Codefest Grant - RubyGems cleanup and enhancement

[Eric Hodel]

--Apple-Mail-6--856733271
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII; format=flowed

Seattle.rb will be hosting a RubyGems cleanup and enhancement codefest!

We would like to solicit your ideas on what you want to see cleaned up
or enhanced in RubyGems.

For our Codefest we do not plan on making large changes to the way
RubyGems works. One thing we would like to focus on is making the gem
command more friendly when you hit ^C, for example.

Combining RPA and Gems is probably not going to be done during our
Codefest.

--
Eric Hodel - (e-mail address removed) - http://segment7.net
FEC2 57F1 D465 EB15 5D6E 7C11 332A 551C 796C 9F04

--Apple-Mail-6--856733271
content-type: application/pgp-signature; x-mac-type=70674453;
name=PGP.sig
content-description: This is a digitally signed message part
content-disposition: inline; filename=PGP.sig
content-transfer-encoding: 7bit

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)

iD8DBQFCMlOVMypVHHlsnwQRAmhlAKCd/gMOFAzvgynux365gGqQShmXmgCguOJX
UHAbyPvkvvyo/17IB7chhZk=
=+yJH
-----END PGP SIGNATURE-----

--Apple-Mail-6--856733271--

 

[Curt Hibbs]

Seattle.rb will be hosting a RubyGems cleanup and enhancement codefest!

We would like to solicit your ideas on what you want to see cleaned up
or enhanced in RubyGems.

For our Codefest we do not plan on making large changes to the way
RubyGems works. One thing we would like to focus on is making the gem
command more friendly when you hit ^C, for example.

This is probably not feasible, but I'll throw it out there anyway -- I'd
like to see a GUI front-end to RubyGems... a RubyGems browser of sorts. Rich
Kilmer started on one, but I don't think it got finished (if it did, I would
already have it in the One-Click Ruby Installer).

Curt

 

[James Britt]

This is probably not feasible, but I'll throw it out there anyway -- I'd
like to see a GUI front-end to RubyGems... a RubyGems browser of sorts. Rich
Kilmer started on one, but I don't think it got finished (if it did, I would
already have it in the One-Click Ruby Installer).

Isn't that what RubySlippers is?

James

 

[Curt Hibbs]

Isn't that what RubySlippers is?

Have you tried it? I would call it a primitive web front-end, and pretty
clunky at that.

What I want is something like a news reader, or windows file explorer. A
desktop GUI with a tree control on the left and a detail panel of sorts on
the right. With the ability to mount and browse multiple repositories as
well as your installed gems.

Curt

 

[gabriele renzi]

Eric Hodel ha scritto:

Seattle.rb will be hosting a RubyGems cleanup and enhancement codefest!

We would like to solicit your ideas on what you want to see cleaned up
or enhanced in RubyGems.

For our Codefest we do not plan on making large changes to the way
RubyGems works. One thing we would like to focus on is making the gem
command more friendly when you hit ^C, for example.

definitely, ATM, rubygems does not feel "fault tolerant".
One little thing I'd appreciate is some kind of feedback for the user
like a progress bar both in downloading and upgrading cache.
And i would greatly appreciate speed, both in the way "gem" works and in
how the require thing work.

Oh, and and it would be cool to be able to use a .gem file the way a
..jar is used in java (but this would require using zip instead of gzip,
AFAICT).

Combining RPA and Gems is probably not going to be done during our
Codefest.

 

[James Britt]

Have you tried it? I would call it a primitive web front-end, and pretty
clunky at that.

I've used it, and it's not bad, but not the native-OS app I would prefer.


James

 

[Richard Lyman]

This is probably not feasible, but I'll throw it out there anyway -- I'd
like to see a GUI front-end to RubyGems... a RubyGems browser of sorts. Rich
Kilmer started on one, but I don't think it got finished (if it did, I would
already have it in the One-Click Ruby Installer).

Curt

Yeah - I made a quick-n-dirty mockup that I gave Rich Kilmer and fully
expected pretty much everything to be redone, but I haven't seen
anything yet. For that matter, I haven't heard too much from him on
this list recently... being busy can be good though...

