Re: fixing rc-version-greater-than-expected-version
On 07/31/2012 10:21 AM, Andreas Tille wrote: On Tue, Jul 31, 2012 at 10:01:53AM +0200, Alex Mestiashvili wrote: I would like to avoid adding the epoch in the middle of the minor releases, i.e. -betaX. Because the only reason is lintian's warning and not a significant change in the software. And I also agree with Charles that it is better to wait until we really need to add the epoch. If we will need to do so at all. Well, you are the person who is actively working on this package and if you have your way to make sure 1. you will notice new upstream releases 2. users will detect new package version this is fine for me whatever method you choose. Kind regards Andreas. Hi Andreas, first of all sorry for the delay. I'll go with overriding for 2.0.0-beta7 now and switch to epoch with the next release -beta8. Thank you and Charles for the useful discussion, Alex -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/502a109d.3020...@biotec.tu-dresden.de
Re: fixing rc-version-greater-than-expected-version
On Tue, Aug 14, 2012 at 10:47:25AM +0200, Alex Mestiashvili wrote: I'll go with overriding for 2.0.0-beta7 now and switch to epoch with the next release -beta8. If I were you I would simply leave the warning - there is no point in overriding anything that should be changed later anyway. Kind regards Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120814092331.gi6...@an3as.eu
Re: fixing rc-version-greater-than-expected-version
On 08/14/2012 11:23 AM, Andreas Tille wrote: On Tue, Aug 14, 2012 at 10:47:25AM +0200, Alex Mestiashvili wrote: I'll go with overriding for 2.0.0-beta7 now and switch to epoch with the next release -beta8. If I were you I would simply leave the warning - there is no point in overriding anything that should be changed later anyway. Kind regards Andreas. This definitely makes sense, but what if the package will be rejected because of this warning? Regards, Alex -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/502a1aa0.3010...@biotec.tu-dresden.de
Re: fixing rc-version-greater-than-expected-version
On Tue, Aug 14, 2012 at 11:30:08AM +0200, Alex Mestiashvili wrote: This definitely makes sense, but what if the package will be rejected because of this warning? As far as I know there is no warning which causes a rejection - this should only happen in case of a defined set of lintian errors (I admit I do not know the place where this set is documented). Kind regards Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120814122157.gb13...@an3as.eu
Re: fixing rc-version-greater-than-expected-version
On Tue, Jul 31, 2012 at 08:24:57AM +0900, Charles Plessy wrote: $ dpkg --compare-versions 2.0.0-beta6-1 gt 1:2.0.0 echo $? There are people who do not like epochs - but this is the clean way. indeed, I think that using an epoch is a bit too much at the moment. Could you please explain why you consider epoch as a bit too much? For instance, we do not know how long 2.0.0 will last, as sometimes minor corrections are released very shortly after a release, and packages with version numbers like 2.0.1 would solve our problem. I would recommend to ignore the warning until 2.0.0 is released. In this case version 2.0.0-beta6 users will not notice the final release. I fail to see your problem in using a technique which was invented exactly for this purpose. Kind regards Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120731061437.ga26...@an3as.eu
Re: fixing rc-version-greater-than-expected-version
Le Tue, Jul 31, 2012 at 08:14:37AM +0200, Andreas Tille a écrit : On Tue, Jul 31, 2012 at 08:24:57AM +0900, Charles Plessy wrote: $ dpkg --compare-versions 2.0.0-beta6-1 gt 1:2.0.0 echo $? There are people who do not like epochs - but this is the clean way. indeed, I think that using an epoch is a bit too much at the moment. Could you please explain why you consider epoch as a bit too much? Hi Andreas, since 2.0.0-beta6-1 is lower than 2.0.0-beta7-1, it is not strictly required to introduce an epoch at the moment, althouth of course you have given good reasons to predict that we may have to do so when 2.0.0 is released. For instance, we do not know how long 2.0.0 will last, as sometimes minor corrections are released very shortly after a release, and packages with version numbers like 2.0.1 would solve our problem. I would recommend to ignore the warning until 2.0.0 is released. In this case version 2.0.0-beta6 users will not notice the final release. I fail to see your problem in using a technique which was invented exactly for this purpose. When Upstream releases 2.0.0 we can package it as 1:2.0.0-1. If, however, 2.0.1 is released before we have time to upload 1:2.0.0-1 or even realise that 2.0.0 is out, then we do not need an epoch anymore. And I have experience of a package, bedtools, which had three minor release within a day: basically by the time I uploaded the package it was already outdated, twice in a row... So I propose to postpone the addition of the eopch for the day where somebody prepares thne next non-beta upload. Cheers, -- Charles -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120731071059.ga11...@falafel.plessy.net
