Add Mbed Configuration Option Range Limits

[kegilbert]

Description

Add optional range limits and accepted value lists to configuration option default and overridden values.

Added some rudimentary tests, let me know if you'd like any more. Will squash after the review period.

Example output with multiple invalid configurations in a Mbed OS build:

Invalid config range for platform.stdio-baud-rate, value 9600 is not in the required range: [115200:-1]    
Invalid config range for drivers.test, value bloep is not an accepted value: [u'bloop', u'bleep']          
Invalid config range for drivers.uart-serial-rxbuf-size, value 256 is not in the required range: [0:100]   
Invalid config range for drivers.uart-serial-txbuf-size, value 256 is not in the required range: [300:-1]                                                                    

You can see the proposed JSON format in the added test files.

Pull request type

[ ] Fix
[ ] Refactor
[ ] Target update
[X] Functionality change
[ ] Docs update
[ ] Test update
[ ] Breaking change

[kegilbert]

@theotherjimmy Updated, rebased, and retested.

New error format:

Building project mbed-os (K64F, GCC_ARM)                                                                                           
Scan: mbed-os                                                                                                                      
[ERROR]                                                                                                                            
Invalid config range for drivers.uart-serial-rxbuf-size = 256 (macro name: "MBED_CONF_DRIVERS_UART_SERIAL_RXBUF_SIZE"), is not an a
ccepted value: [0, 512]                                                                                                            
Invalid config range for drivers.uart-serial-txbuf-size = 256 (macro name: "MBED_CONF_DRIVERS_UART_SERIAL_TXBUF_SIZE"), is not in t
he required range: [-inf:200]                                                                                                      
Invalid config range for platform.stdio-baud-rate = 9600 (macro name: "MBED_CONF_PLATFORM_STDIO_BAUD_RATE"), is not in the required
 range: [115200:inf]                                                                                                               

[bridadan]

Super slick feature! A few questions/suggestions below.

[kegilbert]

@theotherjimmy @bridadan Fairly substantial rework to better handle all string/integer formats:

	Range	Acceptable Value List	Both
None	X	X	X
5	./	./	X
"5"	./	./	X
"0x05"	./	./	X
"Text"	X	./	X

As a result it does look a bit grosser than before, let me know your thoughts.

@theotherjimmy Question in regards to:
https://github.com/ARMmbed/mbed-os/blob/master/tools/test/config/config_test.py#L92

This is checking if the expected exception message is in the actual exception string which means if the test throws additional exceptions but the first part of the message stays the same it will pass. For example:

If I expect

Invalid config range for drivers.uart-serial-c1 = 256 (macro name: "MBED_CONF_DRIVERS_UART_SERIAL_C1"), is not an accepted value: 0, 2560, 312                   
Invalid config range settings for drivers.uart-serial-c2 = text (macro name: "MBED_CONF_DRIVERS_UART_SERIAL_C2"). Range specifiers are not applicable to non-deci
mal/hexadecimal string values                                                                                                                                    
Invalid config range for drivers.uart-serial-rxbuf-size = 256 (macro name: "MBED_CONF_DRIVERS_UART_SERIAL_RXBUF_SIZE"), is not an accepted value: 0, 255, 312    

but get the same exception with the addition of

Invalid config range for drivers.uart-serial-txbuf-size = 256 (macro name: "MBED_CONF_DRIVERS_UART_SERIAL_TXBUF_SIZE"), is not in the required range: [200:250]  

That will pass. Changing that "is in" to a hard check if the strings match one to one causes several other tests to fail such as the duplicate lib test as the actual exception contains path information that we can't hardcode so I'm assuming this was done on purpose. Just a heads up that with how I'm generating exceptions there is a test gap as is.

[bridadan]

Changing that "is in" to a hard check if the strings match one to one causes several other tests to fail such as the duplicate lib test as the actual exception contains path information that we can't hardcode so I'm assuming this was done on purpose.

Huh, very interesting, thanks for checking that. Maybe something like a regular expression would be fitting here?

[bridadan]

Some simplifications and another question or two.

[kegilbert]

@bridadan Refactored, awesome find with the automagic int conversions looks much cleaner. Changed the JSON formats back to the array format, but print out a comma separated string list on exceptions for better readability.

I'll sync with you tomorrow on my thoughts of updating the expected exception message logic.

[bridadan]

Looking really good! Two more little nit picks but I think we're getting there!

[kegilbert]

@bridadan Updated!

[cmonr]

Well darn. @kegilbert Please take a look at the pytest failure.

PR updated.

[kegilbert]

Forgot to update the expected error message in the test case, should be fixed.

[bridadan]

Very nice work! Thanks for being so speedy getting the feedback integrated 👍

[kegilbert]

Thanks for giving the feedback!

[cmonr]

/morph build

[mbed-ci]

Build : SUCCESS

Build number : 3654
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/8673/

Triggering tests

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

[mbed-ci]

Exporter Build : SUCCESS

Build number : 3261
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/8673/

[mbed-ci]

Test : SUCCESS

Build number : 3434
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/8673/3434