I really feel that until we have a _simple to use_ GUI (something like
synaptic), no method of 'making packages available' is going to appeal
to the general populous (sp?) until then... anyway...

I have no idea where my little mockup went, but it only took an hour
or so to do, so if someone wanted to, it wouldn't take them long to do
their own over again.

-Rich

 

[Curt Hibbs]

Yeah - I made a quick-n-dirty mockup that I gave Rich Kilmer and fully
expected pretty much everything to be redone, but I haven't seen
anything yet. For that matter, I haven't heard too much from him on
this list recently... being busy can be good though...

I really feel that until we have a _simple to use_ GUI (something like
synaptic), no method of 'making packages available' is going to appeal
to the general populous (sp?) until then... anyway...

I completely agree. That's why I'm *very* interested in getting a GUI
front-end to RubyGems into the One-Click installer, and on Windows and the
upcoming installer for OSX.

I have no idea where my little mockup went, but it only took an hour
or so to do, so if someone wanted to, it wouldn't take them long to do
their own over again.

Did you use FXRiuby? Anyway, I knew it was a long shot, but it would be
great if they decided to pick this up.

Curt

 

[Martin DeMello]

synaptic), no method of 'making packages available' is going to appeal
to the general populous (sp?) until then... anyway...

'populace'

martin

 

[Marcel Molina Jr.]

Seattle.rb will be hosting a RubyGems cleanup and enhancement codefest!

Glad to see this happen.

We would like to solicit your ideas on what you want to see cleaned up
or enhanced in RubyGems.

*gem query*

will list all gems matching a particular criteria. The
domain may be local, remote or both.

$ gem query -R -n ^m

*gem search*

will list all gems whose name "contains" the given
(case-insensitive) string. If no string is provided, all gems will
be listed.

$ gem search -r log

So one supports regex matches and the other doesn't? One is meant only for
remote queries and the other is indifferent?

There seems to be some overlap here. Perhaps these could be combined into one
command. Perhaps the documentation could be clarified to indicate more
expressly how they *do* differ.

marcel

 

[Richard Lyman]

Did you use FXRiuby? Anyway, I knew it was a long shot, but it would be
great if they decided to pick this up.

Curt

Yeah - FXRuby. If I wasn't so swamped with school stuff I'd be more
than happy to make a little mockup again... it wasn't that hard... an
FXTreeList on the left side, with some textfields and stuff on the
right for packet specific info...

Like I said though - it was a pretty poor mockup. I really hoped they
would take it to a much higher level (which is probably why I've been
content waiting...).

Maybe I shouldn't wait any more... it would be fun... grr... I _have_
to concentrate on school... This couldn't wait until the end of April
could it? I'd have ~20 a week free time then...

-Rich

 

[Vincent Foley]

On top of my head, I can think of a few things:

1) Having a command to remove all, but the most recent version of
installed gems (I had to clean up about 10 versions of Rails,
ActiveRecord and ActionPack the other day)

2) Having a "show" command, where there would be a detailed explanation
of what a particular gem does ("search" could also search that
description), a URL, an author email, etc.

3) Already mentionned by Gabriele, a cute progress bar (using
progressbar available in RubyGems)

Vincent.

 

[Curt Hibbs]

Yeah - FXRuby. If I wasn't so swamped with school stuff I'd be more
than happy to make a little mockup again... it wasn't that hard... an
FXTreeList on the left side, with some textfields and stuff on the
right for packet specific info...

Like I said though - it was a pretty poor mockup. I really hoped they
would take it to a much higher level (which is probably why I've been
content waiting...).

Maybe I shouldn't wait any more... it would be fun... grr... I _have_
to concentrate on school... This couldn't wait until the end of April
could it? I'd have ~20 a week free time then...

Well, it sure seems unlikely that anyone else is going to do it befoe then,
so I'd say go for it!

Curt

 

[George Moschovitis]

1. Resolve dependencies when installing local gems
2. better support for using just 'require' instead of require_gem

-g.

