[tools] Prevent trace-backs from incomplete args

[theotherjimmy]

Also makes all argument errors (from our code and argpase) look the same

[bridadan]

@theotherjimmy yes, this definitely need some TLC!

Any chance we could get a before and after log? (For the unified errors at least)

[theotherjimmy]

example python tools/make.py -m K64F -t GCC_ARM --source .
before

[ERROR] coercing to Unicode: need string or buffer, list found

after

usage: make.py [-h] [-m MCU] [-t TOOLCHAIN] [--color] [-c] [-o OPTIONS]
               [-p PROGRAM] [-n PROGRAM] [-j JOBS] [-v] [--silent] [-D MACROS]
               [-S] [-f GENERAL_FILTER_REGEX] [--automated] [--host HOST_TEST]
               [--extra EXTRA] [--peripherals PERIPHERALS]
               [--dep DEPENDENCIES] [--source SOURCE_DIR]
               [--duration DURATION] [--build BUILD_DIR] [-N ARTIFACT_NAME]
               [-d DISK] [-s SERIAL] [-b BAUD] [-L] [--rtos] [--rpc] [--eth]
               [--usb_host] [--usb] [--dsp] [--fat] [--ublox] [--testlib]
               [-l LINKER_SCRIPT]
make.py: error: argument --build is required when argument --source is provided

[theotherjimmy]

example python tools/get_config.py --source .
before

[ERROR] You should specify an MCU


usage: get_config.py [-h] [-m MCU] [-t TOOLCHAIN] [--color]
                     [--source SOURCE_DIR] [--prefix PREFIX] [-v]

optional arguments:
  -h, --help            show this help message and exit
  -m MCU, --mcu MCU     build for the given MCU (ARCH_BLE, ARCH_BLE_BOOT,
                        ARCH_BLE_OTA, ARCH_GPRS, ARCH_LINK, ARCH_LINK_BOOT,
                        ARCH_LINK_OTA, ARCH_MAX, ARCH_PRO, ARM_BEETLE_SOC,
                        ARM_IOTSS_BEID, ARM_MPS2_M0, ARM_MPS2_M0P,
                        ARM_MPS2_M1, ARM_MPS2_M3, ARM_MPS2_M4, ARM_MPS2_M7,
                        B96B_F446VE, BLUEPILL_F103C8, DELTA_DFCM_NNN40,
                        DELTA_DFCM_NNN40_BOOT, DELTA_DFCM_NNN40_OTA,
                        DISCO_F051R8, DISCO_F100RB, DISCO_F303VC,
                        DISCO_F334C8, DISCO_F401VC, DISCO_F407VG,
                        DISCO_F429ZI, DISCO_F469NI, DISCO_F746NG,
                        DISCO_L053C8, DISCO_L476VG, EFM32GG_STK3700,
                        EFM32HG_STK3400, EFM32LG_STK3600, EFM32PG_STK3401,
                        EFM32WG_STK3800, EFM32ZG_STK3200, ELEKTOR_COCORICO,
                        ELMO_F411RE, HEXIWEAR, HRM1017, HRM1017_BOOT,
                        HRM1017_OTA, K20D50M, K22F, K64F, KL05Z, KL25Z, KL26Z,
                        KL27Z, KL43Z, KL46Z, LPC1114, LPC11C24, LPC11U24,
                        LPC11U24_301, LPC11U34_421, LPC11U35_401,
                        LPC11U35_501, LPC11U35_501_IBDAP, LPC11U35_Y5_MBUG,
                        LPC11U37H_401, LPC11U37_501, LPC11U68, LPC1347,
                        LPC1549, LPC1768, LPC2368, LPC2460, LPC4088,
                        LPC4088_DM, LPC4330_M0, LPC4330_M4, LPC4337, LPC810,
                        LPC812, LPC824, LPCCAPPUCCINO, MAX32600MBED,
                        MAXWSNENV, MICRONFCBOARD, MOTE_L152RC,
                        MTS_DRAGONFLY_F411RE, MTS_GAMBIT, MTS_MDOT_F405RG,
                        MTS_MDOT_F411RE, NRF51822, NRF51822_BOOT,
                        NRF51822_OTA, NRF51822_Y5_MBUG, NRF51_DK,
                        NRF51_DK_BOOT, NRF51_DK_LEGACY, NRF51_DK_OTA,
                        NRF51_DONGLE, NRF51_DONGLE_BOOT, NRF51_DONGLE_OTA,
                        NRF51_MICROBIT, NRF51_MICROBIT_B, NRF51_MICROBIT_BOOT,
                        NRF51_MICROBIT_B_BOOT, NRF51_MICROBIT_B_OTA,
                        NRF51_MICROBIT_OTA, NRF52_DK, NUCLEO_F030R8,
                        NUCLEO_F031K6, NUCLEO_F042K6, NUCLEO_F070RB,
                        NUCLEO_F072RB, NUCLEO_F091RC, NUCLEO_F103RB,
                        NUCLEO_F207ZG, NUCLEO_F302R8, NUCLEO_F303K8,
                        NUCLEO_F303RE, NUCLEO_F334R8, NUCLEO_F401RE,
                        NUCLEO_F410RB, NUCLEO_F411RE, NUCLEO_F429ZI,
                        NUCLEO_F446RE, NUCLEO_F446ZE, NUCLEO_F746ZG,
                        NUCLEO_F767ZI, NUCLEO_L011K4, NUCLEO_L031K6,
                        NUCLEO_L053R8, NUCLEO_L073RZ, NUCLEO_L152RE,
                        NUCLEO_L432KC, NUCLEO_L476RG, NUMAKER_PFM_NUC472,
                        NZ32_SC151, OC_MBUINO, RBLAB_BLENANO,
                        RBLAB_BLENANO_BOOT, RBLAB_BLENANO_OTA, RBLAB_NRF51822,
                        RBLAB_NRF51822_BOOT, RBLAB_NRF51822_OTA, RZ_A1H,
                        SAMD21G18A, SAMD21J18A, SAMG55J19, SAML21J18A,
                        SAMR21G18A, SEEED_TINY_BLE, SEEED_TINY_BLE_BOOT,
                        SEEED_TINY_BLE_OTA, SSCI824, STM32F3XX, STM32F407,
                        TEENSY3_1, TY51822R3, TY51822R3_BOOT, TY51822R3_OTA,
                        UBLOX_C027, UBLOX_C029, VK_RZ_A1H, WALLBOT_BLE,
                        WALLBOT_BLE_BOOT, WALLBOT_BLE_OTA, WIZWIKI_W7500,
                        WIZWIKI_W7500ECO, WIZWIKI_W7500P, XADOW_M0,
                        XBED_LPC1768)
  -t TOOLCHAIN, --tool TOOLCHAIN
                        build using the given TOOLCHAIN (ARM, GCC_ARM, GCC_CR,
                        IAR, uARM)
  --color               print Warnings, and Errors in color
  --source SOURCE_DIR   The source (input) directory
  --prefix PREFIX       Restrict listing to parameters that have this prefix
  -v, --verbose         Verbose diagnostic output

