Remove deprecated flags args #3385
Conversation
|
/morph test-nightly |
|
Reviewers: Could you verify that removing these flags will not break any of your CI, tests, or other stuff? |
Result: SUCCESSYour command has finished executing! Here's what you wrote!
OutputAll builds and test passed! |
How are they advised? I assume these args are removed, and a user needs to search for file changes and find out why they were removed and what to use instead. Shouldn't we accept these arguments and print a warning Nit: This could be helpful to a reader of the commit message (once I find out that this broke my build, I would find this commit and try to get a reason why and how to use new options). I would provide more information to the commit message (migration path - use --profile debug/default/small instead + links -> references to the docs how to use profiles). From one commit I would be able to get all the information I need to migrate and update my codebase. |
|
Agree with @0xc0170 we've seen a number of changes like this recently which are not obvious about how to go about fixing. I think the suggestion about adding a deprecation message with meaningful information and links to any documentation is definitely a good suggestion! |
|
@0xc0170 Thanks for the feedback. I'll update the commit message to be more specific. |
theotherjimmy
force-pushed
the
remove-flags-args
branch
from
800fbef to
8701c36
Compare
|
/morph test |
Result: SUCCESSYour command has finished executing! Here's what you wrote!
OutputAll builds and test passed! |
|
@theotherjimmy One thing that's still not clear to me is if a user supplies any one of the 3 removed arguments, will they be greeted with an argparse error? @0xc0170 and @adbridge brought up a good point about lots of changing parameters recenetly, |
The tools will no longer accept `--cflags`, `--cppflags`, or `--ldflags`. Instead, the ability to modify these flags is provided by the `--profile` argument. Documentation for the `--profile` argument may be found in docs/Toolchain_Profiles.md
theotherjimmy
force-pushed
the
remove-flags-args
branch
from
8701c36 to
d9c8d1c
Compare
|
@mazimkhan how do we re-run uvisor test on this? |
|
/morph test-nightly |
|
/morph export-build |
|
@mbed-bot: TEST HOST_OSES=ALL |
|
retest uvisor |
Result: FAILUREYour command has finished executing! Here's what you wrote!
OutputTest Prep failed! |
Result: FAILUREYour command has finished executing! Here's what you wrote!
Outputmbed Build Number: 53 Exporter Build failed! |
|
nightly failure is a synchronization issue on windows in preparation (two threads trying to access the same file from two contexts). |
|
export build seems to be the duplicate i2c_api.c files issue. |
|
@theotherjimmy Can you comment a bit more on the nightly failure? I admit I haven't seen that issue before. I'll go ahead and rerun the nightly. /morph test-nightly |
|
/morph export-build |
Result: SUCCESSYour command has finished executing! Here's what you wrote!
OutputAll builds and test passed! |
Result: FAILUREYour command has finished executing! Here's what you wrote!
Outputmbed Build Number: 54 Exporter Build failed! |
|
I still don't think that the intermittent uvision failures are related. They seem to affect a random target every time and only one object file each time. |
|
/morph export-build |
Result: SUCCESSYour command has finished executing! Here's what you wrote!
Outputmbed Build Number: 62 All exports and builds passed! |
|
/morph export-build |
|
Dang it. These pages need to stop being cached and not updating. |
Result: SUCCESSYour command has finished executing! Here's what you wrote!
Outputmbed Build Number: 64 All exports and builds passed! |
Status
READY
Migrations
YES
Anyone using the
--cflags,--cppflags, or--ldflagsarguments are advised that they have no affect, and should stop passing them. Please migrate any flag shenanigans to using--profileTODOs