Don't set the build_dir to anything on export #3924

theotherjimmy · 2017-03-10T17:22:16Z

When constructing a toolchain for export, we currently set the
build_dir to the export_dir. When exporting offline, the
export_dir is always set to the root of the project. The toolchains
ignore their build_dir when scanning for sorces, so when the exporters
use the toolchains to scan for their resources, they get nothing.

In this patch we set the build_dir of the toolchain that the exports
use to nothing. A path of nothing should not match anything, and will
therefore not ignore everything when scanning for resources.

TODO

/morph export-build

theotherjimmy · 2017-03-10T17:22:32Z

/morph export-build

mbed-bot · 2017-03-10T18:07:37Z

Result: SUCCESS

Your command has finished executing! Here's what you wrote!

/morph export-build

Output

mbed Build Number: 135

All exports and builds passed!

sg- · 2017-03-11T19:41:44Z

/morph export-build

mbed-bot · 2017-03-11T20:25:57Z

Result: SUCCESS

Your command has finished executing! Here's what you wrote!

/morph export-build

Output

mbed Build Number: 136

All exports and builds passed!

0xc0170 · 2017-03-13T13:34:55Z

I restarted jenkins CI as I could not locate the test results, please @theotherjimmy review once it finishes

0xc0170

LGTM

Small typo in the commit message- bulid (build?). I had to look at the prepare_toolchian definition to see what is being set to "". It was not clear to me from the commit.

0xc0170 · 2017-03-14T08:57:04Z

Restarted CI jenkins to report back the status

adbridge · 2017-03-14T14:44:36Z

@theotherjimmy "We are not using the toolchains to bulid, and they ignore everything in
the build_dir when scanning, so let's not ignore everything" Could you please re-write this to a meaningful comment that everybody can understand.

theotherjimmy · 2017-03-14T15:53:21Z

Ah, I missed the the tool chains have the build dir set to the root of the project bit

When constructing a toolchain for export, we currently set the `build_dir` to the `export_dir`. When exporting offline, the `export_dir` is always set to the root of the project. The toolchains ignore their `build_dir` when scanning for sorces, so when the exporters use the toolchains to scan for their resources, they get nothing. In this patch we set the `build_dir` of the toolchain that the exports use to nothing. A path of nothing should not match anything, and will therefore not ignore everything when scanning for resources.

theotherjimmy · 2017-03-14T16:33:40Z

Commit message changed. Copying to PR description.

theotherjimmy · 2017-03-14T16:34:28Z

/morph export-build

theotherjimmy · 2017-03-14T16:34:40Z

'cause rebase

theotherjimmy · 2017-03-14T16:35:55Z

@adbridge I changed the commit message to be more useful (and updated the PR description to match it). Could you comment again?

adbridge · 2017-03-14T16:43:28Z

@theotherjimmy Thanks for the updated description, much improved.

theotherjimmy · 2017-03-14T16:44:04Z

@adbridge Your welcome. I'm changing it to needs: CI then.

adbridge · 2017-03-14T16:46:10Z

@marhil01 @SeppoTakalo Any idea why pr_head is failing here? I couldn't see an obvious failure?

mbed-bot · 2017-03-14T18:16:18Z

Result: SUCCESS

Your command has finished executing! Here's what you wrote!

/morph export-build

Output

mbed Build Number: 138

All exports and builds passed!

bridadan · 2017-03-14T19:28:02Z

Looks like there's a build failure with K64F for the mbed-client-testapp?

