Re: [Fink-devel] Implicit epoch for distributions

Kyle Moffett wrote: On Nov 10, 2005, at 00:19:59, Blair Zajac wrote: Dave, Can epochs have versions, such as 1.2? We could take the existing epochs and prepend '1.' for 10.1, '2.' for 10.2, etc. That way the existing epochs are smaller than the new epochs automatically. If you read the dp

Re: [Fink-devel] eterm maintainership

Rogue wrote: Hi - can I request maintainership for eterm? I don't suppose there will be much to do but I like to know it is still alive in the fink tree and there is a port of call for users if things break. I'll be working down the Dependencies for unmaintained packages too. Best Regards, Ro

Re: [Fink-devel] ConfFiles and Python packages variants

Andrea Riciputi wrote: [] AFAICU this is the reason why dpkg fails when you try to replace -py24 variant with the -py23 one. [] On Nov 9, 2005, at 17:35 , Sébastien Maret wrote: I've downloaded an update of the pyx-py from the trackers (#1306090). The package installs a conf file in /sw/etc /

Re: [Fink-devel] Implicit epoch for distributions

Chris Zubrzycki wrote: On Nov 8, 2005, at 9:33 PM, Blair Zajac wrote: Do we have an implicit "epoch" for distributions, so that a package built under 10.3 will be implicitly a lower version than the same package version and revision number under 10.4-t? [] Long ago I repeatedly proposed putt

Re: [Fink-devel] Implicit epoch for distributions

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Nov 9, 2005, at 9:34 PM, David R. Morrison wrote: On Nov 8, 2005, at 9:33 PM, Blair Zajac wrote: Do we have an implicit "epoch" for distributions, so that a package built under 10.3 will be implicitly a lower version than the same package

Re: [Fink-devel] Implicit epoch for distributions

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Nov 10, 2005, at 12:22 AM, Kyle Moffett wrote: On Nov 10, 2005, at 00:19:59, Blair Zajac wrote: Dave, Can epochs have versions, such as 1.2? We could take the existing epochs and prepend '1.' for 10.1, '2.' for 10.2, etc. That way the ex

Re: [Fink-devel] Implicit epoch for distributions

On Nov 10, 2005, at 00:19:59, Blair Zajac wrote: Dave, Can epochs have versions, such as 1.2? We could take the existing epochs and prepend '1.' for 10.1, '2.' for 10.2, etc. That way the existing epochs are smaller than the new epochs automatically. If you read the dpkg docs, the answer

Re: [Fink-devel] Implicit epoch for distributions

David R. Morrison wrote: On Nov 8, 2005, at 9:33 PM, Blair Zajac wrote: Do we have an implicit "epoch" for distributions, so that a package built under 10.3 will be implicitly a lower version than the same package version and revision number under 10.4-t? This would be handy, as we wouldn

Re: [Fink-devel] How to determine required dependencies?

On 11/9/05, Trevor Harmon wrote: > Unfortunately, I don't know of any such tool for Java packages, which > is what I've mainly been submitting to Fink. Same problem with Perl, > Python, and so on. Anybody know of otool equivalents for these > languages? For python, try 'python -v your-script-here

[Fink-devel] Re: How to determine required dependencies?

David Bacher <[EMAIL PROTECTED]> writes: > In pseudocode: > > * Build and install your package to an isolated location > > * find | file | grep to create a list of all binary objects > > * otool -L each binary | grep /sw > > * for each dependency in /sw, look up its package > > I vaguely recall a

Re: [Fink-devel] How to determine required dependencies?

On Nov 9, 2005, at 5:19 PM, Rogue wrote: I disagree respectfully. I have in the past avoided (trying out) many packages because the Dep list was unreasonably long. I tend to be minimalist, not the safest approach undoutedly, but it is worth it in the long run IMHO. Indirect dependencies i

Re: [Fink-devel] How to determine required dependencies?

On Wednesday, 09 November 2005 at 22:24, Kevin Horton wrote: > On 9-Nov-05, at 19:30 , David Bacher wrote: > It isn't obvious to me how I interpret the output of otool -L. Some > lines can be connected to obvious fink packages, but many don't > directly match up. Is there some way to find wh

Re: [Fink-devel] How to determine required dependencies?

On 9-Nov-05, at 19:30 , David Bacher wrote:On 11/9/05, Trevor Harmon <[EMAIL PROTECTED]> wrote: I do basically the same thing. It's largely a trial-and-errorprocess, unfortunately. And even then, sometimes you still don't knowwhether there is some dynamic dependency that you don't know aboutuntil i

Re: [Fink-devel] Implicit epoch for distributions

On Nov 8, 2005, at 9:33 PM, Blair Zajac wrote: Do we have an implicit "epoch" for distributions, so that a package built under 10.3 will be implicitly a lower version than the same package version and revision number under 10.4-t? This would be handy, as we wouldn't have to deal with havin

Re: [Fink-devel] eterm maintainership

On 11/9/05, Rogue <[EMAIL PROTECTED]> wrote: > Hi - can I request maintainership for eterm? I don't suppose there will be > much to do but I like to know it is still alive in the fink tree and there is > a port of call for users if things break. I'll be working down the > Dependencies for unmain

Re: [Fink-devel] How to determine required dependencies?