Re: fixing rc-version-greater-than-expected-version
On Tue, Jul 31, 2012 at 04:10:59PM +0900, Charles Plessy wrote: In this case version 2.0.0-beta6 users will not notice the final release. I fail to see your problem in using a technique which was invented exactly for this purpose. When Upstream releases 2.0.0 we can package it as 1:2.0.0-1. Unfortunately also our automatic tools will not inform us about the 2.0.0 release and thus chances are good that we only realise 2.0.1 release ... whenever this will happen. For sure I'm writing this from a perspective of somebody who always needs to rely on what uscan is reporting because I will not manually watch specific download sites. So I propose to postpone the addition of the eopch for the day where somebody prepares thne next non-beta upload. I keep in failing to understand that you are that careful with the epochs method which in my eyes is nothing that should be avoided. Kind regards Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120731080151.ge26...@an3as.eu
Re: fixing rc-version-greater-than-expected-version
Le Tue, Jul 31, 2012 at 10:01:51AM +0200, Andreas Tille a écrit : On Tue, Jul 31, 2012 at 04:10:59PM +0900, Charles Plessy wrote: In this case version 2.0.0-beta6 users will not notice the final release. I fail to see your problem in using a technique which was invented exactly for this purpose. When Upstream releases 2.0.0 we can package it as 1:2.0.0-1. Unfortunately also our automatic tools will not inform us about the 2.0.0 release and thus chances are good that we only realise 2.0.1 release ... whenever this will happen. For sure I'm writing this from a perspective of somebody who always needs to rely on what uscan is reporting because I will not manually watch specific download sites. That is a good point. I am probably biased in that I only use the good old qa.debian.org page, which will display the version number in a different color whenever it mismatches. So let's use an epoch :) -- Charles -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120731081433.gb11...@falafel.plessy.net
Re: fixing rc-version-greater-than-expected-version
On 07/31/2012 08:14 AM, Andreas Tille wrote: On Tue, Jul 31, 2012 at 08:24:57AM +0900, Charles Plessy wrote: $ dpkg --compare-versions 2.0.0-beta6-1 gt 1:2.0.0 echo $? There are people who do not like epochs - but this is the clean way. indeed, I think that using an epoch is a bit too much at the moment. Could you please explain why you consider epoch as a bit too much? For instance, we do not know how long 2.0.0 will last, as sometimes minor corrections are released very shortly after a release, and packages with version numbers like 2.0.1 would solve our problem. I would recommend to ignore the warning until 2.0.0 is released. In this case version 2.0.0-beta6 users will not notice the final release. I fail to see your problem in using a technique which was invented exactly for this purpose. Kind regards Andreas. Hi Andreas, I would like to avoid adding the epoch in the middle of the minor releases, i.e. -betaX. Because the only reason is lintian's warning and not a significant change in the software. And I also agree with Charles that it is better to wait until we really need to add the epoch. If we will need to do so at all. Best regards, Alex -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/501790f1.2020...@gmail.com
Re: fixing rc-version-greater-than-expected-version
On Tue, Jul 31, 2012 at 10:01:53AM +0200, Alex Mestiashvili wrote: I would like to avoid adding the epoch in the middle of the minor releases, i.e. -betaX. Because the only reason is lintian's warning and not a significant change in the software. And I also agree with Charles that it is better to wait until we really need to add the epoch. If we will need to do so at all. Well, you are the person who is actively working on this package and if you have your way to make sure 1. you will notice new upstream releases 2. users will detect new package version this is fine for me whatever method you choose. Kind regards Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120731082156.gg26...@an3as.eu
fixing rc-version-greater-than-expected-version
Hi All, I just imported the new upstream version for bowtie2 and now lintian gives me the following warning: N: Processing binary package bowtie2 (version 2.0.0-beta7-1, arch i386) ... W: bowtie2: rc-version-greater-than-expected-version 2.0.0-beta7 2.0.0 (consider using 2.0.0~beta7) In general the message is totally clear 2.0.0-beta7 is higher than 2.0.0, but if I consider using 2.0.0~beta7 then my updated version will become lower than any previous version. Ffor example: dpkg --compare-versions 2.0.0-beta6-1 gt 2.0.0~beta7-1 echo $? returns 0 exit code... What should I do in such case? override the warning? Thank you, Alex -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/501682e0.5000...@gmail.com
Re: fixing rc-version-greater-than-expected-version
