https://bugzilla.novell.com/show_bug.cgi?id=400857 User coolo@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=400857#c2 Stephan Kulow changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO Info Provider| |mls@novell.com --- Comment #2 from Stephan Kulow 2008-06-18 01:36:40 MDT --- test case? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=400857 User coolo@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=400857#c4 Stephan Kulow changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO Info Provider| |aj@novell.com --- Comment #4 from Stephan Kulow 2008-06-23 08:30:58 MDT --- Aj, please verify you called the applydeltaiso on !ppc -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=400857 User coolo@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=400857#c6 Stephan Kulow changed: What |Removed |Added ---------------------------------------------------------------------------- Severity|Blocker |Minor --- Comment #6 from Stephan Kulow 2008-06-23 09:09:32 MDT --- yeah, the compressor behaves differently on powerpc than on !powerpc. So you need to apply deltas on the architecture they're aimed for. The reason seems to be this: i386: coolo@desdemona#build>./src/lzma/lzma -F alone -7 -v -v -v -v -v --lzma=dict=1 -c ../ltmain.sh | md5sum - bd07d5e67d9ed728930f94e7d1d601a3 - coolo@desdemona#build>./src/lzma/lzma -F alone -7 -v -v -v -v -v --lzma=dict=2 -c ../ltmain.sh | md5sum - 3f5ac41653c7735a442f42625316a15d - coolo@desdemona#build>./src/lzma/lzma -F alone -7 -v -v -v -v -v --lzma=dict=4 -c ../ltmain.sh | md5sum - 53a4f4938590ec243ebb4a5407816ae2 - coolo@desdemona#build>./src/lzma/lzma -F alone -7 -v -v -v -v -v --lzma=dict=8 -c ../ltmain.sh | md5sum - a2c40e5b0e48a944420cc33a618cff96 - On powerpc64 it's _almost_ the same, unless for >= 8. So I guess some long vs. int32 issue in the handling of the dictionary. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=400857 User coolo@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=400857#c8 Stephan Kulow changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|coolo@novell.com |mls@novell.com --- Comment #8 from Stephan Kulow 2008-06-23 23:31:00 MDT --- ok, I talked to lasse and he is aware that the crc tables are endian specific. He thinks, that hardcoding the endianess is going to work, but that in any case the next snapshot of lzma might have better compression anyway, so it will be different from 11.0 and we need to rethink how we create deltaisos anyway, because as it stands, you always need the applydeltaiso of the target system - which is kind of weird and only worked as long as bzip2 didn't change as it was in pure maintenance mode. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.