murdock: enable nrf52dk for on-hardware tests #9095

kaspar030 · 2018-05-07T15:19:57Z

Contribution description

This PR enables on-hardware tests on two nrf52dk boards.

nrf52dk uses $(HEXFILE) as flashfile. One of the commits makes an exception for this. IMO, this is ugly, but will go away as soon as #8838 is merged.

Issues/PRs references

#8838

cladmi · 2018-05-09T16:52:24Z

Looks good.
Still funny that the current hack is set FLASHFILE:=$(HEXFILE) when it's still a .bin :D

Does the patch for the test work ? it could help for other shell based tests that from time to time fail on samr21.

I tried running it but it currently failed, so waiting on your input for re-running it.

tcschmidt · 2018-07-04T19:14:19Z

ping @MrKevinWeiss

MrKevinWeiss

I guess you are just waiting for #8838 to be merged so it gets done cleanly.

MrKevinWeiss · 2018-07-05T07:07:19Z

tests/ps_schedstatistics/tests/01-run.py

@@ -40,8 +41,10 @@ def _check_startup(child):

 def _check_help(child):
    child.sendline('')
-    child.expect('>')
+    child.expect('> ')


Instead of adding sleep can you just child.expect('> ', timeout=0.1)?

No, that would make pexpect give '> ' more time, right? I intended to add a slight delay after receiving the prompt, before sending the next line.

MrKevinWeiss · 2018-07-05T07:07:57Z

tests/ps_schedstatistics/tests/01-run.py

@@ -40,8 +41,10 @@ def _check_startup(child):

 def _check_help(child):
    child.sendline('')
-    child.expect('>')
+    child.expect('> ')
+    time.sleep(0.1)
    child.sendline('help')


tcschmidt · 2018-11-16T11:41:18Z

@MrKevinWeiss @smlng is this PR still needed given the establishing HIL testing technology and deployment?

MrKevinWeiss · 2018-11-19T10:14:57Z

@kaspar030 Do you have the murdock workers connected? Now that we have the compile_and_test_for_board any maintainer can run these tests on there desk. Also I found lots of failures that we due to things unrelated to the test (ie. libfixmath, xtimer_periodic_wakeup). Plus some cases where the flasher either doesn't work or get held up and the only way to flash is by pressing the reset button at the right time.

How would you like to proceed?

kaspar030 · 2018-11-26T10:43:25Z

Do you have the murdock workers connected?

One is connected, but somehow the incoming UART is broken. I haven't gotten around to debugging that, yet.

MrKevinWeiss · 2018-11-26T11:00:37Z

I am not sure but I also think IoT labs supports a whole bunch of boards now. They may have something available.

kaspar030 · 2018-11-26T11:09:14Z

I am not sure but I also think IoT labs supports a whole bunch of boards now. They may have something available.
Now that we have the compile_and_test_for_board any maintainer can run these tests on there desk.
is this PR still needed given the establishing HIL testing technology and deployment?

All of those don't support running tests per-PR as part of the CI.

emmanuelsearch · 2018-11-26T11:13:53Z

@kaspar030 I am in the office today if there's anyway to physically help with the attached nrf52dk

kaspar030 · 2018-11-26T11:29:29Z

@kaspar030 I am in the office today if there's anyway to physically help with the attached nrf52dk

Yes, if you can find a second board, you could replace the nrf that's attached to the Pi. It might be a hardware problem. The nrf is connected to the Pi named "pi-36f90aef".

emmanuelsearch · 2018-11-26T11:45:18Z

@kaspar030 can you try again? I think the USB cable was faulty. I changed to a new one.

kaspar030 · 2018-11-26T11:49:58Z

@kaspar030 can you try again? I think the USB cable was faulty. I changed to a new one.

Seems to work. Thanks!

kaspar030 · 2019-02-27T13:23:00Z

This is not WIP anymore. Anyone?

cladmi · 2019-03-06T11:30:07Z

The last commit with the murdoc.inc.mk can be discarded. The value is found automatically with the current master. The test could be re-run after.

BOARD=nrf52dk make --no-print-directory -C examples/hello-world/ info-debug-variable-FLASHFILE
/home/harter/work/git/RIOT/examples/hello-world/bin/nrf52dk/hello-world.bin

What did you do for the JLink update ? I may need the same thing as sometime my jlink uart is mixed and could be added in the doc (other than this PR but asking as it was mentioned here).

kaspar030 · 2019-03-06T11:33:58Z

The last commit with the murdoc.inc.mk can be discarded.

Nice. rebased & dropped that commit.

kaspar030 · 2019-03-06T12:21:39Z

What did you do for the JLink update ?

