On Thu, May 07, 2009 at 10:05:40AM +0400, Eygene Ryabinkin wrote: > Wed, May 06, 2009 at 09:39:09PM -0700, Tim Kientzle wrote: > > I tried but could not reproduce your bug using > > a recent -CURRENT. I downloaded this file: > > > > http://ftp.cse.yzu.edu.tw/pub/FreeBSD/distfiles/openoffice.org2/OOo_OOH680_m18_source.tar.bz2 > > > > and extracted ooo_crystal_images-1.tar.gz > > and got the same MD5 that Eygene reported. > > > > $ md5 ooo_crystal_images-1.tar.gz > > MD5 (ooo_crystal_images-1.tar.gz) = ff0d576d4b0e71c268b1516095a3d085 > > > > Where did you download your file from? > > Alexey's .tar.gz file should be also fine: he reported the checksum from > the .tar file, not .tar.gz one: > ----- > ~/tmp> md5 crystal.tar > MD5 (crystal.tar) = f9d09511003da0f59d943311a678b94a > ----- > I have the same MD5 for the .tar file: > ----- > $ md5 crystal.tar > MD5 (crystal.tar) = f9d09511003da0f59d943311a678b94a > ----- > > I am not able to reproduce the bug too, so perhaps the output from > ktrace, > ----- > ktrace -f gunzip.trace gunzip -c ooo_crystal_images-1.tar.gz | (ktrace -f tar.trace tar -xf -) > ----- > will be useful for diagnostics. The files of interest will be > gunzip.trace and tar.trace. > Ok, ktraces/kdumps on my system show that tar is exiting before the last gunzip's write :( gunzip.dump.human shows the last successful write on gunzip side and tar.dump.human shows the read sequence on tar side. gunzip has written 65536 bytes whereas tar has read 6 x 10240 = 61440 bytes and has exited thereafter. gunzip has got successful write operation (65536 bytes). At last gunzip tries to write 4932 zeros and fails... The command is exactly as above. FWIW: ~> echo $SHELL /bin/tcsh Alexey.
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:47 UTC