18:27:49 [K64F ARM] "/tmp/4Bnhfm", line 128 (column 3): Warning: L6312W: Empty Execution region description for region RW_IRAM1
18:27:49 [K64F ARM] Error: L6200E: Symbol __asm___12_Ethernet_cpp___Z7__REV16j multiply defined (by Ethernet.o and CAN.o).
18:27:49 [K64F ARM] Error: L6200E: Symbol __asm___12_Ethernet_cpp___Z7__REVSHi multiply defined (by Ethernet.o and CAN.o).
18:27:49 [K64F ARM] Error: L6200E: Symbol __asm___12_Ethernet_cpp___Z5__RRXj multiply defined (by Ethernet.o and CAN.o).
18:27:49 [K64F ARM] Not enough information to list the image map.
18:27:49 [K64F ARM] Finished: 1 information, 1 warning and 3 error messages.
18:27:49 [K64F ARM] [ERROR] "/tmp/4Bnhfm", line 128 (column 3): Warning: L6312W: Empty Execution region description for region RW_IRAM1
18:27:49 [K64F ARM] Error: L6200E: Symbol __asm___12_Ethernet_cpp___Z7__REV16j multiply defined (by Ethernet.o and CAN.o).
18:27:49 [K64F ARM] Error: L6200E: Symbol __asm___12_Ethernet_cpp___Z7__REVSHi multiply defined (by Ethernet.o and CAN.o).
18:27:49 [K64F ARM] Error: L6200E: Symbol __asm___12_Ethernet_cpp___Z5__RRXj multiply defined (by Ethernet.o and CAN.o).

0xc0170 · 2017-03-15T09:53:31Z

@bridadan where is the failure? All tests are now green?

bridadan · 2017-03-15T15:06:21Z

@0xc0170 I had restarted the CI once more after posting that, glad the issue is resolved!

0xc0170 · 2017-03-16T09:58:24Z

@theotherjimmy This is a bugfix for a bug that is on master and even on 5.4 branch (I faced this issue locally and took me a while to figure out that this is not my local setup but upstream bug). The importance of this bug is higher than it seems. It clicked all once I started browsing and saw all the reports (were referenced only 2 days ago, but this patch was sent 6 days ago thus we could already know more about this bug 6 days ago?). This could be easily integrated for the last release but was not as the description was lacking details, and even after improved, I still don't find it enough. For instance, when was this bug introduced? I had to go to the history, this one : theotherjimmy@fbb6f71. So this is fixing already another fix that introduced this bug. How can I reproduce this bug ? what does this bugfix fixes? As in the commit message : Before this patch: description here. After this patch: description here.

Please take more time to describe the issue, how this patch fixes it and reference issues or any relevant information. I assume this bugfix could be already in the last release!

How come this was not discovered with our build test that the project even fails to export or exported project is empty (only root is scanned, and in my case it was just config header file in the workspace view) ? @theotherjimmy Did you check build tests? I checked the PR that introduced this change, we did not run build CI test ! Here's it : #3852 (no mention of running build CI). Would it fail?

bridadan · 2017-03-16T16:12:45Z

To help prevent this from happening in the future, I have a test job that will replace the nightly. It will test the exporters in parallel to running tests on the hardware, so you shouldn't have to run the explicit exporter job any more.

adbridge · 2017-03-20T11:23:35Z

@bridadan Do you mean that when we run /morph test-nightly going forward it will automatically test exporters as well? Or do you mean you have added a new, separate command ? If it's the latter +1 if it's the former +5 !

mbed-bot · 2017-03-20T12:00:25Z

Result: FAILURE

Your command has finished executing! Here's what you wrote!

/morph test-nightly going forward it will automatically test exporters as well? Or do you mean you have added a new, separate command ? If it's the latter +1 if it's the former +5 !

Output

mbed Build Number: 1700

Example Build failed!

bridadan · 2017-03-20T15:52:39Z

@adbridge It's not active now, but going forward the plan is to have the nightly command run the exporters as well. There a still a few issues to iron out. But this really needs to get in first!

sg- · 2017-03-20T21:18:44Z

/morph test-nightly

mbed-bot · 2017-03-21T01:00:36Z

Result: SUCCESS

Your command has finished executing! Here's what you wrote!

