On Fri, Oct 30, 2009 at 06:08:27PM +0300, Sergei Golovan wrote: > Could you test this findings and hopefully confirm that the new expect version > works fine and just needs extra care when working with binary data?
Hi, With LANG=C or the fconfigure lines you suggested, it successfully runs past 32960 and the points at which it failed originally (I got bored somewhere past 50000000), so it looks like the failure at 32960 is encoding-related, and the failure usually at 624906 is fixed. I checked the source package to see if Daniel Jacobowitz's patch from http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=377332#10 had been applied, and it looks like the relevant code has been rewritten. Looking at my original bug report, I see I was still using LANG=C then (instead of en_GB.UTF-8) , which explains why I didn't run into the encoding problem originally. So, I think it is safe to close this bug. Ian. -- Ian Beckwith - i...@debian.org - i...@erislabs.net - http://erislabs.net/ianb/ GPG fingerprint: AF6C C0F1 1E74 424B BCD5 4814 40EC C154 A8BA C1EA Listening to: DJ Food - Raiding The 20th Century (Words & Music Expansion)
signature.asc
Description: Digital signature
