Forums

"Find the file named A"

Started by Corey, Rick December 26, 2002
Hi

When I run the CW debugger, I often get an error message in a file-select
dialog box: "Find the file named A". To appease it, I double-click on any
*.c file. The error repeats 2-5 times. The default directory to search in
seems to be a dir that I haven't used for development for almost two months.

Then, it changes to "Find the file named Ffre...". I find that to make this
one go away without locking up the IDE, I should click on "Cancel" - usually
either three times, or six times.

Often/usually, if I get back into the IDE, my stack crawl window will show
that I am inside two invocations of "@DummyFn1", with two invocations of
"ConfigInterruptVector" above that (I almost never see names of my own
functions when I get these errors).

We use many SDK functions, but are using CANbus without the SDK CANbus
driver. All my self-configured interrupts are at level 1 (non-nestable),
yet when I looked at the "Change of Flow FIFO dump", it showed two lines of
"FarchExitNestedInterruptCommon" on top, followed by FQTimerSuperISRD0,
followed by 6 lines of "F@DummyFn1". I thought I was NOT using nested
interrupts, and that my code was not using timers at all (I have undef-ed
INCLUDE_QUAD_TIMER and INCLUDE_TIMER.

Sometimes, after "Finding" A and "Canceling" Ffre, I get another problem
when trying to continue: "Unhandled Exception C0000005 at 00d5cdea" .
Once, I followed this back into Visual C (as if trying to debug the IDE),
and saw references to "MWCORE!" and "MWCommBase.dll" and thread 458.

Other people in my row are using the same Metrowerks IDE, but never see this
error - yet they are using my code in addition to their own. I have had
some hardware problems with my PC, and have been hoping that these are
symptoms of a sickly PC, rather than an IDE problem.

The few times I've checked, using stackcheckSizeUsed(), I have not seemed to
be overflowing my stack, but that would be my next avenue to pursue. This
eval board is rev 0115.

I tend to see the "Find the file named A" message more often while I either
have bugs, or the code is acting weirdly. Then the error messages and
"bugs" seem to go away for a few days, or come back, without correlating
clearly with code or test bed changes. Sometimes rebooting my PC will make
the error message stop recurring for a while. I can't yet rule out an
intermittent bug in my '803 code or some test condition being the triggering
factor.

Does anyone else see this error, and/or know what tends to cause it, or have
any suggestions to try? Often, when I get the error message, I can't get
back into the debugger to look at the aftermath ("program not responding").
I've had this off and on for months, but it is getting more frequent and
more annoying. I don't know whether to blame the IDE or my code (or both).

Thanks in advance!

Rick Corey, Software Development
DPC Instrument Systems Division, Flanders NJ



Rick,

To get rid of this problem immediately, please go and rebuild your
MSL library project which you can find at the following path:

{install path}\M56800 Support\msl\MSL_C\DSP_56800\Project\

If you open up this project (MSL C 56800.mcp) and rebuild this
project it will hopefully remove this problem from your machine. I
believe this problem was addressed in the 5.0.3 patch as it pertains
to the MSL libraries so you might also want to go ahead and download
this patch as well at:

