Hash mismatch between build types about rawspeed HOT 8 CLOSED

That looks like mRaw/sRaw related. I guess the log file says that the md5sums are different but everything else is the same? Do you know when this problem appeared?

from rawspeed.

I guess the log file says that the md5sums are different but everything else is the same?

yes

Do you know when this problem appeared?

zero clues, may have been like that since always. this was the first time i did such a comparison

from rawspeed.

may have been like that since always.

Or maybe it is not there even now: I can't reproduce this. I am testing my 'own' release build based on some quick and dirty qmake file and that produces the same as the upstream cmake build when set to release type ASAN.

from rawspeed.

1. build with gcc, -DCMAKE_BUILD_TYPE=ASAN
2. delete all hashes, generate them
3. build with gcc, -DCMAKE_BUILD_TYPE=Release
4. verify hashes
5. ???
6. profit

from rawspeed.

Done. All good. Now where is my money? ;)

from rawspeed.

that is bad :(
this is gcc version 6.3.0 20170124 (Debian 6.3.0-5)

from rawspeed.

I am using c++ (Ubuntu 5.2.1-22ubuntu2) 5.2.1 20151010 right now.

here is the md5sum from the one failing 80D file: data md5sum: 65086dcc586860f0feadf1ac7ace6369 Maybe this fits with one or yours.

from rawspeed.

Apparently was fixed by 870daf9
Canon EOS 5D Mark IV m/s-raws are just completely broken anyway

from rawspeed.