I tried examining GDI.EXE with MiTeC EXE Explorer, and there didn't seem to be anything like a timestamp or compilation date anywhere inside. Also tried PEDUMP with the syntax you posted in the other threads, it generated no output whatsoever, even tried redirecting output to a .TXT file, and it was blank. eXeScope didn't find anything either.
GDI32.DLL did yield some results however. Each version seems to have two instances of "TimeDateStamp," but they are usually within a few seconds of each other.
GDI32.GLL version 4.00.1112 for OSR2 from Q229670:
TimeDateStamp: 398F4DE4 -> Mon Aug 07 20:01:40 2000
TimeDateStamp: 398F4DDD -> Mon Aug 07 20:01:33 2000
GDI32.DLL version 4.00.954 for 95 Original from Q229670:
TimeDateStamp: 39C2632F -> Fri Sep 15 13:58:07 2000
TimeDateStamp: 39C2632C -> Fri Sep 15 13:58:04 2000
GDI32.DLL version 4.00.955 from Q247578:
TimeDateStamp: 39D4D9EF -> Fri Sep 29 14:05:35 2000
TimeDateStamp: 39D4D9EC -> Fri Sep 29 14:05:32 2000
GDI32.DLL version 4.00.970 from MSAA:
TimeDateStamp: 32346E9C -> Mon Sep 09 15:23:08 1996
TimeDateStamp: 32091696 -> Wed Aug 07 18:20:06 1996
Going by these results it would seem that the MSAA versions are actually the oldest.
Q247578 is the newest, but it supposedly not for OSR2.
The Q229670 KB article claims it should have versions 4.01.971 for 95 Original (but doesn't?), so does that mean that Q229670 version 4.00.1112 for OSR2 is the OSR2 equivalent of Original 4.01.971? If so, then the Q229670 HotFix must contain the last version for OSR2.
The version discrepancy between the Q229670 KB article and the actual files it contains is puzzling. It should contain 4.01.971 files for 95 Original, but does not. In fact, there does not seem to be any GDI32.DLL version 4.01.971, and the only GDI.EXE with that version is in the W95EURO package and has an earlier timestamp than all of these other files.
Edited by LoneCrusader, 23 January 2011 - 02:36 PM.