]> git.proxmox.com Git - mirror_ubuntu-eoan-kernel.git/commit
kconfig: tests: test randconfig for choice in choice
authorMasahiro Yamada <yamada.masahiro@socionext.com>
Tue, 13 Mar 2018 09:12:10 +0000 (18:12 +0900)
committerMasahiro Yamada <yamada.masahiro@socionext.com>
Sun, 25 Mar 2018 17:04:04 +0000 (02:04 +0900)
commit3e4888c2e3d77d70edb905364cd70059a52a1343
tree2ae6e086d4442cc153be0e4d73c787fd3c98c45c
parentbeaaddb625400e4a561d2d8443f9df2aa3f9a899
kconfig: tests: test randconfig for choice in choice

Commit 3b9a19e08960 ("kconfig: loop as long as we changed some symbols
in randconfig") fixed randconfig where a choice contains a sub-choice.
Prior to that commit, the sub-choice values were not set.

I am not sure whether this is an intended feature or just something
people discovered works, but it is used in the real world;
drivers/usb/gadget/legacy/Kconfig is source'd in a choice context,
then creates a sub-choice in it.

For the test case in this commit, there are 3 possible results.

Case 1:
  CONFIG_A=y
  # CONFIG_B is not set

Case 2:
  # CONFIG_A is not set
  CONFIG_B=y
  CONFIG_C=y
  # CONFIG_D is not set

Case 3:
  # CONFIG_A is not set
  CONFIG_B=y
  # CONFIG_C is not set
  CONFIG_D=y
  CONFIG_E=y

So, this test iterates several times, and checks if the result is
either of the three.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
scripts/kconfig/tests/rand_nested_choice/Kconfig [new file with mode: 0644]
scripts/kconfig/tests/rand_nested_choice/__init__.py [new file with mode: 0644]
scripts/kconfig/tests/rand_nested_choice/expected_stdout0 [new file with mode: 0644]
scripts/kconfig/tests/rand_nested_choice/expected_stdout1 [new file with mode: 0644]
scripts/kconfig/tests/rand_nested_choice/expected_stdout2 [new file with mode: 0644]