On Mon, Jul 30, 2012 at 02:49:36PM +0200, Alex Mestiashvili wrote: I just imported the new upstream version for bowtie2 and now lintian gives me the following warning: N: Processing binary package bowtie2 (version 2.0.0-beta7-1, arch i386) ... W: bowtie2: rc-version-greater-than-expected-version 2.0.0-beta7 2.0.0 (consider using 2.0.0~beta7) In general the message is totally clear 2.0.0-beta7 is higher than 2.0.0, but if I consider using 2.0.0~beta7 then my updated version will become lower than any previous version. Ffor example: dpkg --compare-versions 2.0.0-beta6-1 gt 2.0.0~beta7-1 echo $? returns 0 exit code... Well, at some point in time you somehow did a bad choice for the RC versions (and to prevent this the lintian check was invented - probably to late for your case). What should I do in such case? override the warning? I'd call this a dangerous way because it will close your eyes in such cases in the future. You should decide about your plan how to number the real bowtie 2.0.0 release because also $ dpkg --compare-versions 2.0.0-beta6-1 gt 2.0.0 echo $? 0 is the case. If you simply override the warning users of 2.0.0-beta6-1 do not see any upgrade path on their machines before, say 2.0.1~alpha1. The proper way to deal with such problems is using an epoch: $ dpkg --compare-versions 2.0.0-beta6-1 gt 1:2.0.0 echo $? There are people who do not like epochs - but this is the clean way. Kind regards Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120730141714.gd27...@an3as.eu
Re: fixing rc-version-greater-than-expected-version
On 07/30/2012 04:17 PM, Andreas Tille wrote: On Mon, Jul 30, 2012 at 02:49:36PM +0200, Alex Mestiashvili wrote: I just imported the new upstream version for bowtie2 and now lintian gives me the following warning: N: Processing binary package bowtie2 (version 2.0.0-beta7-1, arch i386) ... W: bowtie2: rc-version-greater-than-expected-version 2.0.0-beta7 2.0.0 (consider using 2.0.0~beta7) In general the message is totally clear 2.0.0-beta7 is higher than 2.0.0, but if I consider using 2.0.0~beta7 then my updated version will become lower than any previous version. Ffor example: dpkg --compare-versions 2.0.0-beta6-1 gt 2.0.0~beta7-1 echo $? returns 0 exit code... Well, at some point in time you somehow did a bad choice for the RC versions (and to prevent this the lintian check was invented - probably to late for your case). True, you read my mind :) What should I do in such case? override the warning? I'd call this a dangerous way because it will close your eyes in such cases in the future. You should decide about your plan how to number the real bowtie 2.0.0 release because also $ dpkg --compare-versions 2.0.0-beta6-1 gt 2.0.0 echo $? 0 is the case. If you simply override the warning users of 2.0.0-beta6-1 do not see any upgrade path on their machines before, say 2.0.1~alpha1. I see, thank you for the hint. The proper way to deal with such problems is using an epoch: $ dpkg --compare-versions 2.0.0-beta6-1 gt 1:2.0.0 echo $? this depend on the way upstream will call the non-beta version. In case of version clash it is fine for me to start use epochs. But what to do if I want to upload the package ? Will it be fine to upload it with the warning ? There are people who do not like epochs - but this is the clean way. Kind regards Andreas. Thank you, Alex -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/5016a374.1030...@gmail.com
Re: fixing rc-version-greater-than-expected-version
On Mon, Jul 30, 2012 at 05:08:36PM +0200, Alex Mestiashvili wrote: $ dpkg --compare-versions 2.0.0-beta6-1 gt 1:2.0.0 echo $? this depend on the way upstream will call the non-beta version. In case of version clash it is fine for me to start use epochs. But what to do if I want to upload the package ? Will it be fine to upload it with the warning ? The warning will vanish because the version including epoch is higher. Kind regards Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120730200427.ga12...@an3as.eu
Re: fixing rc-version-greater-than-expected-version
On Mon, Jul 30, 2012 at 02:49:36PM +0200, Alex Mestiashvili wrote: N: Processing binary package bowtie2 (version 2.0.0-beta7-1, arch i386) ... W: bowtie2: rc-version-greater-than-expected-version 2.0.0-beta7 2.0.0 (consider using 2.0.0~beta7) Le Mon, Jul 30, 2012 at 04:17:14PM +0200, Andreas Tille a écrit : The proper way to deal with such problems is using an epoch: $ dpkg --compare-versions 2.0.0-beta6-1 gt 1:2.0.0 echo $? There are people who do not like epochs - but this is the clean way. Hi Alex and Andreas, indeed, I think that using an epoch is a bit too much at the moment. For instance, we do not know how long 2.0.0 will last, as sometimes minor corrections are released very shortly after a release, and packages with version numbers like 2.0.1 would solve our problem. I would recommend to ignore the warning until 2.0.0 is released. Have a nice day, -- Charles Plessy Debian Med packaging team, http://www.debian.org/devel/debian-med Tsurumi, Kanagawa, Japan -- To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120730232457.ga10...@falafel.plessy.net