Advance notice of folder structure change in under "/targets/TARGET_RENESAS"

[TomoYamanaka]

@mbedNoobNinja

Hello, we are Renesas mbed team.

We will release a new mbed board that mounts RZ/A1.
Related to that, we consider that would like to share common codes in the code below HAL in mbed-os.
The structure that we would like to change is under /targets/TARGET_RENESAS, we show its structure in the below.

New folder structure will be merged by end of Sep.
So, please let me know if there is any problem about new folder structure by end of Aug.
Also, we would appreciate it if you could execute the test by using new folder structure.

Current state:
The folder are separated by mbed board unit, and there are many same files in these folders.
https://github.com/ARMmbed/mbed-os/tree/master/targets/TARGET_RENESAS

proposed change:
The same files are shared, only difference files are separated in folder by mbed board unit.
In this way, maintenance and new board supporting will be very easy.
https://github.com/TomoYamanaka/mbed/tree/Feature_LYCHEE/targets/TARGET_RENESAS

[theotherjimmy]

@TomoYamanaka That seems fine. I don't forseee the build system having any problems with that.

Would you also like an advance review of your changes to targets.json?

[theotherjimmy]

master...TomoYamanaka:Feature_LYCHEE That's the diff with master. Your branch seems to be a bit behind, and it would not be accepted if you submitted a PR right now.

[TomoYamanaka]

@theotherjimmy

Would you also like an advance review of your changes to targets.json?

Thank you.
Please let me know if there is any problem.
I will reflect its comments when I PR.

That's the diff with master. Your branch seems to be a bit behind, and it would not be accepted if you submitted a PR right now.

That's right. I understand that this branch is not up to date.
This branch is based on mbed-os-5.4.3 for some development.
For now, I will PR by end of Sep. At that time I will PR this branch after merge the latest master.

[mbedNoobNinja]

@TomoYamanaka that's great, we are glad you are considering that.

Also, we would appreciate it if you could execute the test by using new folder

We will be able to run the tests not sooner than 2-3 weeks.

[TomoYamanaka]

@mbedNoobNinja

Thank you.
We are looking forward to your reply.

[mbedNoobNinja]

Finally ...
Sorry for the delay here are the results for the master branch of @TomoYamanaka :
cli_tst_summary.txt
old_tst_summary.txt
here are the complete logs:
old_tst_arm_iar_gcc.txt
cli_tst_arm.txt
cli_tst_gcc.txt
cli_tst_iar.txt

Not bad, except the IAR ones ... but this is just for reference.

The branch in interest however is Feature_LYCHEE of @TomoYamanaka and here I am getting strange pyton error:
mbed test -m VK_RZ_A1H -t ARM
Traceback (most recent call last):
File "C:\Projects\kitVKRZA1H\mbed\tools\test.py", line 247, in
status = print_report_exporter.report(build_report)
File "C:\Projects\kitVKRZA1H\mbed\tools\test_exporters.py", line 92, in report
return self.exporter_print(test_summary_ext, print_log_for_failures=print_log_for_failures)
File "C:\Projects\kitVKRZA1H\mbed\tools\test_exporters.py", line 343, in exporter_print
raise Exception("'test_run' did not have a 'result' value")
Exception: 'test_run' did not have a 'result' value
[ERROR] 'test_run' did not have a 'result' value
[mbed] ERROR: "python" returned error code 1.
[mbed] ERROR: Command "python -u C:\Projects\kitVKRZA1H\mbed\tools\test.py -t ARM -m VK_RZ_A1H --source C:\Projects\kitVKRZA1H\mbed --build C:\Projects\kitVKRZA1H\mbed\BUILD\tests\VK_RZ_A1H\ARM --test-spec C:\Projects\kitVKRZA1H\mbed\BUILD\tests\VK_RZ_A1H\ARM\test_spec.json" in "C:\Projects\kitVKRZA1H\mbed"

The error is the same if I say:
mbed test -m RZ_A1H -t ARM

I am not really familiar with the mbed os testing procedure, the mbed 2 testing was more simple, but this looks far more advanced and complicated for me (maybe because I am too rookie with the python).
Anyway ... Any suggestions what to do next will be gladly appreciated !

[theotherjimmy]

@TomoYamanaka @mbedNoobNinja The branch Feature_LYCHEE seems to be based on a release branch. We do not accept PRs to release branches at the moment, so I recommend that you base your work on master.

[mbedNoobNinja]

... Well the error was in my TV! (more specifically in execution of switch command of the TortoiseGit) In attempt to switch from master to Feature_LYCHEE that mess up the whole project and cause the strange python issue. After fresh & clean checkout of the Feature_LYCHEE branch with TortoiseGit command (Git Clone) there were no python errors anymore.
... That's fine, but new issue come up:
[Error] i2c_api.c@64,0: #154: expression must have struct or union type
[Error] spi_api.c@47,0: #154: expression must have struct or union type
[Error] serial_api.c@151,0: #136: struct "serial_s" has no field "serial"
...Anyway after adding "I2C_ASYNCH", "SPI_ASYNCH" & "SERIAL_ASYNCH" interfaces as "device_has" ittems for the "VK_RZ_A1H" platform in targets.json, the errors disappeared and the test finally managed to run.

