Closed microbit-carlos closed 3 months ago
is this on hardware or in the simulator?
@carlosperate is this on hardware? Can CODAL update be the reason if on hardware?
This difference is on hardware.
It could be CODAL, but I can't think of something specific in the v0.2.63 to v0.2.67 transition that could cause this. We are using v0.2.67 in MicroPython and we don't have the same memory limitation (although MicroPython is using a lower level API).
The CODAL_DEFAULT_STREAM_RECORDING_MAX_LENGTH
in StreamRecording
is still set to 50K codal-core/inc/streams/StreamRecording.h#L9, which should result in the 4.5 seconds we see in MakeCode v6.
I've created a branch with MakeCode v6.0.28 and CODAL v0.2.67 (latest). The CI has created this test deployment:
On initial test it looks like it yields similar results to MakeCode v7 with the same latest CODAL. So it looks related to the CODAL update.
I've pushed branches with the in-between CODAL versions to see if we can narrow it down, but there has been several issues:
Code: 504 / 504.0 GatewayTimeout
in the console(Edit: Restarting everything that failed several times eventually worked, added the results in here to keep a single place with the summary.)
I've tried building locally the v0.2.66 branch and that worked: https://github.com/microsoft/pxt-microbit/commit/86cd987df1a4c53ad9287a2b13261041915cc539
Using pxt buildtarget --local
and pxt buildtarget --cloud
both finish successfully.
pxt ci
also run successfully.
Not sure why the pxt-buildpush
is failing with:
https://github.com/microsoft/pxt-microbit/actions/runs/10282431056/job/28454622812
waiting 8s for C++ build...
polling C++ build https://makecode.com/compile/3bec9d0618871b4b8c832c205460247f433817a995fb2b80d96a84a89416f387.json (attempt #31)
build log https://makecode.com/compile/3bec9d0618871b4b8c832c205460247f433817a995fb2b80d96a84a89416f387.log
build failed
Code: 504 / 504.0 GatewayTimeout
TypeError: Cannot read property 'length' of null
at compressHex (/home/runner/work/pxt-microbit/pxt-microbit/node_modules/pxt-core/built/pxt.js:105209:37)
at /home/runner/work/pxt-microbit/pxt-microbit/node_modules/pxt-core/built/pxt.js:105138:36
{
sha: '3bec9d0618871b4b8c832c205460247f433817a995fb2b80d96a84a89416f387'
}
at processTicksAndRejections (internal/process/task_queues.js:95:5)
at async fillExtInfoAsync (/home/runner/work/pxt-microbit/pxt-microbit/node_modules/pxt-core/built/pxt.js:1[137](https://github.com/microsoft/pxt-microbit/actions/runs/10282431056/job/28454622812#step:5:138)73:51)
at async MainPackage.getCompileOptionsAsync (/home/runner/work/pxt-microbit/pxt-microbit/node_modules/pxt-core/built/pxt.js:1[138](https://github.com/microsoft/pxt-microbit/actions/runs/10282431056/job/28454622812#step:5:139)19:33)
at async /home/runner/work/pxt-microbit/pxt-microbit/node_modules/pxt-core/built/pxt.js:162510:19
at async /home/runner/work/pxt-microbit/pxt-microbit/node_modules/pxt-core/built/pxt.js:99437:42
at async Promise.all (index 0)
at async promisePoolAsync (/home/runner/work/pxt-microbit/pxt-microbit/node_modules/pxt-core/built/pxt.js:99443:21)
at async buildTargetCoreAsync (/home/runner/work/pxt-microbit/pxt-microbit/node_modules/pxt-core/built/pxt.js:162492:5)
at async Object.ciAsync [as _callback] (/home/runner/work/pxt-microbit/pxt-microbit/node_modules/pxt-core/built/pxt.js:160821:9)
@riknoll is it possible the cloud compiler is still having sporadic issues?
(Edit: There seems to be some flakiness as well, restarting the CI jobs or trying to rebuild MakeCode project fails a few times in a row, but sometimes it succeeds.)
PR with the fix raised:
This simple programme measures how long we can record audio until hitting the default limit:
On MakeCode live it records around 4.5 seconds of audio: microbit-recording-live.hex.zip
On MakeCode beta less than 1.5 seconds: microbit-recording-beta.hex.zip