Bug#338743: [Pkg-bluetooth-maintainers] Re: Bug#338743: [bluez-utils #338743] More info
On Fri, 2005-12-23 at 21:49 +0100, Filippo Giunchedi wrote: You are supposed to use the real MAKEDEV, not the one in experimental. reportedly, unstable (2.3.1-79) MAKEDEV bluetooth works as expected. Indeed this still holds as a bug for both bluez-utils and experimental MAKEDEV, but frankly, as long as I report version of packages I'm using, if I'm running experimental and if I'm supposed to, is not your business. Well, actually, it might be. If a bug exists in a package version in experimental but not in that package's version in unstable, then the maintainer of that package (which is me for makedev) probably wants to know about it, but it is not relevant to the maintainer of any other package, and many would prefer not to be bothered hearing about it. Part of the definition of experimental in Debian is that it may break the rest of your system to use things that are in experimental. If you plan to file a bug against the makedev package, please make sure to specify correctly the version that you're running in the bug submission, so that I don't waste time chasing ghosts. Suggested patches to improve makedev are always welcome... Bdale -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#338743: [bluez-utils #338743] More info
tags 338743 + patch
thanks
* Marco d'Itri ([EMAIL PROTECTED]) disait :
cd /dev ./MAKEDEV bluetooth
less /sbin/MAKEDEV
Thanks, here is attached a patch for the postinst script.
I plan to make a sponsored NMU for closing this RC bug in bluez-utils'
maintainer does not apply it.
Best regards,
--
Alexis Sukrieh [EMAIL PROTECTED]
0x1EE5DD34
Debian http://www.debian.org
Backup Manager http://www.backup-manager.org
diff -ubBr bluez-utils-2.19/debian/bluez-utils.postinst
bluez-utils-2.19.orig/debian/bluez-utils.postinst
--- bluez-utils-2.19/debian/bluez-utils.postinst2005-12-23
09:36:11.0 +0100
+++ bluez-utils-2.19.orig/debian/bluez-utils.postinst 2005-12-23
09:35:40.0 +0100
@@ -24,36 +24,9 @@
# Amended by Edd Dumbill [EMAIL PROTECTED] for
# the Debian distribution and to support devfs
- VHCI_MAJOR=10
- VHCI_MINOR=250
-
- RFCOMM_MAJOR=216
-
- # if devfs isn't mounted on /dev/ or we're not
- # using udev, then we need to create the device names
-
- if [ ! -e /dev/.devfsd -a ! -e /dev/.udevdb ]; then
- echo Checking and creating device nodes ...
- #
- # Create device for VHCI
- #
- if [ ! -c /dev/vhci ]; then
- mknod /dev/vhci c ${VHCI_MAJOR} ${VHCI_MINOR}
- chmod 660 /dev/vhci
- fi
-
- #
- # Create devices for RFCOMM
- #
- C=0;
- while [ $C -lt 256 ]; do
- if [ ! -c /dev/rfcomm$C ]; then
- mknod -m 660 /dev/rfcomm$C c
${RFCOMM_MAJOR} $C
- chgrp dialout /dev/rfcomm$C
- fi
- C=`expr $C + 1`
- done
- fi
+# Policy says we must not use mknod in maintainer scripts,
+# using MAKEDEV instead.
