Tools: Omit include path parents when they're scan rules

[theotherjimmy]

Description

Realistically, people don't do the following:

#include "TARGET_Freescale/mbed_rtx.h"

instead, they do this:

#include "mbed_rtx.h"

Knowing this, I remove collecting of include paths "through label
directories". Label directories are the things we scan for when building.
They start with TARGET_, TOOLCHAIN_ and FEATURE_. This PR collects
include path parents up to the first label dir.
I think an example is in order. Here is the Mbed OS targets directory:

targets
├── TARGET_ARM_FM
├── TARGET_ARM_SSG
├── TARGET_Analog_Devices
├── TARGET_Atmel
├── TARGET_Freescale
│  ├── TARGET_K20XX
│  ├── TARGET_KLXX
│  ├── TARGET_MCUXpresso_MCUS
│  └── mbed_rtx.h
├── TARGET_Maxim
├── TARGET_NORDIC
├── TARGET_NUVOTON
├── TARGET_NXP
├── TARGET_ONSEMI
├── TARGET_RENESAS
├── TARGET_Realtek
├── TARGET_STM
├── TARGET_Silicon_Labs
├── TARGET_TOSHIBA
├── TARGET_WICED
├── TARGET_WIZNET
├── TARGET_ublox
└── targets.json

Notice how there is no header file in the targets directory. When we
add include paths for the K64F, we would all parents of
targets/TARGET_Freescale/mbed_rtx.h. Which would be:

targets/TARGET_Freescale
targets

That's overkill. Instead this PR reduces this example to the following
include paths:

targets/TARGET_Freescale

While this may seem great, you would be hard pressed to measure this
improvement. It reduce the include paths by an entire 0.5% on MCC by
my measurements (20346, with this change; 20439 without).

Pull request type

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

[theotherjimmy]

@deepikabhavnani @cmonr This is not as good as I hoped. All well.

[cmonr]

Interesting.

[cmonr]

/morph build

[mbed-ci]

Build : SUCCESS

Build number : 2711
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/7623/

Triggering tests

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

[mbed-ci]

Exporter Build : FAILURE

Build number : 2340
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/7623/

[mbed-ci]

Test : SUCCESS

Build number : 2442
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/7623/2442

[cmonr]

@studavekar @ARMmbed/mbed-os-test Are y'all able to confirm that the SPIF lib is being setup correctly? Unsure if this is a PR, examples.py, or CI config issue.

[studavekar]

@studavekar @ARMmbed/mbed-os-test Are y'all able to confirm that the SPIF lib is being setup correctly? Unsure if this is a PR, examples.py, or CI config issue.

Looks like it was a GitHub issue https://mbed-os.mbedcloudtesting.com/job/build_wrapper/5352/console

00:05:07.027 ssh: Could not resolve hostname github.com: Temporary failure in name resolution
00:05:07.028 fatal: Could not read from remote repository.
00:05:07.028 
00:05:07.028 Please make sure you have the correct access rights
00:05:07.028 and the repository exists.
00:05:07.637 abort: HTTP Error 406: Not Acceptable

[cmonr]

/morph export-build

[mbed-ci]

Exporter Build : SUCCESS

Build number : 2352
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/7623/

[adbridge]

Relies on #7183 which is scheduled for 5.10