Fixed float-abi linker option for CCES exporter.

[mharringADI]

Description

This pull request fixes issue #6977. The CCES exporter was incorrectly specifying the float-abi linker option in the .json file. There was also an issue with handling the toolchain cpu options. The CCES exporter was supposed to remove the toolchain cpu options from the linker and compiler flags, but there was a bug preventing it from removing them.

Pull request type

[x] Fix
[ ] Refactor
[ ] New target
[ ] Feature
[ ] Breaking change

[li-ho]

@mharringADI,

Those #'s in ADuCM4050.ld needs to be converted and removed by arm-none-eabi-cpp.

• "arm-none-eabi-cpp" -E -P -DMBED_DEBUG -DMBED_TRAP_ERRORS_ENABLED=1 -Wl,-n -Wl,--start-group -lstdc++ -lsupc++ -lm -lc -lgcc -lnosys -Wl,--end-group ../mbed-os/targets/TARGET_Analog_Devices/TARGET_ADUCM4X50/TARGET_ADUCM4050/TOOLCHAIN_GCC_ARM/ADuCM4050.ld -o linker-script-debug.ld
• "arm-none-eabi-cpp" -E -P -DNDEBUG -Wl,-n -Wl,--start-group -lstdc++ -lsupc++ -lm -lc -lgcc -lnosys -Wl,--end-group ../mbed-os/targets/TARGET_Analog_Devices/TARGET_ADUCM4X50/TARGET_ADUCM4050/TOOLCHAIN_GCC_ARM/ADuCM4050.ld -o linker-script-release.ld
• "arm-none-eabi-cpp" -E -P -Wl,-n -Wl,--start-group -lstdc++ -lsupc++ -lm -lc -lgcc -lnosys -Wl,--end-group ../mbed-os/targets/TARGET_Analog_Devices/TARGET_ADUCM4X50/TARGET_ADUCM4050/TOOLCHAIN_GCC_ARM/ADuCM4050.ld -o linker-script-develop.ld

For Debug configuration, linker-script-debug.ld is chosen.

If ADuCM4050.ld is not converted, arm-none-eabi-g++ reports ignoring invalid character `#' in expression syntax error and does not generate elf or bin files

[andrew-mclachlan]

Hi @li-ho - the issue regarding the LD script is a separate issue IMHO (#6954). This Pull Request is to set the float-abi option properly. Re: LD script issue, we need an update to CrossCore Embedded Studio (CCES) to handle the pre-processing. A JIRA issue has been logged internally and we'll work to update CCES for the next release.

[li-ho]

Hi @andrew-mclachlan, I only share some thoughts for your next PR. Please ignore me.

[theotherjimmy]

Could you split this rename into another commit?

[andrew-mclachlan]

For my own education, could you clarify why it should be a separate commit? Thanks!

[theotherjimmy]

This change (as far as I can tell) is unrelated. I would like to keep the changes related to the fix separate from format/naming changes.

[mharringADI]

The rename was added because the bug was caused by a confusion over what the two arguments referred to. By renaming the arguments it should avoid such a confusion in the future. You'll notice that there was more change here than just the rename; the container flags are now removed instead of the flags_to_remove flags

[theotherjimmy]

👍 Thanks for the explanation!

[0xc0170]

/morph build

[mbed-ci]

Build : SUCCESS

Build number : 2178
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/7011/

Triggering tests

/morph test
/morph uvisor-test
/morph export-build
/morph mbed2-build

[mbed-ci]

Test : SUCCESS

Build number : 1966
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/7011/1966

[mbed-ci]

Exporter Build : SUCCESS

Build number : 1803
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/7011/

[theotherjimmy]

@0xc0170 @cmonr Is there a reason to hold this up? Are we making a release candidate ATM?

[cmonr]

It's coming in during gatekeeping, along with all of the other PRs that are marked "ready for merge".