ivman does not automount devices with the hal versions greater than 0.5.5. Solution: echo ">sys-apps/hal-0.5.5.1-r3" >> /etc/portage/package.mask
Which ivman version?
all versions. the maintainer of ivman provided a patch to me, which should make it work again. with the new hal versions, i will try this and report back.
Ivman-0.6.10 + hal-0.5.7 was not working here neither for automounting cdroms, despite this standard rule: <ivm:Match name="ivm.mountable" value="true"> <ivm:Option name="mount" value="true" /> </ivm:Match> In debug mode, it complains about lack of a mount policy for the device. This is because "10-storage-policy.fdi" (which used to be part of the defaults with <=hal-0.5.5) has been dropped recently: http://lists.freedesktop.org/archives/hal/2006-January/004368.html Hence the device doesn't get the "storage.policy.should_mount" property, which i guess is what Ivman was looking for. Restoring this policy file from the CVS attic, as suggested in the above email, makes Ivman happy again. Maybe some einfo about that in hal ebuild, and/or a copy of this file in /usr/share/doc/hal-*/conf/ could be useful? But sure, if Ivman is the only package suffering from this change and it can be fixed, then it's even better. Florian, could you attach the patch you've got so that i test it too?
Created attachment 84812 [details, diff] Patch to make ivman work again with hal-0.5.7 Rohan McGovern sent me this patch, but it does not work for me.
The patch works for me (tested an usbstick and a cdrom). What it does is to allow mounting when the "storage.policy.should_mount" property is missing on both the volume and it's parent device, so sure it solves the problem i've described above with recent HAL not setting this property anymore. That said, this patch would have some drawback. In particular, the old "10-storage-policy.fdi" was setting "should_mount=false" on devices which were not hotpluggable or removable. With this property not beeing set anymore, and with Ivman accepting this as an implicit "yes", it will mount such devices. That's an important change of behavior which will have some unexpected consequencies for many users imho. To avoid that, i would suggest to also install, by default in HAL ebuild, the "storage-non-fixed.fdi" which is currently in its doc directory. This way: - hotpluggable devices (usb sticks, etc.) still won't have the 'should_mount' property, and will be mounted thanks to the Ivman patch - cdroms and floppies (which are not hotpluggable, but are special cases of the "storage-non-fixed.fdi" policy) will have 'should_mount' set to 'true', so Ivman will mount them too. - other not-hotpluggable devices (harddisk partitions, etc.) will have 'should_mount' set to 'false', so Ivman will continue to ignore them, like it was doing with old HAL versions because of the "10-storage-policy.fdi". In my opinion, this "ivman patch + storage-non-fixed.fdi" solution is still not has good as restoring the dead "10-storage-policy.fdi" (because this file had other subtilities, like hints for naming mount points, etc.), but it should make things work reasonably well, while staying very close to upstream's defaults. (In reply to comment #4) > but it does not work for me. What is you test case and what debug output do you get with an "ivman --nofork --debug" launched in a terminal (with no other ivman instance running)?
You need to reopen the bug after providing the requested information.
I think it is best, when you have a look at this yourself rohan.
Created attachment 84874 [details, diff] A better patch for Ivman with HAL 0.5.7 Here is a patch I committed to Ivman CVS just now. I have tested it with a CDROM and a digital camera with both HAL 0.5.5.1-r3 and 0.5.7, and it works for me. It should avoid the caveat mentioned in comment #5; the behaviour is now: 1) check for (volume|storage).policy.should_mount as usual; if it exists, do what it says (so HAL 0.5.5 works) 2) if (volume|storage).policy.should_mount is missing on volume and parent device, look at storage.removable on parent device and mount if it is true (so HAL 0.5.7 works, and non-removable drives aren't mounted) This patch doesn't require any changes to any config files in HAL or Ivman, and I can't think of any situations in which it wouldn't work. Please try this patch, or wait for Ivman 0.6.11 to be released (which should be once I fix Bug #129309).
thanks, I bumped the version to 0.6.11, should be available through rsync in an hour.
