Running the validation test suite (see environment details below), failed. But after some quick investigation, it seems this is a false negative.
Running run.sh ended with a
&nbps;&nbps;&nbps; $ ./run.sh -q
&nbps;&nbps;&nbps; => tfw_check.txt&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps; FAILED&nbps;&nbps;&nbps;&nbps;&nbps;&nbps;&nbps; <=
&nbps;&nbps;&nbps; FAILED (1)
Now, a comparison of test/ref/tfw_check.txt and test/res/tfw_check.txt, shows the difference in this:
From test/ref/tfw_check.txt:
&nbps;&nbps;&nbps; conf/t_no_operator_usage.aru: No syntax error
&nbps;&nbps;&nbps; conf/t_non_static.aru: No syntax error
From test/res/tfw_check.txt:
&nbps;&nbps;&nbps; conf/t_non_static.aru: No syntax error
&nbps;&nbps;&nbps; conf/t_no_operator_usage.aru: No syntax error
Actual result is the finally the same as the one expected from the reference result, except the two incriminated lines are swapped.
False negative are always confusing, especially if one do not check deeper in this case.
Environment details (if ever this is relevant):
* GCC 4.7.
* ASIS compiled from AdaCore Libre sources.
* AdaControl recompiled for that context.
I don't know if others see the same, so I don't mark it as reproducible.
Anonymous
This is a known issue due to different order of files between Windows and Linux. It is documented in the UG, parag. 2.1.6