable, the window appears >> >as you would normally expect. >> >>It looks as if no window is created at all, nothing in the taskbar. > >If it were me at this point, I would start sprinkling debug statements >around to try to find out more info. I would do this with Alastair's >hello.lisp since it is such a small amount of code to work with. >Check that REGISTERCLASSEX is not returning zero, and that >CREATEWINDOW is returning a non-null foreign pointer. Check >that his WINDOW-PROCEDURE function is getting called, >and print out the argument values, so that you can see what >events get processed. Ok, I know what you mean, but that will take some time. In nearly any other language I would have done it already, but Lisp is ... different ;-) and I have to learn so much >If those debug statements indicate everything is succeeding, >but you still don't see any window, then I would suggest installing >a Spy utility (I use Winspector Spy) so that you can check whether >the window is getting created but not being made visible. Alastair's >window classname is "hello". I have a Spy software around and can check >Is it safe to assume that you don't have any desktop mods or >virtual desktops or anything of that sort? No, nothing. I'm thinking of setting up a plain new VM just for training on Lisp development just in case any of my other environments is causing troubles. Michael Jung

1 0

Re: [graphic-forms-devel] graphic-forms
by Michael Jung 04 Sep '06

04 Sep '06

Hello Jack, At 17:55 04.09.2006, you wrote: >OK, that's helpful info. Check your taskbar and see if in fact a window >did get created but at the bottom of the desktop Z-order instead of >the top. I notice this happening no matter whether I use SBCL or >CLISP, and it has something to do with the console window. Whereas, >if I run an application delivered as an executable, the window appears >as you would normally expect. It looks as if no window is created at all, nothing in the taskbar. Regards Michael Jung

2 2

[graphic-forms-devel] graphic-forms
by Michael Jung 04 Sep '06

04 Sep '06

Hello Jack, first of all, thanks for offering help to a newbie ! Nothing unusal happens so far, everything compiles fine. When I load and start the tests it seems to evaluate and hangs, so that means it doesn't return to the normal prompt. I thought that I'm doing something wrong with that patch for sbcl, tried the demo from lisphacker.com - the same, nothing happens. Regards Michael Jung Senior Software-Architect | Designer | Consultant | Trainer Borland Delphi, LISP, Linux-Workstations, Server, Virtual-, Cluster-, Grid-Computing fon: 0163 - 517 66 02 openBC - Get together - wherever you are -> http://www.openbc.com/go/invita/5622521

2 1