It was updated automatically when I flashed the board with JLink on my workstation, which had a very recent JLinkExe (I think from this year even).

cladmi · 2019-03-06T12:39:18Z

Ok, so my board may have been updated already.

When looking in the murdock output, some files are not recognized as being utf-8 and are forced to be downloaded:

https://ci.riot-os.org/RIOT-OS/RIOT/9095/e3729be0358ec5a5e952c254fc97867c4acd85a4/output/run_test/tests/ssp/nrf52dk:gnu.txt

I get a file download.It opens as chinese in gedit, but gvim handles it correctly.

When opening the url for samr21-xpro it works correctly.

https://ci.riot-os.org/RIOT-OS/RIOT/9095/e3729be0358ec5a5e952c254fc97867c4acd85a4/output/run_test/tests/ssp/samr21-xpro:gnu.txt

cladmi · 2019-03-06T12:46:46Z

"file" does not recognize the file as utf-8 It looks like there are ^@ characters that it does not like. Apparently null characters https://en.wikipedia.org/wiki/Null_character

kaspar030 · 2019-03-06T12:57:20Z

"file" does not recognize the file as utf-8 It looks like there are ^@ characters that it does not like. Apparently null characters https://en.wikipedia.org/wiki/Null_character

At least it is defined how Firefox does this:

From here:

When the Content-Type sent by the server is one of (case-sensitively)

    text/plain
    text/plain; charset=ISO-8859-1
    text/plain; charset=iso-8859-1

and the server did not send a Content-Encoding header, Mozilla will sniff the first block of data it gets and check for non-text bytes. Text bytes are 9-13, 27, and 31-255. When encountering a non-text byte, the helper app dialog will be shown, showing the MIME type corresponding to the extension of the file.

kaspar030 · 2019-03-06T12:58:09Z

@smlng can we try forcing all *.txt to be text/plain; charset=utf-8?

kaspar030 · 2019-03-07T13:38:22Z

Anyhow, the output issue is unrelated to this PR.

@cladmi @smlng do you see any reason not to merge this?

MrKevinWeiss

Simple and clean, I ACK. Maybe we can rerun the ci (on hardware) to make sure everything is good before merge.

MrKevinWeiss · 2019-03-07T14:52:21Z

@kaspar030 hmmm I can't see the tests being run on the hardware in the murdock output, shouldn't the samr21 and nrf52dk be triggered and show up in the worker stats?

--- worker stats:
e5 total: 3326 pass: 3326 fail: 0 avg: 2.1s
e52 total: 3602 pass: 3602 fail: 0 avg: 2.0s
haw-icc total: 6968 pass: 6968 fail: 0 avg: 2.3s
ifive total: 602 pass: 602 fail: 0 avg: 3.2s
mobi3.inet.haw-hamburg.de total: 4759 pass: 4759 fail: 0 avg: 3.3s
mobi7.inet.haw-hamburg.de total: 2897 pass: 2897 fail: 0 avg: 3.1s
mobi8.inet.haw-hamburg.de total: 2661 pass: 2661 fail: 0 avg: 4.4s
riot-ci.inet.haw-hamburg.de total: 1144 pass: 1144 fail: 0 avg: 3.4s
riotbuild total: 8504 pass: 8504 fail: 0 avg: 3.3s

cladmi · 2019-03-07T16:00:20Z

The output issue will be seen on the first failed test for nrf52dk and somebody clicking on the test result. But if they open it in a good editor, it should be ok.

kaspar030 · 2019-03-07T17:34:45Z

@cladmi yes but this issue is also present for samr21-xpro, it is not nrf52dk specific.

kaspar030 · 2019-03-07T17:37:49Z

@MrKevinWeiss results are there now. Care to confirm? there's a race when setting both CI labels at the same time, it sometimes causes one to not be recognized by murdock.

MrKevinWeiss · 2019-03-07T19:54:29Z

All good, thanks @kaspar030

cladmi · 2019-03-11T13:50:43Z

@cladmi yes but this issue is also present for samr21-xpro, it is not nrf52dk specific.