On Wed, 9 Nov 2005 16:12:33 -0800 Trevor Harmon <[EMAIL PROTECTED]> wrote: > Also, I generally try to overestimate in deciding what packages go > into Depends. If I see something only looks like it *might* be a > dependency, I go ahead and add it. Better to have too many > dependencies than

Re: [Fink-devel] How to determine required dependencies?

On 11/9/05, Trevor Harmon <[EMAIL PROTECTED]> wrote: I do basically the same thing. It's largely a trial-and-errorprocess, unfortunately. And even then, sometimes you still don't knowwhether there is some dynamic dependency that you don't know aboutuntil it fails at runtime. I'm sure it doesn't

[Fink-devel] mldonkey

hi, could you please update mldonkey? fink has version 2.5.21, current is 2.6.7 which requires an updated ocaml. regards kp --- SF.Net email is sponsored by: Tame your development challenges with Apache's Geronimo App Server. Download it for

Re: [Fink-devel] Re: Fink binary distributions

On 11/8/05, Brendan Cully <[EMAIL PROTECTED]> wrote: > On Tuesday, 08 November 2005 at 18:21, Alexander K. Hansen wrote: > > Speaking for myself, I can now try to expedite some packages out of > > the tracker. Anybody reading this thread who's got a package sitting > > around, please email me (bet

[Fink-devel] eterm maintainership

Hi - can I request maintainership for eterm? I don't suppose there will be much to do but I like to know it is still alive in the fink tree and there is a port of call for users if things break. I'll be working down the Dependencies for unmaintained packages too. Best Regards, Rogue PS: does a

Re: [Fink-devel] How to determine required dependencies?

On Nov 9, 2005, at 3:12 PM, Koen van der Drift wrote: I'm not an 'old pro', but what I usually do for a new package is to remove all non-essential packages (eg with FinkCommander). And then start from scratch to compile the package. Then one by one I start to add packages based on the error

Re: [Fink-devel] How to determine required dependencies?

On Nov 9, 2005, at 6:01 PM, Kevin Horton wrote: I have created a few fink packages, but I always find it a struggle to determine what to list as Depends, and BuildDepends. Sometimes I can glean a few hints from studying the output of the build process, and sometimes a failed build is a go

[Fink-devel] How to determine required dependencies?

I have created a few fink packages, but I always find it a struggle to determine what to list as Depends, and BuildDepends. Sometimes I can glean a few hints from studying the output of the build process, and sometimes a failed build is a good clue. But I am never sure to have captured ev

Re: [Fink-devel] ConfFiles and Python packages variants

Hi Sébastian, I've added the ConfFiles field to the pyx info file exactly with the idea that it would have been shared between different variants. Anyway re-reading the packaging manual I've noticed that (citing from the manual): When a package is "remove"d, the configuration files will re

[Fink-devel] ConfFiles and Python packages variants

I've downloaded an update of the pyx-py from the trackers (#1306090). The package installs a conf file in /sw/etc / that is listed in the ConfFiles: field of the info file. However, when trying to two variant (-py23 and -py24), dpkg complains about overwriting the file: /sw/bin/dpkg: error

Re: [Fink-devel] perlmodule example files

Koen van der Drift wrote: Hi, What's a proper location to install example files that come with a perl module? The package bioperl-pm that I maintain stores a bunch of those files in /sw/bin. Which prompts the validator to give the following message: ... A possibility could be /sw/share/%n/

[Fink-devel] perlmodule example files

Hi, What's a proper location to install example files that come with a perl module? The package bioperl-pm that I maintain stores a bunch of those files in /sw/bin. Which prompts the validator to give the following message: Error: File in a perl-versioned package is neither versioned nor

Re: [Fink-devel] Re: Packages in the tracker to kickstart ...

On 9-Nov-05, at 07:09 , Kevin Horton wrote: On 8-Nov-05, at 22:20 , Alexander K. Hansen wrote: On 11/8/05, Alexander K. Hansen <[EMAIL PROTECTED]> wrote: On 11/8/05, Kevin Horton <[EMAIL PROTECTED]> wrote: On 8-Nov-05, at 18:21 , Alexander K. Hansen wrote: Speaking for myself, I can now tr

Re: [Fink-devel] Re: Packages in the tracker to kickstart ...

On 8-Nov-05, at 22:20 , Alexander K. Hansen wrote: On 11/8/05, Alexander K. Hansen <[EMAIL PROTECTED]> wrote: On 11/8/05, Kevin Horton <[EMAIL PROTECTED]> wrote: On 8-Nov-05, at 18:21 , Alexander K. Hansen wrote: Speaking for myself, I can now try to expedite some packages out of the tracke

Re: [Fink-devel] Re: Fink binary distributions

I've also a package sitting in the tracker for a long time: It is an update for the new upstream version of IPython (ipython.scipy.org/>) that has been in the Fink distribution for a long time. I've been u