Here are the results:
xx_cli_tst_summary.txt
xx_old_tst_summary.txt
and the logs:
xx_cli_tst_arm.txt
xx_cli_tst_gcc.txt
xx_cli_tst_iar.txt
xx_old_tst_arm_iar_gcc.txt

Definitely there is some issue with the IAR binaries, it looks they stuck somewhere.
The other total failure is features-feature_lwip-tests-mbedmicro-net-tcp_hello_world

That's the situation for now ... any help will be welcome!

[TomoYamanaka]

@mbedNoobNinja

Thank you for your comments, and I appreciate that you execute the test.

Summarizing the above, is it right that the rest problems is the below?

Definitely there is some issue with the IAR binaries, it looks they stuck somewhere.
The other total failure is features-feature_lwip-tests-mbedmicro-net-tcp_hello_world

This Feature_LYCHEE branch is based on the release branch of mbed-os-5.4.3.
So, can you execute the test again by using mbed-os-5.4.3?

[TomoYamanaka]

@theotherjimmy

Thank you for your comments.

@TomoYamanaka @mbedNoobNinja The branch Feature_LYCHEE seems to be based on a release branch. We do not accept PRs to release branches at the moment, so I recommend that you base your work on master.

That's right.
This branch is based on release branch of mbed-os-5.4.3 for advance development.
Therefore, I will PR, after I merge Feature_LYCHEE branch to the latest master branch in advance.

[mbedNoobNinja]

I don't know how to explicitly checkout mbed-os-5.4.3.
The tags in the branch Feature_LYCHEE are:

feature_LYCHEE_rev1
feature_LYCHEE_rev2
feature_LYCHEE_rev3
mbed-os-5.1.0
mbed-os-5.1.0-rc1
mbed-os-5.1.0-rc2
mbed-os-5.1.0-rc3
mbed-os-5.1.0-rc4
mbed-os-5.1.0-rc5
mbed-os-5.2.0-rc1
mbed-os-5.3.0-rc1
mbed-os-5.4.0
mbed-os-5.4.0-rc1
mbed-os-5.4.0-rc2
mbed-os-5.4.1
mbed-os-5.4.2
mbed_lib_rev100
mbed_lib_rev101
mbed_lib_rev102
mbed_lib_rev103
mbed_lib_rev104
mbed_lib_rev105
mbed_lib_rev106
mbed_lib_rev107
mbed_lib_rev108
mbed_lib_rev110
mbed_lib_rev111
mbed_lib_rev112
mbed_lib_rev113
mbed_lib_rev114
mbed_lib_rev116
mbed_lib_rev117
mbed_lib_rev118
mbed_lib_rev119
mbed_lib_rev120
mbed_lib_rev121
mbed_lib_rev122
mbed_lib_rev138
mbed_lib_rev139
mbed_lib_rev65
mbed_lib_rev67
mbed_lib_rev68
mbed_lib_rev69
mbed_lib_rev70
mbed_lib_rev71
mbed_lib_rev73
mbed_lib_rev74
mbed_lib_rev75
mbed_lib_rev76
mbed_lib_rev77
mbed_lib_rev78
mbed_lib_rev79
mbed_lib_rev80
mbed_lib_rev81
mbed_lib_rev82
mbed_lib_rev83
mbed_lib_rev84
mbed_lib_rev85
mbed_lib_rev86
mbed_lib_rev87
mbed_lib_rev88
mbed_lib_rev89
mbed_lib_rev90
mbed_lib_rev91
mbed_lib_rev92
mbed_lib_rev93
mbed_lib_rev94
mbed_lib_rev95
mbed_lib_rev96
mbed_lib_rev97
mbed_lib_rev98
mbed_lib_rev99
mbed_liv_rev108
mbed_private_fsl_initial

There is no 5.4.3 version in it.
From official ARMmbed master's branch point of view, there is tag mbed-os-5.4.3 and if I checkout that the targets folder structure obviously won't have TARGET_RZA1XX. Should I run the test no matter what the folder structure is ?

If Yes, here are the results:
5.4.3_cli_tst_summary.txt
5.4.3_cli_tst_arm.txt
5.4.3_cli_tst_gcc.txt

features-feature_lwip-tests-mbedmicro-net-tcp_echo is fine, I accidentally pull the RJ cable out and the test failed !

I suppose if the old tests live in folder named unsupported they are not interesting anymore so I skipped them.

[theotherjimmy]

I will PR, after I merge Feature_LYCHEE branch to the latest master branch in advance.

Please do not merge. There are several reasons not to:

• We do not accept PRs to master with merge commits
• You would be introducing duplicate history (master's and the release branch's)

Instead, you will have to cherry pick the commit range. You cannot rebase because there are duplicated commits on the release branch with different hashes, because they have a different ancestor. If you were to rebase, git would try to apply all of the release commits onto master in your branch. This is why I recomend doing development off master whenever possible: it allows you to rebase to bring things up to date.

[TomoYamanaka]

@mbedNoobNinja

Thank you for test of mbed OS 5.4.3.
Sorry for lack of description.
Because I wanted to comfirm whether the new folder structure has affected, I asked for test to you.

As far as the test result, it seems that features-feature_lwip-tests-mbedmicro-net-tcp_hello_world test is fail in mbed OS 5.4.3 too.
I was relieved that the new folder structure did not seem to be affected.