Fix `scalar diagnose` failures (#543)

This resolves #541.

The root cause here is that this loop that intended to get all files
from the `info` directory of the shared object cache _never worked_, but
previously we did not propagate that to a complete failure.

The reason it didn't work is that we tried to read file contents from
the directory path.

I also noticed a silly retry issue because we are now using `run_git()`
which does three retries due to concurrency issues in the functional
tests. We don't want to repeat failures three times in `scalar
diagnose`.
This commit is contained in:
Derrick Stolee 2022-11-08 10:58:35 -05:00 коммит произвёл Johannes Schindelin
Родитель 6bbfe3b836 71a6f10266
Коммит 2a17a368fb
1 изменённых файлов: 1 добавлений и 0 удалений

Просмотреть файл

@ -1351,6 +1351,7 @@ BASIC_CFLAGS += -DSHA1DC_FORCE_ALIGNED_ACCESS
endif
ifneq ($(filter leak,$(SANITIZERS)),)
BASIC_CFLAGS += -DSUPPRESS_ANNOTATED_LEAKS
BASIC_CFLAGS += -O0
SANITIZE_LEAK = YesCompiledWithIt
endif
ifneq ($(filter address,$(SANITIZERS)),)