https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97090
--- Comment #8 from David Malcolm <dmalcolm at gcc dot gnu.org> ---
I tested with a cross build on x86_64-pc-linux-gnu with
target==powerpc64-unknown-linux-gnu after various fixes for non-determinism
(g:f635f0ce87d687b177b734968f18226d50499e75) and I'm not seeing the bogus
diagnostic at line 170.
I ran 100 iterations (with address-space randomization enabled), via this
script:
SRC_DIR=$(dirname $1)
FILENAME=$(basename $1)
rm $FILENAME.*
for i in `seq 1 100`; do
echo "iteration: $i"
./xgcc -B. -fanalyzer -c $SRC_DIR/$FILENAME -m64
-fdiagnostics-plain-output -fanalyzer -Wanalyzer-too-complex
-fanalyzer-call-summaries -fdump-analyzer -fanalyzer-call-summaries -S
-fdump-analyzer-supergraph -fdump-analyzer-exploded-graph -fdump-analyzer
-fdump-noaddr -fdump-analyzer-exploded-nodes-2
mv $FILENAME.supergraph.dot $FILENAME.$i.supergraph.dot
mv $FILENAME.analyzer.txt $FILENAME.$i.analyzer.txt
mv $FILENAME.supergraph-eg.dot $FILENAME.$i.supergraph-eg.dot
mv $FILENAME.eg.txt $FILENAME.$i.eg.txt
mv $FILENAME.eg.dot $FILENAME.$i.eg.dot
done
md5sum $FILENAME.* | sort | less
and got identical md5sums for every iteration of each of the various logs.
So perhaps the determinism fixes have removed flakiness from this test.
Is anyone still seeing that test failure on any configuration, after
g:f635f0ce87d687b177b734968f18226d50499e75 ?
(I still need to investigate the other failures reported by Jakub in comment
#3)
