build system tests: Add some more notes about test internals

This commit is contained in:
Angus Gratton 2016-10-04 15:54:27 +11:00
parent 477d71e589
commit eccf54b939

View file

@ -2,7 +2,7 @@
#
# Test the build system for basic consistency
#
# Just a bash script that tests some likely make failure scenarios in a row
# A bash script that tests some likely make failure scenarios in a row
# Creates its own test build directory under TMP and cleans it up when done.
#
# Environment variables:
@ -11,6 +11,17 @@
# ESP_IDF_TEMPLATE_GIT - Can override git clone source for template app. Otherwise github.
# NOCLEANUP - Set to '1' if you want the script to leave its temporary directory when done, for post-mortem.
#
#
# Internals:
# * The tests run in sequence & the system keeps track of all failures to print at the end.
# * BUILD directory is set to default BUILD_DIR_BASE
# * The "print_status" function both prints a status line to the log and keeps track of which test is running.
# * Calling the "failure" function prints a failure message to the log and also adds to the list of failures to print at the end.
# * The function "assert_built" tests for a file relative to the BUILD directory.
# * The function "take_build_snapshot" can be paired with the functions "assert_rebuilt" and "assert_not_rebuilt" to compare file timestamps and verify if they were rebuilt or not since the snapshot was taken.
#
# To add a new test case, add it to the end of the run_tests function. Note that not all test cases do comprehensive cleanup
# (although very invasive ones like appending CRLFs to all files take a copy of the esp-idf tree.)
# Set up some variables
#