I did not know, I only saw it for nrf52dk.

kaspar030 requested a review from cladmi May 7, 2018 15:19

kaspar030 force-pushed the murdock_enable_nrf52dk branch 2 times, most recently from 5b09ac8 to 0913103 Compare May 7, 2018 21:38

cladmi added CI: ready for build If set, CI server will compile all applications for all available boards for the labeled PR and removed CI: ready for build If set, CI server will compile all applications for all available boards for the labeled PR labels May 8, 2018

tcschmidt requested a review from MrKevinWeiss June 21, 2018 11:18

MrKevinWeiss reviewed Jul 5, 2018

View reviewed changes

kaspar030 force-pushed the murdock_enable_nrf52dk branch from 0913103 to 111f8f7 Compare August 3, 2018 13:38

kaspar030 added the CI: ready for build If set, CI server will compile all applications for all available boards for the labeled PR label Aug 3, 2018

murdock: add nrf52dk to available test boards

e3729be

kaspar030 force-pushed the murdock_enable_nrf52dk branch from befb352 to e3729be Compare March 6, 2019 11:32

kaspar030 mentioned this pull request Mar 7, 2019

nrf52: Add riotboot support #11126

Merged

4 tasks

MrKevinWeiss approved these changes Mar 7, 2019

View reviewed changes

kaspar030 added CI: ready for build If set, CI server will compile all applications for all available boards for the labeled PR and removed CI: ready for build If set, CI server will compile all applications for all available boards for the labeled PR labels Mar 7, 2019

MrKevinWeiss merged commit 282991d into RIOT-OS:master Mar 7, 2019

danpetry added this to the Release 2019.04 milestone Mar 11, 2019

kaspar030 deleted the murdock_enable_nrf52dk branch March 11, 2019 13:53

cladmi mentioned this pull request Jun 13, 2019

murdock and tests using utf-8 characters #11691

Closed

cladmi mentioned this pull request Sep 16, 2019

tests: re-enable utf-8 tests on native #12246

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

murdock: enable nrf52dk for on-hardware tests #9095

murdock: enable nrf52dk for on-hardware tests #9095

kaspar030 commented May 7, 2018

cladmi commented May 9, 2018

tcschmidt commented Jul 4, 2018

MrKevinWeiss left a comment

MrKevinWeiss Jul 5, 2018

kaspar030 Jul 5, 2018

MrKevinWeiss Jul 5, 2018

tcschmidt commented Nov 16, 2018

MrKevinWeiss commented Nov 19, 2018

kaspar030 commented Nov 26, 2018

MrKevinWeiss commented Nov 26, 2018

kaspar030 commented Nov 26, 2018

emmanuelsearch commented Nov 26, 2018

kaspar030 commented Nov 26, 2018

emmanuelsearch commented Nov 26, 2018

kaspar030 commented Nov 26, 2018

kaspar030 commented Feb 27, 2019

cladmi commented Mar 6, 2019

kaspar030 commented Mar 6, 2019

kaspar030 commented Mar 6, 2019

cladmi commented Mar 6, 2019

cladmi commented Mar 6, 2019

kaspar030 commented Mar 6, 2019

kaspar030 commented Mar 6, 2019

kaspar030 commented Mar 7, 2019

MrKevinWeiss left a comment

MrKevinWeiss commented Mar 7, 2019

cladmi commented Mar 7, 2019

kaspar030 commented Mar 7, 2019

kaspar030 commented Mar 7, 2019

MrKevinWeiss commented Mar 7, 2019

cladmi commented Mar 11, 2019

murdock: enable nrf52dk for on-hardware tests #9095

murdock: enable nrf52dk for on-hardware tests #9095

Conversation

kaspar030 commented May 7, 2018

Contribution description

Issues/PRs references

cladmi commented May 9, 2018

tcschmidt commented Jul 4, 2018

MrKevinWeiss left a comment

Choose a reason for hiding this comment

MrKevinWeiss Jul 5, 2018

Choose a reason for hiding this comment

kaspar030 Jul 5, 2018

Choose a reason for hiding this comment

MrKevinWeiss Jul 5, 2018

Choose a reason for hiding this comment

tcschmidt commented Nov 16, 2018

MrKevinWeiss commented Nov 19, 2018

kaspar030 commented Nov 26, 2018

MrKevinWeiss commented Nov 26, 2018

kaspar030 commented Nov 26, 2018

emmanuelsearch commented Nov 26, 2018

kaspar030 commented Nov 26, 2018

emmanuelsearch commented Nov 26, 2018

kaspar030 commented Nov 26, 2018

kaspar030 commented Feb 27, 2019

cladmi commented Mar 6, 2019

kaspar030 commented Mar 6, 2019

kaspar030 commented Mar 6, 2019

cladmi commented Mar 6, 2019

cladmi commented Mar 6, 2019

kaspar030 commented Mar 6, 2019

kaspar030 commented Mar 6, 2019

kaspar030 commented Mar 7, 2019

MrKevinWeiss left a comment

Choose a reason for hiding this comment

MrKevinWeiss commented Mar 7, 2019

cladmi commented Mar 7, 2019

kaspar030 commented Mar 7, 2019

kaspar030 commented Mar 7, 2019

MrKevinWeiss commented Mar 7, 2019

cladmi commented Mar 11, 2019