[M3devel] [M3commit] CVS Update: cm3

[Tony Hosking]

Do we really want to build all the Windows interfaces even on non- 
Windows hosts?  Yuck.

On Jan 22, 2008, at 5:36 PM, Jay Krell wrote:

> CVSROOT:	/usr/cvs
> Changes by:	jkrell at birch.	08/01/22 17:36:01
>
> Modified files:
> 	cm3/m3-sys/cm3/src/: M3Backend.m3 m3makefile
> 	cm3/scripts/: backup-pkgs.sh boot-cm3-build-on-target.sh
> 	              boot-cm3-core.sh boot-cm3-with-m3.sh
> 	              copy-bootarchives.sh do-cm3-core.sh
> 	              make-bin-dist-min.sh pack-crossbuild.sh pkginfo.sh
> Removed files:
> 	cm3/m3-sys/cm3/src/: M3BackPosix.m3 M3BackWin32.m3
>
> Log message:
> 	put integrated backend into all hosts, so that cross builds work
> 	a bit more; built on PPC_DARWIN (ie: built on Posix, where it's
> 	an actual diff; what this will enable for me is a "semi-cros"
> 	from a NT386GNU cm3.exe to NT386/NT386MINGNU. And it's fairly cheap,
> 	the integrated backend is nothing compared to cm3cg.
> 	did not run all the .sh files, just upgrade.sh
> 	Note that m3staloneback is relatively unused, probably for  
> debugging, left alone.
> 	mklib should come in as well for cross purposes but left that  
> alone too.
> 	There are warnings in WinDef.m3 about <*WINAPI*> on function pointer
> 	types. Perhaps they can be deferred and only trigger if the types
> 	are used? If the types are called? ie: make the pragma understood,  
> but
> 	don't support calling using calling conventions not supported by  
> target