Well, I thought it would be a good idea to post here first and to receive feedback from other developers before contacting support. I found a similar discussion about the older PSoC Creator.
Also, I don't like contacting support because everything gets handled behind the scenes and other developers can't benefit from it. Actually I have the impression that there are very few people posting here which is the bad sign for a very small community.
Thanks anyway, if I don't get any other replies I will have to contact support as usual.
Postings here in this forum are relatively common, not only a few. Just add the post-numbers from the first tree contributors showing about 250 posts per month.
A technical answer may take a bit longer than answers in the forum due to different time-zones and a one-to-one conversation.
Of course you are invited to keep us informed of the result.
Maybe I should have elaborated why I did send you to the Cypress support team...
Normally Creator is stable - most issues with debugging are due to power supply issues. But none of the contributors here is able to debug it (there are some Cypress Emplyoees here, though). And since you provided not more information, it looks like we cannot give you proper help to solve your problems. Opening a support case is the fastest way to get help, since you (hopefully) directly get a contact with the capability to help you.
Btw: its a good style here to post the results of a support case, so others can look at it in the future.
Update: Looks like the problem is not USB related as it also appears on the VoltageDisplay_DelSigADC example for the PSoC 5LP Kit wit PSoC Creatror 3. After halting / resuming the program a couple of times in the debugger it cannot be halted anymore, so I have to stop the debug session. I either get no error message or I get:
Error: dbg.M0015: Debugger exited unexpectedly during run. Encountered error (GDB encountered an error and had to close)
This is on PSoC Creator 3 with Win 7 64 bit. I have now tested in PSoC 2.2 on my old Win XP and everything seems to work fine there. I will now install PSoC Creator 3 on XP to test again. I have also contacted Cypress but they have not provided any hints so far.
BTW: Today I received a Creator update. If I remember correctly, before the update the debugger first stopped in some init file instead of main(). After the update this seems to be fixed.
So I checked the debugging with PSoC Creator 3 on my old Win XP Laptop, again with the original VoltageDisplay_DelSigADC example project.
After a couple of halts and resumes the debugger stops responding, i.e. it cannot be halted anymore. So it must be a bug in the Creator software or the PSoC 5 LP Kit, not my PC as the problem occurs on Win7 64bit as well as on WinXP 32 bit on completely different computers.
UPDATE: Cypress confirmed the problem, looks like a race condition in the debugger. They are working on it.
I use GCC4,41, and now 4,73, with different machine and project, write from me or others guy..
With creator 3, in debug mode i see the following problem.
A)The debug don't start on " main" routine , but in an fixed point in some part of the code.
With F10, become jump from on line to anothers without any significance.
In Window dissasembly, in effect with F10 follow a line by line without jump, but the code is unrelated with the code write on the file, that i try debugging.
B)Sometimes a code that is debugged correctly, without any problem, became also do the same things.
FOR COME BACK TO DEBUG , i begining cut some part of code.and the problem persist.
BUT, if you paste the code that , is cutted before , all come to work.
A very generator of ansiety.
C)I see that gcc, don' like some bitfield definition, like uint8 foo:1; , after this cure, the things run a little better.
D)The solution of open a case, don't work with me , the guy was able to see the problem, but the solution was, use the MDK compiler (2-3000 $).
I want share this expereience with others people, maybe some ligth come on the oscure world of gcc and creator.
Maybe some take some help from what i see.
Best regards to all
Since most of us do not have that bad experience I would suggest you to upload an example-üroject (with Creator's "Create Workspace Bundle) that shows your mentioned error.
The fact that sometimes the disassembly is out of sync with the sourcecode is a known bug and seemingly has to do with optimization compiler flags,
Yes, the optimization does cause problem and the problme with the problem with debugging has been there for a while.
One thing we do is turn all optimization off and just use the bigger chip if running out of room. ( Not ideal but not big problem for our project, may be more important for other projects that every cents counts).
I have always used level optimitation = None, and the problem is present, also in this moment.
I make the project bundle, in this project i use the keypad from Bob with some little modifie (i have a keyboard matrix with diode).
But this is influent.
I Have plenty room of rom and ram.
I become suspect that, i have some basic mode of programming, that gcc don't digest.
I hope that looking on my code, maybe come some evidence.
Any suggestion be very appreciate.
Best to all.
Can't find anything but would think that using longer variable names to include more details of the usage may help.
About the debugging issue.
I remember that most oftern happen near the end of a IF statement,
My debugger can't be paused during debugging and hence, I can't test and debug my code. I am using CY8C4124AXI-443 PSoC. Is everyone facing the same problem with mine?