Suspected graphics linking problem in mixed Fortran/C program

Ok,

I tried running it from the SF console, rather than from the windows console, and it runs fine. If I try to run it from the windows console, nothing happens at all, and it goes back to the command prompt. Thoughts?

Thanks,
Jason

Yes, but it runs fine from the SF console. So using the debugger won't tell me why it doesn't work using the windows command prompt. It won't bomb from the SF console.

Ok, If I start the debugger, does that run it from the SF console or the windows command prompt? How do I ensure that it's being launched from the windows command prompt and stepping through the debugger?

When the task manager can't kill it, you have to kill GDB.exe.
That does the trick, and avoids the need for a reboot.
---
John

Ok, I've done that and now I'm noticing something new. Currently, I have it set to launch in an external console, but that doesn't seem to actually show the console if I start the debugger. My first read statements happen in a subroutine. If I try to step over that subroutine, SF crashes. If I step into the subroutine up to the first read statement, then step over the read statement, I get this message, but only if I open the SF console first (since a console isn't coming up on its own otherwise):
"*stopped,reason="end-stepping-range",frame={addr="0x0040c113",func="askc",args=[{name="prompt",value="<error reading variable: Cannot access memory at address 0x49b398>"},{name="comand",value="<error reading variable: Cannot access memory at address 0x28ed49>"},{name="cargs",value="<error reading variable: Cannot access memory at address 0x28ecc9>"},{name="_prompt",value="8"},{name="_prompt@entry",value="8"},{name="_comand",value="7"},{name="_comand@entry",value="7"},{name="_cargs",value="128"},{name="_cargs@entry",value="128"}],file=".\\src\\userio.f",fullname="P:\\Tools\\Xrotor\\src\\userio.f",line="132"},thread-id="1",stopped-threads="all""

I *think this is telling me the variable COMAND is not long enough for the data that's trying to get entered into it. If I remember right, Linux and Unix only do a carriage return or line feed at the end of a line, but Windows does a carriage return and line feed at the end of a line. Does this seem plausible? I'm going to try some changes on that front.

Jason,

Before you continue any further, let's try to clear up what you're working with. 

1. Which version of Simply Fortran are you currently using?

2. Does your program use the console for any input/output whatsoever?  If it does, you need to leave Windows GUI unchecked.   That box should be used only if you have absolutely no console input/output.

If your program is using a "Read *," or anything like that, you'll need to ensure that Windows GUI is not checked.  The long message you're seeing:

is due to a bug in Simply Fortran (didn't know about that one) where you're seeing the communications between Simply Fortran and the debugger.  I'll have to look into further suppressing this output.

Let me know the answer to the above questions.  I'm interested in what's actually happening with your code.

Jason,

I think I'm starting to understand what you're asking.  Actually, the Windows GUI box does not cause any linking against Windows libraries to occur.  You'll need to open the Project Options dialog in your executable project, switch to the Compiler Flags tab, and add the following:

-lgdi32

You might have to add some other flags, but we'll tackle those as compiler errors arise.  Click Ok, and save your project. 

Next, perform a Clean from the Build menu.  Finally, click Build Now! in the Build menu.  Your program will either compile or, more likely, complain about another linker error.

Out of curiosity, did you compile "libPlot.lib" using Simply Fortran?  If the library was compiled for/by another compiler, it may not be compatible with Simply Fortran's compiler.  However, if you've gotten as far as seeing undefined references to BitBlt, I think everything is working okay.

Let me know what you get next based on these comments.

Jason,

That's great to hear!  When you do get a chance, maybe you could provide a quick screenshot to show us the graphics portion.  I know I'd be interested in seeing it, and I think a good portion of users are interested in any sort of graphical programs written in Fortran.

John,

You would need to host the image elsewhere and then provide a link.  If you look at the help entry on images, you'll see that it is providing a complete URL to the image online.

Usually when I post images here, I first upload them to Imgur, which is simple, fast, and uncluttered.  You don't need an account or anything, so it makes posting an image easy.  They'll provide a per-image link to manage it in case you wish to delete it later.

Jason, it doesn't display.
Also, may I suggest you place it in the forum under the new thrread I've just created for GUI screenshots?
---
John

Yes, it's an airfoil analysis program from MIT and it couples boundary layer modifications into the panel model to make it able to do viscous analysis (panel methods are inherently inviscid - i.e.: no boundary layer). It's available to the public under the GPL here:

http://web.mit.edu/drela/Public/web/xfoil/

I wouldn't need to compile it on my own, except I want to make some additions to it.

Thanks for the help!
Jason