http://www.metrowerks.com/MW/Support/Download/default.htm

Regards,
John

--- In , "Corey, Rick" <rcorey@d...> wrote:
> Hi
>
> When I run the CW debugger, I often get an error message in a file-
select
> dialog box: "Find the file named A". To appease it, I double-click
on any
> *.c file. The error repeats 2-5 times. The default directory to
search in
> seems to be a dir that I haven't used for development for almost
two months.
>
> Then, it changes to "Find the file named Ffre...". I find that to
make this
> one go away without locking up the IDE, I should click on "Cancel" -
usually
> either three times, or six times.
>
> Often/usually, if I get back into the IDE, my stack crawl window
will show
> that I am inside two invocations of "@DummyFn1", with two
invocations of
> "ConfigInterruptVector" above that (I almost never see names of my
own
> functions when I get these errors).
>
> We use many SDK functions, but are using CANbus without the SDK
CANbus
> driver. All my self-configured interrupts are at level 1 (non-
nestable),
> yet when I looked at the "Change of Flow FIFO dump", it showed two
lines of
> "FarchExitNestedInterruptCommon" on top, followed by
FQTimerSuperISRD0,
> followed by 6 lines of "F@DummyFn1". I thought I was NOT using
nested
> interrupts, and that my code was not using timers at all (I have
undef-ed
> INCLUDE_QUAD_TIMER and INCLUDE_TIMER.
>
> Sometimes, after "Finding" A and "Canceling" Ffre, I get another
problem
> when trying to continue: "Unhandled Exception C0000005 at
00d5cdea" .
> Once, I followed this back into Visual C (as if trying to debug the
IDE),
> and saw references to "MWCORE!" and "MWCommBase.dll" and thread 458.
>
> Other people in my row are using the same Metrowerks IDE, but never
see this
> error - yet they are using my code in addition to their own. I
have had
> some hardware problems with my PC, and have been hoping that these
are
> symptoms of a sickly PC, rather than an IDE problem.
>
> The few times I've checked, using stackcheckSizeUsed(), I have not
seemed to
> be overflowing my stack, but that would be my next avenue to
pursue. This
> eval board is rev 0115.
>
> I tend to see the "Find the file named A" message more often while
I either
> have bugs, or the code is acting weirdly. Then the error messages
and
> "bugs" seem to go away for a few days, or come back, without
correlating
> clearly with code or test bed changes. Sometimes rebooting my PC
will make
> the error message stop recurring for a while. I can't yet rule out
an
> intermittent bug in my '803 code or some test condition being the
triggering
> factor.
>
> Does anyone else see this error, and/or know what tends to cause
it, or have
> any suggestions to try? Often, when I get the error message, I
can't get
> back into the debugger to look at the aftermath ("program not
responding").
> I've had this off and on for months, but it is getting more
frequent and
> more annoying. I don't know whether to blame the IDE or my code
(or both).
>
> Thanks in advance!
>
> Rick Corey, Software Development
> DPC Instrument Systems Division, Flanders NJ


--- In , "Corey, Rick" <rcorey@d...> wrote:
> Hi
>
> When I run the CW debugger, I often get an error message in a file-
select
> dialog box: "Find the file named A". To appease it, I double-click
on any
> *.c file. The error repeats 2-5 times. The default directory to
search in
> seems to be a dir that I haven't used for development for almost
two months.
>
> Then, it changes to "Find the file named Ffre...". I find that to
make this
> one go away without locking up the IDE, I should click on "Cancel" -
usually
> either three times, or six times.

