Hi! I tried to reproduce the bug in version zzuf_0.12-1.1 with a lenny machine (kernel 2.6.26-2-686).
When I compile it directly, all tests pass. When I run it in a sid cowbuilder environment, some tests pass, but even more tests fail compared to the test run described in the original bug report. I tried to look closer at the problem and found the interesting detail (in cowbuilder): r...@rigel:/tmp/533977/zzuf-0.12/test# ./../src/zzuf -s 58202476 -r 0.001 -P '\n' < ./file-text > testa.txt r...@rigel:/tmp/533977/zzuf-0.12/test# ./../src/zzuf -s 58202476 -r 0.001 -P '\n' sed -e n ./file-text > testb.txt When comparing testa.txt and testb.txt they differ. r...@rigel:/tmp/533977/zzuf-0.12/test# diff test?.txt | head -n 12 2c2 < tell him pinned to in appeared bleak and tHe was billowing out had bawned --- > tell him pinned to in appeared bleak and thE was billowing out had bawlgd 5c5 < get lynched by the voice said, - shou�ed - he knew that the main Universe --- > get lynched by the voice said, - shoute� - he knew that the main Universe 8c8 < Deep Thought is that morning, O�people born? Why not? - Funny, - said Ford. - --- > Deep Thought is that morning, O peo�le born? Why not? - Funny, - said Ford. - Outside cowbuilder, the two commands from above produce both the same output as testa.txt. When changing the parameter -r 0.001 to -r 0, the output files are in all cases (both commands in and outside cowbuilder) identical to the input file ./file-text In my case, not only the sed tests fail but also other tests (but not all of them) when I do the tests inside cowbuilder. In some cases, there appears an interesting error (inside cowbuilder): r...@rigel:/tmp/533977/zzuf-0.12/test# ./../src/zzuf -s 25155168 -r 10 -P '\n' < ./file-text | wc -c 32768 r...@rigel:/tmp/533977/zzuf-0.12/test# ./../src/zzuf -s 25155168 -r 10 -P '\n' cat ./file-text | wc -c E: cowdancer: .ilist header unexpected cowdancer: Fatal, initialize_functions failed E: cowdancer: .ilist header unexpected cat: ./file-text: Cannot allocate memory 0 r...@rigel:/tmp/533977/zzuf-0.12/test# ./../src/zzuf -m -s 25155168 -r 10 -P '\n' cat ./file-text E: cowdancer: .ilist header unexpected cowdancer: Fatal, initialize_functions failed E: cowdancer: .ilist header unexpected cat: ./file-text: Cannot allocate memory zzuf[s=25155168,r=10]: d41d8cd98f00b204e9800998ecf8427e The "result" MD5 sum of many tests in my cowbuilder case is d41d8cd98f00b204e9800998ecf8427e, which is the MD5 sum of an empty string. Only the sed test is different in my case: There are no cowbuilder error messages and the output has the same length as the original output as it should be; however, the result is not identical as shown above. Without cowbuilder, every tests passes as mentioned before. I am using a cowbuilder from the stable cowdancer 0.47 package. Best regards, Philipp Spitzer -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
