[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
Fw: [cdt-dev] MBS wizard extension debugging
|
Hey, I think I isolated what was causing the problem.
When I try to get a page property that doesn't exist, everything
degrades...
propID = "mpiIncludePath";// bogus?
obj = MBSCustomPageManager.getPageProperty(pageID, propID);
When I set propID to the "proper" value (the value that was set in the
wizardPage) things start to work
predictably... OK now I can step further...stay tuned.
Very strange.
...Beth
Beth Tibbitts (859) 243-4981 (TL 545-4981)
IBM T.J.Watson Research Center
Mailing Address: IBM Corp., 455 Park Place, Lexington, KY 40511
----- Forwarded by Beth Tibbitts/Watson/IBM on 05/31/2006 03:47 PM -----
Beth
Tibbitts/Watson/I
BM To
"CDT General developers list."
05/31/2006 03:34 <cdt-dev@xxxxxxxxxxx>
PM cc
Subject
Re: [cdt-dev] MBS wizard extension
debugging(Document link: Beth
Tibbitts)
The Runnable works with some code, just not with "more."
In particular the hunk of code i uncomment which causes it to fail is:
Object obj = MBSCustomPageManager.getPageProperty(pageID, propID);
String mpiIncludePath = null;
if (obj != null) {
mpiIncludePath = obj.toString();
}
System.out.println("Found MPI include path: " +
mpiIncludePath);
But, I'm not convinced the specific code is the culprit.
I'm on Windows XP.
I'm trying to isolate it to a smaller test case that will fail... I will
post that if I can make it small and intelligible...
...Beth
Beth Tibbitts (859) 243-4981 (TL 545-4981)
IBM T.J.Watson Research Center
Mailing Address: IBM Corp., 455 Park Place, Lexington, KY 40511
Chris Recoskie
<recoskie@xxxxxx.
com> To
Sent by: "CDT General developers list."
cdt-dev-bounces@e <cdt-dev@xxxxxxxxxxx>
clipse.org cc
Subject
05/31/2006 02:19 Re: [cdt-dev] MBS wizard extension
PM debugging
Please respond to
"CDT General
developers list."
<cdt-dev@eclipse.
org>
Without seeing the code in question I can't really help too much I'm
afraid. "Works for me."
How minimal is your Runnable? Also, what platform are you running on?
===========================
Chris Recoskie
Team Lead, IBM CDT Team
IBM Toronto
http://www.eclipse.org/cdt
Beth Tibbitts
<tibbitts@xxxxxx.
com> To
Sent by: cdt-dev@xxxxxxxxxxx
cdt-dev-bounces@e cc
clipse.org
Subject
[cdt-dev] MBS wizard extension
31/05/2006 02:09 debugging
PM
Please respond to
"CDT General
developers list."
<cdt-dev@eclipse.
org>
I am having strange problems trying to debug some MBS extensions... I added
a wizard page for C Managed Build projects, with a Runnable that runs at
the end
of the wizard, to incorporate user's selections made in my wizard page.
Thanks to Chris Recoskie for some pointers to know what to try... I'll
document all this when I get it working...
But I can't even get all of the Runnable to run...
I have had problems with cygpath.exe errors, and JVM segfaults.
I'm on Windows XP.
I have upgraded cygwin to the most recent everything, got Eclipse 3.2 RC6
(was RC4) and CDT RC2 (nightly build from 5/30) plus CDT code from head on
some projects yesterday.
Finally to try to avoid segfaults I got the Sun JVM 1.5 for 'installed JRE"
(I was using the IBM JVM).
The segfaults & "unknown bytecode" errors have stopped but now the
intermittent behavior in the wizard's execution seems to have returned...
I'm not certain that the JVM change really 'fixed' things. Just changed
the symptoms.
My wizardPage ctor always runs... I can see from a println...
but often the Whole "New C Managed Build Project" wizard doesn't run. Then
sometimes the "New C Managed Build Project" wizard runs, but my wizard page
doesn't show up.
It seems to vary depending on if I select :
(1)New.. Project then select " C > Managed Make C Project " from the first
wizard page
or
(2) New > " Managed Make C Project" directly from the context menu.
Do these take two different routes?
Route (1) seems more reliable... at least the wizard has already started so
it proceeds. But then sometimes my page isn't at the end... even though
the ctor has already run.
I can dramatically (consistently) change the behavior of the above wizard
if i simply comment out part of the code in my Runnable.
So simply having the code compiled in seems to affect whether or not the
wizard, or the wizardpage, that it's attached to (via extensionPt) actually
gets shown.
This is very strange, 'feels' like a JVM problem still, to me.
I also have discovered that it seems I need to catch all errors in my
wizard page that extends MBSCustomPage, and the Runnable, regardless of if
the compiler forces me to,
or else it dies down in the bowels of Eclipse platform code somewhere.
try/catch around entire method seems safer now. Should this be so?
But right now, I can't even get it to fail this way, since changing the
JVM.
I just switched to using Sun's JDK 1.5 for both my development version of
Eclipse, *and* the runtime JRE. Same behavior.
So, i seem to be stuck. I can't get much of anything useful to run in a
Runnable() attached to my wizard page.
Am I going crazy???
...Beth
Beth Tibbitts (859) 243-4981 (TL 545-4981)
High Productivity Tools / Parallel Tools http://eclipse.org/ptp
IBM T.J.Watson Research Center
Mailing Address: IBM Corp., 455 Park Place, Lexington, KY 40511
_______________________________________________
cdt-dev mailing list
cdt-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cdt-dev
_______________________________________________
cdt-dev mailing list
cdt-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cdt-dev