I have experienced a similar problem when my C/W project is buried
too deeply within my directory structure. I believe your problem
will go away if you move the project to the root of your hard drive.
Then the fully qualified path names to your files will will be
shorter. > Often/usually, if I get back into the IDE, my stack crawl window
will show
> that I am inside two invocations of "@DummyFn1", with two
invocations of
> "ConfigInterruptVector" above that (I almost never see names of my
own
> functions when I get these errors).
>
> We use many SDK functions, but are using CANbus without the SDK
CANbus
> driver. All my self-configured interrupts are at level 1 (non-
nestable),
> yet when I looked at the "Change of Flow FIFO dump", it showed two
lines of
> "FarchExitNestedInterruptCommon" on top, followed by
FQTimerSuperISRD0,
> followed by 6 lines of "F@DummyFn1". I thought I was NOT using
nested
> interrupts, and that my code was not using timers at all (I have
undef-ed
> INCLUDE_QUAD_TIMER and INCLUDE_TIMER.
>
> Sometimes, after "Finding" A and "Canceling" Ffre, I get another
problem
> when trying to continue: "Unhandled Exception C0000005 at
00d5cdea" .
> Once, I followed this back into Visual C (as if trying to debug the
IDE),
> and saw references to "MWCORE!" and "MWCommBase.dll" and thread 458.
>
> Other people in my row are using the same Metrowerks IDE, but never
see this
> error - yet they are using my code in addition to their own. I
have had
> some hardware problems with my PC, and have been hoping that these
are
> symptoms of a sickly PC, rather than an IDE problem.
>
> The few times I've checked, using stackcheckSizeUsed(), I have not
seemed to
> be overflowing my stack, but that would be my next avenue to
pursue. This
> eval board is rev 0115.
>
> I tend to see the "Find the file named A" message more often while
I either
> have bugs, or the code is acting weirdly. Then the error messages
and
> "bugs" seem to go away for a few days, or come back, without
correlating
> clearly with code or test bed changes. Sometimes rebooting my PC
will make
> the error message stop recurring for a while. I can't yet rule out
an
> intermittent bug in my '803 code or some test condition being the
triggering
> factor.
>
> Does anyone else see this error, and/or know what tends to cause
it, or have
> any suggestions to try? Often, when I get the error message, I
can't get
> back into the debugger to look at the aftermath ("program not
responding").
> I've had this off and on for months, but it is getting more
frequent and
> more annoying. I don't know whether to blame the IDE or my code
(or both).
>
> Thanks in advance!
>
> Rick Corey, Software Development
> DPC Instrument Systems Division, Flanders NJ




Hi Rick,

Yes, we (every developer here) see this bug all the time, so you're not alone in
seeing it. We didn't report it because it wasn't a "critical bug report", and
it very seldom causes a fatal error. It's just annoying. We also often see 3
or 4 files named "A" in the file selection area of the Global Variables window.

As for the bad stack frame information, this is a bug that Metrowerks hopefully
will have fixed in the next major release. You can "fix" it if you select
optimization level 0, but for our applications this isn't possible, because we
run out of P Flash on the '807 chip, which has the maximum (60K) amount of P
Flash that is possible in the 56800 family. So, we have to live with no stack
trace ability for now. You should also be aware that this bug often causes the
local variables window to show nonsense values for the function's local
variables (which of course are saved on the stack). This caused us a lot of
wasted time until we realized what was happening. When we reported this
problem, we were told that it is "normal" behavior! Well, I can tell you that
this is the very first IDE I have ever used, that gave you nonsense results
unless you totally disabled all optimizations. And I have used a LOT of IDEs in
the last 26+ years.

Art -----Original Message-----
From: Corey, Rick [mailto:]
Sent: Thursday, December 26, 2002 7:26 AM
To: '
Subject: [motoroladsp] "Find the file named A" Hi

When I run the CW debugger, I often get an error message in a file-select
dialog box: "Find the file named A". To appease it, I double-click on any
*.c file. The error repeats 2-5 times. The default directory to search in
seems to be a dir that I haven't used for development for almost two months.

Then, it changes to "Find the file named Ffre...". I find that to make this
one go away without locking up the IDE, I should click on "Cancel" - usually
either three times, or six times.

Often/usually, if I get back into the IDE, my stack crawl window will show
that I am inside two invocations of "@DummyFn1", with two invocations of
"ConfigInterruptVector" above that (I almost never see names of my own
functions when I get these errors).

We use many SDK functions, but are using CANbus without the SDK CANbus
driver. All my self-configured interrupts are at level 1 (non-nestable),
yet when I looked at the "Change of Flow FIFO dump", it showed two lines of
"FarchExitNestedInterruptCommon" on top, followed by FQTimerSuperISRD0,
followed by 6 lines of "F@DummyFn1". I thought I was NOT using nested
interrupts, and that my code was not using timers at all (I have undef-ed
INCLUDE_QUAD_TIMER and INCLUDE_TIMER.

Sometimes, after "Finding" A and "Canceling" Ffre, I get another problem
when trying to continue: "Unhandled Exception C0000005 at 00d5cdea" .
Once, I followed this back into Visual C (as if trying to debug the IDE),
and saw references to "MWCORE!" and "MWCommBase.dll" and thread 458.

