This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

GCC produces slightly different .o files on two consecutive runs in certain conditions

Hi all!

I am trying to do a setup in my C / gnu11
/ cmake based FOSS project so I can cache valgrind results. The project is and it has a git repo at .

Now, I've been running this fuzz building script -
- like "perl6 ../scripts/fuzz-build.p6 --rb -g" and while in clang it
reached seed=24 while producing identical recursive-build-dir output - in
gcc it failed in seed=6 with these two different outputs:

You can find the invocation of my cmake wrapping script in run-t-6.bash

--dbm-single-thread --dbm=kaztree --disable-err-strs --disable-flares
--disable-moves-track --disable-ni --disable-resume
--hard-code-reparent-states --judy --nfc=0 --nobuild-docs --num-stacks=13
--omit-frame --rcs --test-suite --unsafe --without-iter-handler
--without-max-num-states -r

this calls cmake like:


cmake -DCMAKE_BUILD_TYPE=release -DDATADIR=/usr/share


can anyone tell me how I can get gcc to produce identical .o files on
subsequent runs?


Shlomi Fish

Buddha has the Chuck Norris nature.

Please reply to list if it's a mailing list post - .

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]