COMPILE: Getting the source to compile

[Peter Triggs]

Duh, that was it. RalphTricky and Vovan mentioned back on page 1 of this thread that flex and bison don't like paths with embedded spaces. How could I have not noticed that "Program Files" has an embedded space.

[child of Thor]

Originally posted by Peter Triggs
Duh, that was it. RalphTricky and Vovan mentioned back on page 1 of this thread that flex and bison don't like paths with embedded spaces. How could I have not noticed that "Program Files" has an embedded space.

And i guess here lies the big problem people will face converting any code over to different versions - its a lot of code to get through, and one little thing can be a huge headache! Well done on spotting it

[The Big Mc]

CivCTP_dbg.exe - 0 error(s), 0 warning(s)

Success

HAS anybody else seen the button with no words in it on the main Manu if you click on it. it goes to a sprite editor thing which does not seem to be of any use until you look into the civ3 to ctp unit conversion thread it will allow us to find the hotspots easier

[Martin Gühmann]

Originally posted by The Big Mc
HAS anybody else seen the button with no words in it on the main Manu if you click on it. it goes to a sprite editor thing which does not seem to be of any use until you look into the civ3 to ctp unit conversion thread it will allow us to find the hotspots easier

If I remember correctly the button says "Test".

Good to know you got it to compile.

-Martin

[The Big Mc]

it allows you to do more you can open a spr file based on its sprit number then find the perfect hotspot location

[J Bytheway]

I have just updated my source (after a long break) and have been trying to compile it. There's something very strange about the compilation of the .cdb files (using dbgen) - each one stops the compilation process with an error, but the compilation actually appears to have completed successfully. Does anyone know how the VS6 IDE determines whether a custom build step suffered an error, since it seems to be drawing the wrong conclusion in this case.

[Martin Gühmann]

Originally posted by J Bytheway
I have just updated my source (after a long break) and have been trying to compile it. There's something very strange about the compilation of the .cdb files (using dbgen) - each one stops the compilation process with an error, but the compilation actually appears to have completed successfully. Does anyone know how the VS6 IDE determines whether a custom build step suffered an error, since it seems to be drawing the wrong conclusion in this case.

Probably because there are already output files but unfortunatly these files do have the wrong content, I provided my files generated by dbgen but Mr. Baggins didn't and therefore you don't have the government functionality in.

I suspect you didn't compile a debug.exe, you have to do this to generate the output of dbgen first.

-Martin

[J Bytheway]

No, this happened during compilation of Win32 Debug. The problem wasn't that it wasn't working, just that it thought it wasn't working when in fact it was... It's almost certainly due to MrBaggins changes, but it's still strange...

[Fromafar]

Did you happen to do a Release or Final build in between 2 Debug builds? For me, this sequence will almost always cause a failure report for each dbgen generated file.

Because no input (.cdb) files have been changed, it should not be necessary to compile anything. The object files of the the different build types are in different directories. But maybe the update of the .stamp file is causing the problem?
Stubbornly pressing F7, or doing a full build, will eventually create a correct executable.

Anyway, I don't think it is something we have changed. I had this problem with the original Activision package as well.

[J Bytheway]

I can't remember for sure but I think what I did was:
Switch to Win32 - Final
Clean
Build (get errors because I don't have the stuff you get from a debug build)
Switch to Win32 - Debug
Clean
Build (get all the above-mentioned errors, but get through them by stubbornly pressing F7)
Switch to Win32 - Final
Build (Success with no problems)

Perhaps the problem is because dbgen returns a non-zero value when it finishes with no errors (a missing return 0 statement somewhere...). I'll have a look and see if I can find that out.

[J Bytheway]

I've just cleaned and recompiled again, and it works fine if you don't do a (failed) Final compile before the Debug compile, so I guess that's the problem.

[Locutus]

Edit: nm, I'm an idiot

[J Bytheway]

Edit: nm, he's an idiot

[Solver]

Yeah, indeed, from my observations: Doing a Debug, and then a Final build is fine. If you first do a failed Final build, then you later need to do a Debug build, clean the Final build, and rebuild Final again, and you wil have to press F7 quite a number of times, till it shuts up about the error.

[Ekmek]

Bump, I made need this thread...