Re: Bug#800014: multipath-udeb: not installable: depends on libsystemd0
[ Adding debian-boot@lists.debian.org back. ] Ritesh Raj Sarraf(2015-09-28): > Adding systemd maintainers > > > Hello Cyril, > > THank you for the bug report. > > > On Fri, 2015-09-25 at 11:59 +0200, Cyril Brulebois wrote: > > Package: multipath-udeb > > Version: 0.5.0+git0.770e6d0d-3 > > Severity: grave > > Justification: renders package unusable > > > > > > [ Please keep debian-boot@lists.debian.org in the loop. ] > > > > Hi, > > > > Your package in unstable gained a dependency which cannot be > > satisfied: > > libsystemd0 has no udeb. It's therefore no longer installable, along > > with partman-multipath which depends on it. > > This was intentionally done because upstream added systemd dependency > some time ago. For Jessie, I had disabled systemd support, but now, it > makes sense to have it, and align with upstream and other > distributions. > > So, I think the real question is about why libsystemd0 does not have a > udeb ? > > systemd maintainers should provide that input. Some input from someone who doesn't know a thing about systemd or multipath: libmultipath.so.0 has a NEEDED entry on libsystemd.so.0, which seems to only define symbols named sd_* (which seems fair). The only use I see in libmultipath.so.0 is sd_listen_fds. Its manpage says: sd_listen_fds() shall be called by a daemon to check for file descriptors passed by the init system as part of the socket-based activation logic. There's no systemd in d-i, so there's no socket-based activation logic, and you can't rely on that in your multipath udeb. Mraw, KiBi. signature.asc Description: Digital signature
Re: Bug#800014: multipath-udeb: not installable: depends on libsystemd0
On Mon, 2015-09-28 at 18:53 +0200, Cyril Brulebois wrote: > [ Adding debian-boot@lists.debian.org back. ] > Sorry about that. > > This was intentionally done because upstream added systemd > > dependency > > some time ago. For Jessie, I had disabled systemd support, but now, > > it > > makes sense to have it, and align with upstream and other > > distributions. > > > > So, I think the real question is about why libsystemd0 does not > > have a > > udeb ? > > > > systemd maintainers should provide that input. > > Some input from someone who doesn't know a thing about systemd or > multipath: libmultipath.so.0 has a NEEDED entry on libsystemd.so.0, > which seems to only define symbols named sd_* (which seems fair). The > only use I see in libmultipath.so.0 is sd_listen_fds. > > Its manpage says: > >sd_listen_fds() shall be called by a daemon to check for file >descriptors passed by the init system as part of the socket > -based >activation logic. > > There's no systemd in d-i, so there's no socket-based activation > logic, > and you can't rely on that in your multipath udeb. Hmmm... I'll look into it later. But when you say "No systemd in d-i", does it mean that is how it is going to remain ? -- Ritesh Raj Sarraf | http://people.debian.org/~rrs Debian - The Universal Operating System signature.asc Description: This is a digitally signed message part
Re: Bug#800014: multipath-udeb: not installable: depends on libsystemd0
On Mon, Sep 28, 2015 at 11:17:42PM +0530, Ritesh Raj Sarraf wrote: > On Mon, 2015-09-28 at 18:53 +0200, Cyril Brulebois wrote: > > [ Adding debian-boot@lists.debian.org back. ] > > > > Sorry about that. > > > > > This was intentionally done because upstream added systemd > > > dependency > > > some time ago. For Jessie, I had disabled systemd support, but now, > > > it > > > makes sense to have it, and align with upstream and other > > > distributions. For the sake of derivatives without systemd, you might see if it's practical to leave the systemd dependency optional. -- hendrik > > > > > > So, I think the real question is about why libsystemd0 does not > > > have a > > > udeb ? > > > > > > systemd maintainers should provide that input. > > > > Some input from someone who doesn't know a thing about systemd or > > multipath: libmultipath.so.0 has a NEEDED entry on libsystemd.so.0, > > which seems to only define symbols named sd_* (which seems fair). The > > only use I see in libmultipath.so.0 is sd_listen_fds. > > > > Its manpage says: > > > >sd_listen_fds() shall be called by a daemon to check for file > >descriptors passed by the init system as part of the socket > > -based > >activation logic. > > > > There's no systemd in d-i, so there's no socket-based activation > > logic, > > and you can't rely on that in your multipath udeb. > > > Hmmm... I'll look into it later. But when you say "No systemd in d-i", > does it mean that is how it is going to remain ? > > -- > Ritesh Raj Sarraf | http://people.debian.org/~rrs > Debian - The Universal Operating System
Re: Bug#800014: multipath-udeb: not installable: depends on libsystemd0
Ritesh Raj Sarraf(2015-09-28): > Hmmm... I'll look into it later. But when you say "No systemd in d-i", > does it mean that is how it is going to remain ? I'm not sure why there would be a need to change that in the near future. Mraw, KiBi. signature.asc Description: Digital signature
Bug#800014: multipath-udeb: not installable: depends on libsystemd0
Package: multipath-udeb Version: 0.5.0+git0.770e6d0d-3 Severity: grave Justification: renders package unusable [ Please keep debian-boot@lists.debian.org in the loop. ] Hi, Your package in unstable gained a dependency which cannot be satisfied: libsystemd0 has no udeb. It's therefore no longer installable, along with partman-multipath which depends on it. Mraw, KiBi.