Other people in my row are using the same Metrowerks IDE, but never see this
error - yet they are using my code in addition to their own. I have had
some hardware problems with my PC, and have been hoping that these are
symptoms of a sickly PC, rather than an IDE problem.

The few times I've checked, using stackcheckSizeUsed(), I have not seemed to
be overflowing my stack, but that would be my next avenue to pursue. This
eval board is rev 0115.

I tend to see the "Find the file named A" message more often while I either
have bugs, or the code is acting weirdly. Then the error messages and
"bugs" seem to go away for a few days, or come back, without correlating
clearly with code or test bed changes. Sometimes rebooting my PC will make
the error message stop recurring for a while. I can't yet rule out an
intermittent bug in my '803 code or some test condition being the triggering
factor.

Does anyone else see this error, and/or know what tends to cause it, or have
any suggestions to try? Often, when I get the error message, I can't get
back into the debugger to look at the aftermath ("program not responding").
I've had this off and on for months, but it is getting more frequent and
more annoying. I don't know whether to blame the IDE or my code (or both).

Thanks in advance!

Rick Corey, Software Development
DPC Instrument Systems Division, Flanders NJ
_____________________________________
Note: If you do a simple "reply" with your email client, only the author of this
message will receive your answer. You need to do a "reply all" if you want your
answer to be distributed to the entire group.

_____________________________________
About this discussion group:

To Join:

To Post:

To Leave:

Archives: http://www.yahoogroups.com/group/motoroladsp

More Groups: http://www.dsprelated.com/groups.php3 ">http://docs.yahoo.com/info/terms/



Art Johnson wrote:

> Hi Rick,
>
> Yes, we (every developer here) see this bug all the time, so you're not alone
> in seeing it. We didn't report it because it wasn't a "critical bug report",
> and it very seldom causes a fatal error. It's just annoying. We also often
> see 3 or 4 files named "A" in the file selection area of the Global Variables
> window.

We have narrowed this down to the libraries needing rebuilding. I don't
know positively that this is not a Metrowerks error but it appears not to be
since it seems that the libraries and headers are out of sync.

There are many ways this can be accomplished by the end developer especially
when they have to use back ups. Ron

--
Free Programming Courses at CodeWarrior U
<http://www.codewarrioru.com>

Metrowerks, maker of CodeWarrior - "Software Starts Here"
Ron Liechty - - http://www.metrowerks.com




<red faced>

This is indeed our problem...

what happened:
5.02 libs were built on local build machine but in location with longer path
names than the usual local location. Since the entire path of the file
location is stored, 5.02 path info was coincidentally truncated for most (if
not all) asm files and the file names were lost. The build location (also
coincidentally) did not result in too-long paths for compiler-stored paths
(which do have a limit as well).

So the solutioni s if you rebuild the libraries from 5.0.2 it fixes this,
the 5.0.3 should not have the problem and we will lift the cause of the
problem totally with the 5.1 version.

I apologize for my erroneous conclusion, I'm too used to desktop
programmers.

Ron >> Yes, we (every developer here) see this bug all the time, so you're not alone
>> in seeing it. We didn't report it because it wasn't a "critical bug report",
>> and it very seldom causes a fatal error. It's just annoying. We also often
>> see 3 or 4 files named "A" in the file selection area of the Global Variables
>> window.
>
> We have narrowed this down to the libraries needing rebuilding. I don't
> know positively that this is not a Metrowerks error but it appears not to be
> since it seems that the libraries and headers are out of sync.
>
> There are many ways this can be accomplished by the end developer especially
> when they have to use back ups.

--
Free Programming Courses at CodeWarrior U
<http://www.codewarrioru.com>

Metrowerks, maker of CodeWarrior - "Software Starts Here"
Ron Liechty - - http://www.metrowerks.com