/morph test-nightly

Output

mbed Build Number: 1706

All builds and test passed!

theotherjimmy force-pushed the fix-export-build-dir branch from bc0aa18 to bea8287 Compare March 10, 2017 19:30

0xc0170 approved these changes Mar 13, 2017

View reviewed changes

0xc0170 added the needs: CI label Mar 13, 2017

0xc0170 added ready for merge and removed needs: CI labels Mar 14, 2017

adbridge added needs: work and removed ready for merge labels Mar 14, 2017

theotherjimmy added needs: clarification and removed needs: work labels Mar 14, 2017

theotherjimmy force-pushed the fix-export-build-dir branch from bea8287 to 4219d9c Compare March 14, 2017 16:33

bridadan approved these changes Mar 14, 2017

View reviewed changes

This was referenced Mar 14, 2017

[OoB_5.4.1] [EXPORTER] : export to uvision5 and make_armc5 fails on NuMaker_PFM_NUC472. ARMmbed/mbed-os-example-tls#68

Closed

Cannot export to uvision5, platform UBLOX_EVK_ODIN_W2 ARMmbed/mbed-os-example-tls#67

Closed

theotherjimmy added needs: CI and removed needs: clarification labels Mar 14, 2017

0xc0170 added ready for merge and removed needs: CI labels Mar 15, 2017

0xc0170 mentioned this pull request Mar 16, 2017

Export - scan only root folder, only config file in the workspace #3938

Closed

chrissnow mentioned this pull request Mar 20, 2017

EFM32 Pearl Gecko export broken #3963

Closed

theotherjimmy mentioned this pull request Mar 20, 2017

Export for GNU ARM Eclipse fails with exception #3964

Closed

sg- merged commit 16304ae into ARMmbed:master Mar 22, 2017

sg- added release-version: 5.4.2 and removed ready for merge labels Mar 22, 2017

theotherjimmy mentioned this pull request Apr 6, 2017

Project exported error ARMmbed/mbed-os-example-client#227

Closed

Don't set the build_dir to anything on export #3924

Don't set the build_dir to anything on export #3924

Uh oh!

Conversation

theotherjimmy commented Mar 10, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

TODO

Uh oh!

theotherjimmy commented Mar 10, 2017

Uh oh!

mbed-bot commented Mar 10, 2017

Result: SUCCESS

Output

Uh oh!

sg- commented Mar 11, 2017

Uh oh!

mbed-bot commented Mar 11, 2017

Result: SUCCESS

Output

Uh oh!

0xc0170 commented Mar 13, 2017

Uh oh!

0xc0170 left a comment

Choose a reason for hiding this comment

Uh oh!

0xc0170 commented Mar 14, 2017

Uh oh!

adbridge commented Mar 14, 2017

Uh oh!

theotherjimmy commented Mar 14, 2017

Uh oh!

theotherjimmy commented Mar 14, 2017

Uh oh!

theotherjimmy commented Mar 14, 2017

Uh oh!

theotherjimmy commented Mar 14, 2017

Uh oh!

theotherjimmy commented Mar 14, 2017

Uh oh!

adbridge commented Mar 14, 2017

Uh oh!

theotherjimmy commented Mar 14, 2017

Uh oh!

adbridge commented Mar 14, 2017

Uh oh!

mbed-bot commented Mar 14, 2017

Result: SUCCESS

Output

Uh oh!

bridadan commented Mar 14, 2017

Uh oh!

0xc0170 commented Mar 15, 2017

Uh oh!

bridadan commented Mar 15, 2017

Uh oh!

0xc0170 commented Mar 16, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

bridadan commented Mar 16, 2017

Uh oh!

adbridge commented Mar 20, 2017

Uh oh!

mbed-bot commented Mar 20, 2017

Result: FAILURE

Output

Uh oh!

bridadan commented Mar 20, 2017

Uh oh!

sg- commented Mar 20, 2017

Uh oh!

mbed-bot commented Mar 21, 2017

Result: SUCCESS

Output

Uh oh!

Uh oh!

theotherjimmy commented Mar 10, 2017 •

edited

Loading

0xc0170 commented Mar 16, 2017 •

edited

Loading