+cd /dev ./MAKEDEV bluetooth
# handle the devfs case
if [ -e /dev/.devfsd ]; then
Bug#338743: [bluez-utils #338743] More info
Can you verify the MAKEDEV constructs the appropriate devices exactly as the mknod from the postinst does? This was the reason I didn't use MAKEDEV in the first instance, albeit some time ago. On Fri, 2005-12-23 at 09:41 +0100, Alexis Sukrieh wrote: tags 338743 + patch thanks * Marco d'Itri ([EMAIL PROTECTED]) disait : cd /dev ./MAKEDEV bluetooth less /sbin/MAKEDEV Thanks, here is attached a patch for the postinst script. I plan to make a sponsored NMU for closing this RC bug in bluez-utils' maintainer does not apply it. Best regards, -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#338743: [bluez-utils #338743] More info
On Dec 23, Edd Dumbill [EMAIL PROTECTED] wrote: Can you verify the MAKEDEV constructs the appropriate devices exactly as the mknod from the postinst does? This was the reason I didn't use MAKEDEV in the first instance, albeit some time ago. If it does not then it would be a MAKEDEV bug, and your package would need to depend on the fixed version. -- ciao, Marco signature.asc Description: Digital signature
Bug#338743: [Pkg-bluetooth-maintainers] Re: Bug#338743: [bluez-utils #338743] More info
On Fri, Dec 23, 2005 at 07:21:30PM +0100, Marco d'Itri wrote: On Dec 23, Edd Dumbill [EMAIL PROTECTED] wrote: Can you verify the MAKEDEV constructs the appropriate devices exactly as the mknod from the postinst does? This was the reason I didn't use MAKEDEV in the first instance, albeit some time ago. If it does not then it would be a MAKEDEV bug, and your package would need to depend on the fixed version. it does not create devices, at least on powerpc with makedev 3.3.8.2-0: # cd /dev ./MAKEDEV bluetooth don't know how to make device bluetooth I'm CC'ing makedev maintainer, would be the appropriate for bluez-utils to drop bluetooth [1] into /etc/makedev.d/ or is this to be done by makedev package? thanks, filippo [1]: something like c 660 root root 10 250 1 1 vhci c 660 root dialout 216 1 1 256 rfcomm%d -- Filippo Giunchedi PGP key: 0x6B79D401 random quote follows: If only God would give me some clear sign! Like making a large deposit in my name in a Swiss bank. -- Woody Allen signature.asc Description: Digital signature
Bug#338743: [Pkg-bluetooth-maintainers] Re: Bug#338743: [bluez-utils #338743] More info
On Dec 23, Filippo Giunchedi [EMAIL PROTECTED] wrote: it does not create devices, at least on powerpc with makedev 3.3.8.2-0: You are supposed to use the real MAKEDEV, not the one in experimental. -- ciao, Marco signature.asc Description: Digital signature
Bug#338743: [Pkg-bluetooth-maintainers] Re: Bug#338743: [bluez-utils #338743] More info
On Fri, Dec 23, 2005 at 09:25:28PM +0100, Marco d'Itri wrote: On Dec 23, Filippo Giunchedi [EMAIL PROTECTED] wrote: it does not create devices, at least on powerpc with makedev 3.3.8.2-0: You are supposed to use the real MAKEDEV, not the one in experimental. reportedly, unstable (2.3.1-79) MAKEDEV bluetooth works as expected. Indeed this still holds as a bug for both bluez-utils and experimental MAKEDEV, but frankly, as long as I report version of packages I'm using, if I'm running experimental and if I'm supposed to, is not your business. filippo -- Filippo Giunchedi PGP key: 0x6B79D401 random quote follows: Gretchen: You're weird. Donnie: Sorry. Gretchen: No, that was a compliment. -- from Donnie Darko (2001) signature.asc Description: Digital signature
Bug#338743: [bluez-utils #338743] More info
Hello, I was looking for some RC bugs and found that one. I'd like to help if I can and I started investigating. The problem is that bluez-utils' postinst script creates two devices nodes: - /dev/vhci (major 10, minor 250) - /dev/rfcomm (major 216, minor 0-255) I looked at the MAKEDEV manpage but didn't find any information about how to create those nodes. Where can I find more information about that? Thanks. Regards, -- Alexis Sukrieh [EMAIL PROTECTED] 0x1EE5DD34 Debian http://www.debian.org Backup Manager http://www.backup-manager.org -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#338743: [bluez-utils #338743] More info
On Dec 22, Alexis Sukrieh [EMAIL PROTECTED] wrote: The problem is that bluez-utils' postinst script creates two devices nodes: - /dev/vhci (major 10, minor 250) - /dev/rfcomm (major 216, minor 0-255) I looked at the MAKEDEV manpage but didn't find any information about how to create those nodes. Where can I find more information about that? cd /dev ./MAKEDEV bluetooth less /sbin/MAKEDEV -- ciao, Marco signature.asc Description: Digital signature
