- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Howdy:
I'm firing up a CVF(6.5.0) EXE through CreateProcess redirecting standard output and standard error devices to the same file. The EXE then explicitly links to a CVF dll which has suspect code. Everything works fine with the default error handler - all output, including stack trace go to the file. However, when I modify the dll to wrap the suspect code in a C try-except block (where I call a handler that calls TRACEBACKQQ) no stack trace is produced. The only way I've managed to get the trace is by using the FOR_DIAGNOSTIC_LOG_FILE environment variable. This is bad news because I need to see the trace(s) in the context of outputin the log file.
Any ideas?
Keith
Link Copied
2 Replies
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I think I'll write myself a letter,
And make believe it came from you ...
Sure is lonely out here.
Okay, TRACEBACKQQ doesn't seem to turn anyone's crank. (Me too).
Would greatly appreciate some kind of advice here though - dealing with it through the environment variable directing it to a separate file is going to turn a fairly tidy piece of code into a coat-hanger and duct tape job.
Is this something I should be directing to CVF support?
Basically the default error handler has some kind of magic that gets its output to STDERROR and mine (that does a straight call to TRACEBACKQQ) doesn't.
Waiting for a nod in the some kind of direction or at least a "give-it-up" note.
All dressed up and no place to go ...
Keith
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Please forward your letter to Intel Premier Support. It will make its way to the engineers who know this stuff (I don't.)

Reply
Topic Options
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page