after

usage: get_config.py [-h] [-m MCU] [-t TOOLCHAIN] [--color] --source
                     SOURCE_DIR [--prefix PREFIX] [-v]
get_config.py: error: argument -m/--mcu is required

[0xc0170]

Why did you remove [ERROR] ? wasn't that an intention warning/error/etc ? Does this PR clean all [ERROR] messages ?

[theotherjimmy]

args_error already prints an error message.

[0xc0170]

args_error already prints an error message.

Yes, that's obvious from the outputs you shared but it changes from [ERROR] to command: error:. I am not aware though how was that used (if there any tool that was parsing this or was just an unification of messages we print to have a format. That was my concern, looks fine to me as it is now ;)

[theotherjimmy]

Ah, that makes sense. Thanks for the clarification. I only changed things that were "argument errors".

[bridadan]

@theotherjimmy Thanks! I like the change :)

[bridadan]

/morph test

[mbed-bot]

Result: FAILURE

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

/morph test

Output

mbed Build Number: 619

Test failed!

[theotherjimmy]

/morph test

[mbed-bot]

Result: FAILURE

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

/morph test

Output

mbed Build Number: 620

Test failed!

[bridadan]

The only failing test was the UDP nist test, which may have actually failed due to the UDP packets being lost. Any chance you can run that test locally just to double check? I'm about 98.9% sure that this tooling change didn't cause this test failure, so LGTM.

Side note, the nist test should probably retry multiple times if it doesn't hear back within a time frame, maybe I'll go look at that...

[theotherjimmy]

Thanks for explaining the failure @bridadan ! LGTM?

[theotherjimmy]

Resolves #1997

[0xc0170]

LGTM

cc @bogdanm @screamerbg

[bogdanm]

Is this corrrect? Do we really require the user to specify a build dir if he specifies one or more source dirs? Feels a bit weird.

[theotherjimmy]

Yes, We do. I agree, it does feel a bit weird. If you don't specify a build directory, you get a traceback about None not having startswith, I think.

[sg-]

Any more feedback or is this ready for CI and merge?

[theotherjimmy]

I guess it's ready for CI and merge.

[theotherjimmy]

So, I still think that this is good to go. Let's get this merged?

[bogdanm]

LGTM