Re: Shouldn't update-manager's "check for updates" setting have an "hourly" option?
On Tue, 2010-06-22 at 22:49 -0400, Nathan Dorfman wrote: > Personally, I would prefer it, and I think it's quite reasonable. Thoughts? > You have to remember that not everyone has broadband and not everyone can download/upload that much. Checking hourly does add up to a lot of network activity and if you are in a third world country you wouldnt like it. Daily is fine we dont get that many updates to justify anything else. --fagan -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Shouldn't update-manager's "check for updates" setting have an "hourly" option?
On Wed, 2010-06-23 at 00:38 -0400, Nathan Dorfman wrote: > A fair point, but I think that up to 24 hours without a critical > security update could be undesirable in some situations. Certainly, I > think the default should remain "daily." For what it's worth, Fedora's > default is daily but it does provide an hourly option. I still think > it would be preferable for Ubuntu to have one as well... Wouldn't it be awesome if our machines had an optional xmpp connection to each of the servers and when there is a critical security update it could tell each of the machines to update their packages. Just throwing out a solution there, no reason machines couldn't be async. I wonder if xmpp can cope with millions of connections. Martin -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Shouldn't update-manager's "check for updates" setting have an "hourly" option?
On Tue, Jun 22, 2010 at 21:38, Nathan Dorfman wrote: > > On Tue, Jun 22, 2010 at 11:28 PM, Dylan McCall wrote: > > Keep in mind that checking for updates involves a non-trivial download > > of package lists from all repositories the user is subscribed to. > > Unfortunately, it is a much more intense operation than it appears. > > > > A fair point, but I think that up to 24 hours without a critical > security update could be undesirable in some situations. Certainly, I > think the default should remain "daily." For what it's worth, Fedora's > default is daily but it does provide an hourly option. I still think > it would be preferable for Ubuntu to have one as well... > > -- What would happen if the package lists were retrieved using zsync? Wouldn't that make it so there was less downloading, and the differences calculations would be on the client, not the server. -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Shouldn't update-manager's "check for updates" setting have an "hourly" option?
On Tue, Jun 22, 2010 at 11:28 PM, Dylan McCall wrote: > Keep in mind that checking for updates involves a non-trivial download > of package lists from all repositories the user is subscribed to. > Unfortunately, it is a much more intense operation than it appears. > A fair point, but I think that up to 24 hours without a critical security update could be undesirable in some situations. Certainly, I think the default should remain "daily." For what it's worth, Fedora's default is daily but it does provide an hourly option. I still think it would be preferable for Ubuntu to have one as well... -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Shouldn't update-manager's "check for updates" setting have an "hourly" option?
On Tue, Jun 22, 2010 at 7:49 PM, Nathan Dorfman wrote: > Personally, I would prefer it, and I think it's quite reasonable. Thoughts? > > -- > Ubuntu-devel-discuss mailing list > Ubuntu-devel-discuss@lists.ubuntu.com > Modify settings or unsubscribe at: > https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss > Keep in mind that checking for updates involves a non-trivial download of package lists from all repositories the user is subscribed to. Unfortunately, it is a much more intense operation than it appears. -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Shouldn't update-manager's "check for updates" setting have an "hourly" option?
million machines hitting the update servers every hour. hm On Jun 23, 2010, at 10:49 AM, Nathan Dorfman wrote: > Personally, I would prefer it, and I think it's quite reasonable. Thoughts? > > -- > Ubuntu-devel-discuss mailing list > Ubuntu-devel-discuss@lists.ubuntu.com > Modify settings or unsubscribe at: > https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Shouldn't update-manager's "check for updates" setting have an "hourly" option?
Personally, I would prefer it, and I think it's quite reasonable. Thoughts? -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: librxtx-java dependencies
"Onkar Shinde" wrote: >On Tue, Jun 22, 2010 at 11:56 PM, Scott Howard wrote: >> On Tue, Jun 22, 2010 at 9:24 AM, Augusto Brito wrote: >>> Hello >>> >>> The librxtx-java package lists java-runtime as a dependency, which in >>> turn brings along a lot of packages. Would it be reasonable if it >>> depended only on the java-runtime-headless? This new dependency could >>> reduce a lot of unused packages when installed on a server machine as is >>> my case. >> >> Thanks for pointing this out. This package has just been adopted by >> the Debian Java Packaging team, and a new version has just been >> uploaded to debian and maverick [1]. This new version does not depend >> on java-runtime. It actually does not depend on any java (which is a >> bug). I agree that it should at least depend on: default-jre-headless. > >No, it is not a bug. It was decided by Debian Java team that library >packages should not depend on runtime unless the package needs a >specific version of runtime (1.5 or 1.6). Do I understand correctly that you are saying the package depends on non-essential functionality, but doesn't declare that dependency and it's somehow OK? Scott K -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: librxtx-java dependencies
> No, it is not a bug. It was decided by Debian Java team that library > packages should not depend on runtime unless the package needs a > specific version of runtime (1.5 or 1.6). > > > Onkar > Thanks, I just got this clarified on the debian mailing list simultaneously. javahelper "fixed" that for me last upload without my noticing when I removed all runtimes and replaced them with java:Depends -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: librxtx-java dependencies
On Tue, Jun 22, 2010 at 11:56 PM, Scott Howard wrote: > On Tue, Jun 22, 2010 at 9:24 AM, Augusto Brito wrote: >> Hello >> >> The librxtx-java package lists java-runtime as a dependency, which in >> turn brings along a lot of packages. Would it be reasonable if it >> depended only on the java-runtime-headless? This new dependency could >> reduce a lot of unused packages when installed on a server machine as is >> my case. > > Thanks for pointing this out. This package has just been adopted by > the Debian Java Packaging team, and a new version has just been > uploaded to debian and maverick [1]. This new version does not depend > on java-runtime. It actually does not depend on any java (which is a > bug). I agree that it should at least depend on: default-jre-headless. No, it is not a bug. It was decided by Debian Java team that library packages should not depend on runtime unless the package needs a specific version of runtime (1.5 or 1.6). Onkar -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: librxtx-java dependencies
On Tue, Jun 22, 2010 at 9:24 AM, Augusto Brito wrote: > Hello > > The librxtx-java package lists java-runtime as a dependency, which in > turn brings along a lot of packages. Would it be reasonable if it > depended only on the java-runtime-headless? This new dependency could > reduce a lot of unused packages when installed on a server machine as is > my case. Thanks for pointing this out. This package has just been adopted by the Debian Java Packaging team, and a new version has just been uploaded to debian and maverick [1]. This new version does not depend on java-runtime. It actually does not depend on any java (which is a bug). I agree that it should at least depend on: default-jre-headless. I'll upload the fix for this. Thanks again. [1] http://packages.debian.org/sid/librxtx-java -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
librxtx-java dependencies
Hello The librxtx-java package lists java-runtime as a dependency, which in turn brings along a lot of packages. Would it be reasonable if it depended only on the java-runtime-headless? This new dependency could reduce a lot of unused packages when installed on a server machine as is my case. -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss