genprog-code
genprog-code copied to clipboard
Error when running on Coreutils
Hi,
I have a similar issue with #20 , but I'm running on docker machine. The cil in installed by 1.7.3 and the developer's test can be run successfully, e.g. gcd-test-string.
However, when I use the same environment to apply on Coreutils, it throws an error.
After run ./bootstrap, CC="gcc --save-temps" ./configure, and make to generate .i file, we also generate a simple test.sh:
timeout 3m make check TESTS=tests/split/suffix-auto-length.sh VERBOSE=yes SUBDIRS=. ; [[ $? =~ "FAIL: 1" ]] && exit 1
By running ./genprog --compiler gcc --program src/split.c --test-script ./test.sh, it shows:
... cilRep: preprocessing IO function signatures: /tmp/tmpde04f0.c /tmp/tmp5fe916.c cilRep: done preprocessing IO function signatures Exception in loading: Sys_error("src/split.cache: No such file or directory") cilrep: from_source: pre: stmt_count = 0 cilRep: src/split.c: parsing Warning: Parsing raised Invalid_argument("index out of bounds")
split.c[49:0-0] : syntax error Parsing error./genprog aborting: Frontc.ParseError("Parse error") Rep: saving test cache
Variant Test Case Queries: 0 "Test Suite Evaluations": 0
Compile Failures: 0 Wall-Clock Seconds Elapsed: 0.016623 Activity Name Count Seconds = Percent of Total Time TOTAL 0.000 = 0% (avg CPU usage) Fatal error: exception Frontc.ParseError("Parse error")
It's interesting for only running ./genprog --program src/split.c with the same error.
Do you have any idea?
Hey there! Sorry for the delay in responding, took a back burner between the pandemic and my maternity leave. Sorry for asking what sounds like an obvious question, but: src/split.c is the preprocessed version, yes?