gdbtrace.init does not seem to work with newer version of arm-none-eabi-gdb about orbuculum

[zyp]

Thanks, reproduced and fixed in 7809380. describeETM was missing an end at the end, so it ate the rest of the script, and the way it was indented didn't make it obvious.

[salkinium]

Weird, which platform did you try this on? Windows perhaps?
I use both the ARM (non-Python) and the xpack (Python3) GDB v12 on macOS and it works fine.

[ClemensElflein]

PopOS 22.04, so basically Ubuntu 22.04

[niklaut]

This exact problem is now also happening with my GDB after I updated the gdbtrace.init to the latest version.

[zyp]

Which platform are you on? I haven't been able to reproduce it yet.

[niklaut]

I'm on macOS 13.5.2 with xpack arm-none-eabi-gdb-py3

GNU gdb (xPack GNU Arm Embedded GCC arm64) 12.1.90.20221210-git

It's weird, cos parts of the script are shown with autocompletion:

(gdb) _s
_setAddressesEFR32MG12  _setAddressesETM4       _setAddressesIMXRT      _setAddressesNRF        _setAddressesNRF52      _setAddressesNRF53      _setAddressesSTM32      _setAddressesTM4C       _startETMv35            _startETMv4
(gdb) start
start     startETM  starti
(gdb) stop
stop     stopETM
(gdb) describeETM
(gdb) prepa

but then it stops after describeETM, so something in it breaks it. (there's a duplicate stopETM function also btw)
if I remove the describeETM function, I can call the enableSTM32TRACE function.

I can see that it configures the pins, but the orbtrace's trace led doesn't turn red. I suspect that this is a hardware issue with my wiring though.

[niklaut]

Yup, that fixed it for me, thanks!