miss the file bscmake when compile TGB 1.7.5

#1 A few things to try:

Are you loading the correct solution? /engine/VisualStudio2005/...
Try a release build.
Make sure you have TGB and your game closed when building.
Try Rebuild Solution.
Get the most up to date platform sdk from Microsoft.

#2 Ok, Thanks! I had been solved this problem, it lies in the platform sdk.

#3 i had the same mistake.but what's means of "platform SDk"

#4 The Windows Platform SDK.

Reference: http://en.wikipedia.org/wiki/Microsoft_Windows_SDK

#5 WTH?

I'll try RE-installing the platform SDK. However, I did a build last week, and THIS week is when I see the error?

*sigh*

[edit:

Digging through the library references the project description for TGB holds, very first library reference is "../../Lib/SDL/win32". Odd, there IS no win32 directory in there.

As for installing the platform SDK, I'm a bit nervous about tossing in the Win*7* SDK atop VS2005 which already has the WinSDK installed. Hope installing it doesn't mean I'm going to be un&re-installing VS2005.]

[edit2:

I re-installed the SDK to no avail.

I did some more digging and tracing, searching for the complained-of strings in the platform SDK. They were there, alright -- defined in winUser.h ... so, why weren't they being included.

The needed #defines are enclosed within some conditional #ifdef that check _WIN32_WINNT to see if it is greater than or equal to 0x0400. OK, I go do some confirming research on the 'Net and confirm what, yes, this is what it looks like. Not sure what got knocked out of whack since I was able to compile before, but it appears that a fairly standard version-targeting setting got knocked out of the project.

No problem, easy enough to force it. I just drop this in as a test inside TGB's platformWin32.h, right before "#include <windows.h>":

// test modification
#define WINVER 0x0500
#define _WIN32_WINNT 0x0500
// end test modification



SUDDENLY EVERYTHING COMPILES OK.]


[edit 3:

ARGH.

Use 0x0501.

0x0500 compiles fine, but generates an executable that dies upon launch.

*sigh*

On the upside, I'm now 20% less trepidatious about editing the engine code. I'd BEGUN recompiling to edit a console method into simBase. I did. It worked. Then I found out another way to do what I'd wanted to without that custom method. (I created "getDynamicFieldValue" -- symmetrical with "getFieldValue". I can get at that value with "getFieldValue" so long as I specify a dynamic name rather than an index. Oh, well, it was worth the experience, however frustrating.]

#6 Hmm. This is a more interesting problem:

It appears 0x0500 was just as valid as 0x0501. Rather, the behavior I'm seeing is that a recompiled TGB -- stock source files with the above-mentioned test modification -- will indeed run just fine. However, the executable will not be able to launch a second executable, as TGB is designed to do when you hit the "play level" button. Looking at the console log, it seems that the second executable isn't able to create a canvas -- after which of course since it can't do a damned thing without the canvas it shuts down. Interesting, because that same executable has no difficulty if launched on its own.

Anyone else out there have an idea what might be going wrong? Kinda makes using the builder a little difficult, though for my present work I'm mostly going at it via Torsion to build some library code I'll eventually need some of the things TGB can demonstrate.

#7 @liangguoqiang: can you post that SDK link, i have same problem with you (and i also use windows xp sp3)

#8 @Playboom
Windows SDK