[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Chicken-users] Need to do an install, any changes in the pipe line

From: Mario Domenech Goulart
Subject: Re: [Chicken-users] Need to do an install, any changes in the pipe line I should wait for?
Date: Sun, 10 Nov 2013 22:31:33 +0000
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/24.3.50 (gnu/linux)

Hi John,

On Sun, 10 Nov 2013 16:27:22 -0500 John Cowan <address@hidden> wrote:

> Mario Domenech Goulart scripsit:
>> If the user specified he/she wants intarweb 0.9, it makes little sense
>> to me fetching version 1.0.  If egg-pack-sources were to guess
>> combinations of versions that match egg requirements, another
>> alternative would be fetching some previous spiffy version that accepts
> Sure.  I was just saying that you now have sufficient logic to support
> pulling versionless eggs in the correct versions relative to eggs
> whose versions *are* specified.  I think this should be the default
> behavior when at least one egg version is given, actually.  Why would
> you want the latest versions of eggs when they are inconsistent with
> other eggs, except in the situations where --force-versions is
> appropriate?

They should not be inconsistent, unless there are bugs in eggs'
dependencies specification.  chicken-install will not cope with
inconsistencies anyway (modulo bugs).

--force-versions is a bit pointless in that sense, unless you want to
fetch a bunch of sources that will be installed separately.

> But if you don't like that, you could have two new switches
> --latest-versions (for the present behavior) and --consistent-versions
> (for the behavior I am proposing).

Eggs don't specify the exact version of dependencies they need.  You can
specify a requirement like "at least <version>", but not "exactly <version>".

So, as far as I can see, you always have a consistent set, unless eggs
(or chicken-install) are broken.

>> But in this particular case, I think the best thing to do is just abort
>> and leave the decision to the user.
> Fair enough: an explicit inconsistency without --force-versions should
> abort with an error.
> Another minor suggestion: a switch --install-script to specify the name of
>, when you have multiple possibilities in the same egg directory.

Thanks.  I've added that switch.

Best wishes.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]