Oh, man! Someone of developers updated ebuild but did not rename it from captive-1.1.7.ebuild to captive-1.1.7-r1.ebuild. Please do not do so! emerge told be that there was no update for package :( I lost a work day for finding out reason. Shame on this developer :(
Uhm? Why should we revbump it? There's no need to revbump an ebuild because of dependencies update. http://sources.gentoo.org/viewcvs.py/gentoo-x86/sys-fs/captive/captive-1.1.7.ebuild?r1=1.2&r2=1.3 I don't see what sandbox crash and restart are you talking about, sorry.
Ok. At first I'm sorry for my rude tone/ I was very disapointed... I mean that old version of ebuild contained link to source package named captive-1.1.7cvs.tar.gz And the new one references captive-1.1.7.tar.gz New package contains a lot of fixes. But name of ebuild have not change and when I run emerge -avu world emerge told me nothing about improvements in this package. I hope it will help.
Of cource this ebuild works now, but is someone installed it earlier... I think that it is needed to publish new ebuild with newer version but the same content in order to force rebuild :( (but it is not programming, it's hack)
there have been only some utf8 changes (ie. minor, optical improvements) that do not justify a revision bump in my opinion. Can you tell me about your problem? I will sure revbump it if I see a reason for it.
Here is my post to captive-list: ================================================================== I'm trying to use captive with fuse. And it works for me quite good. But sometimes it crashes :( For example during unpacking sources of captive to ntfs drive. In syslog appears message. Filesystem crash broke dirty object As far as I understand from documentation sandbox master component must restart sandbox and continue from last succeded point. but it does not. Is iiit normal behaviour? Additional info: mount.captive-ntfs /dev/hda1 /mnt/tmp/ -o
Here is my post to captive-list: ================================================================== I'm trying to use captive with fuse. And it works for me quite good. But sometimes it crashes :( For example during unpacking sources of captive to ntfs drive. In syslog appears message. Filesystem crash broke dirty object As far as I understand from documentation sandbox master component must restart sandbox and continue from last succeded point. but it does not. Is iiit normal behaviour? Additional info: mount.captive-ntfs /dev/hda1 /mnt/tmp/ -o rw,--sandbox-server=/usr/libexec/captive-sandbox-server tar -xzf /usr/portage/distfiles/captive-1.1.7cvs.tar.gz ================================================================= And here is the ansver: ================================================================= On Sat, 25 Mar 2006 17:27:33 +0100, Antonio wrote: ... > Filesystem crash broke dirty object > > As far as I understand from documentation sandbox master component > must restart sandbox and continue from last succeded point. Yes... > but it does not. Is iiit normal behaviour? It was a bug in 1.1.6.1 only, it should be fixed in 1.1.7 already, I hope. ... > tar -xzf /usr/portage/distfiles/captive-1.1.7cvs.tar.gz ... > http://gentoo-wiki.com/User:Antonio Please check if your Gentoo-vendor package is really the final 1.1.7 - best try the "captive-static" from my web (if it is compatible enough with Gentoo). The "1.1.7cvs" looks as some pre-1.1.7 snapshot which may be without the "restart" fix - check "src/client/fuse/main.c" to be revision >=1.15. ================================================================= I checked and found this bug. Version of src/client/fuse/main.c in captive-1.1.7cvs.tar.gz was 1.14
thanks, I bumped the revision, thansk for reporting.
