SaabCentral Forums banner

21 - 23 of 23 Posts

·
Registered
Joined
·
17,204 Posts
Ah, wait, maybe my question does still stand. The workaround is good for the single stage procedures. When I look up a code like B2441 (an issue I currently have), there's a prompt to select a specific code (B2441, B2442, B2443). Once you select the specific code you go into fault trace - but there's no way to get directly back to that page. So, the workaround doesn't really get to execute directly - you get sent back to the start.
 

·
Registered
Joined
·
392 Posts
Yeah, there's no telling how the WIS code has been altered or patched since it was in its official original form, when it presumably worked. I did a quick symbol scan through the binaries, and I can see where image fetch could hiccup. ISOView isn't the program called to fetch an archive member (which is what's really failing). They do some things with a UNIX program (apparently ported and compiled into the Visual C/C++ code) called funzip to open the archive, and then it looks like other code steps through the archive till it finds a match for the desired member, or maybe vice versa. I was never a Windows development guy, though, just other stuff. If it were an ECU or a *NIX app, I'd be better equipped to fix it. It could be something related to faster multicore machines today vs. what they had when they wrote the original T5 WIS code in the early 90s. It might always work on a 486 or old old Pentium running Windows 98 or even NT/XP. ;)
 

·
Registered
Joined
·
392 Posts
FWIW, if there IS a Windows person out here who wants to play with this here's what I think I know:

There's a possible race condition between threads either when opening or indexing archives (.zip). Funzip has been ported from *NIX, embedded in the VC++ source, and possibly modified to do at least one of these functions.

These are SOBs to fix sometimes, and even worse w/o source code.
 
21 - 23 of 23 Posts
Top