 

[Ben Giddings]

Combining RPA and Gems is probably not going to be done during our
Codefest.

Is there any chance you could start this process a little bit? Choose
some of the main features of RPA that are missing from RubyGems and find
a way of integrating them?

The ones that most interest me are:

== A way of differentiating "stable" and "unstable" packages ==

I don't think this should be a property of the gem, but maybe a
meta-property of some kind. The way Debian, Gentoo and the various BSDs
do things is an example. Someone determines which packages are stable
and which are not, and the user can choose to install the stable version
or the unstable one.

A given package with a given version number may get tested and be
determined to be stable, and be designated "stable". On the other hand,
it may be determined that it is unstable, and so a particular version
might never be made "stable".

== A way of interfacing with various other package managers ==

This may be pie-in-the-sky thinking, but wouldn't it be great if there
were no distinction between installing something via rubygems and
installing it via a package manager? In other words, a debian user
could choose to "gem install foolib" or "apt-get install foolib" and
both would do the same thing? A gentoo user could "gem install foolib"
then later decide he/she doesn't need it and "emerge -C foolib". A
windows user could remove gems using "add/remove programs"?

I think this would be so amazing, but I have no idea how difficult it
would be. If it could work though, I would absolutely love it.

(as an aside, what is the current RubyGems method of removing a gem?)

Ben

 

[Jim Weirich]

Ben Giddings said:

The ones that most interest me are:

== A way of differentiating "stable" and "unstable" packages ==

I don't think this should be a property of the gem, but maybe a
meta-property of some kind. The way Debian, Gentoo and the various BSDs
do things is an example. Someone determines which packages are stable
and which are not, and the user can choose to install the stable version
or the unstable one.

Part of this can be resolved by the ability to list different Gem sources.
This source (identified by URL) contains stable packages and that source
(a different URL) containts unstable packages. Gems can manage different
sources today, although the interface is primitive. The gems team is
thinking about how to make this more flexible (e.g. get gems from this
source, unless its not available, then get them from that source). This is
certainly something that the Seattle team could work on if they decided
to.

Beyond that, you need a team of people who are willing to test gems,
catogorized them, and then migrate them from unstable to stable.

(as an aside, what is the current RubyGems method of removing a gem?)

gem uninstall GEMNAME

(http://docs.rubygems.org/read/chapter/10#page38)

 

[vruz]

Is there any chance you could start this process a little bit? Choose

some of the main features of RPA that are missing from RubyGems and find
a way of integrating them?

Please make a list of features of RubyGems not present in RPA or the
rpa-base client, do the inverse operation and take note.

Now have a look at the codebase of both projects.

Now you understand why not.

 

[Ben Giddings]

Part of this can be resolved by the ability to list different Gem sources.
This source (identified by URL) contains stable packages and that source
(a different URL) containts unstable packages. Gems can manage different
sources today, although the interface is primitive. The gems team is
thinking about how to make this more flexible (e.g. get gems from this
source, unless its not available, then get them from that source). This is
certainly something that the Seattle team could work on if they decided
to.

Right, but I think you also need some "slots" on the user side. If they
want to keep up to date with the most current stable and unstable gem,
at the same time there should be a way to update them. It seems to me
that if you just have URLs once you've downloaded the gem from the
stable server, your system won't have any memory that that foolib-0.39
is the stable version of foolib.

Beyond that, you need a team of people who are willing to test gems,
catogorized them, and then migrate them from unstable to stable.

Right. I don't think that's too difficult though. If we don't have
people who can do this full-time, we could probably also come up with
some kind of a rating system where people can rate the libraries they
use. Libraries passing a certain threshold can be considered 'stable'.

gem uninstall GEMNAME

Ok, I see that option when I say 'gem help commands'. I guess I
expected to see it with 'gem help' or 'gem -h'. My bad.

Ben

 

[David Heinemeier Hansson]

Part of this can be resolved by the ability to list different Gem

sources.
This source (identified by URL) contains stable packages and that
source
(a different URL) containts unstable packages. Gems can manage
different
sources today, although the interface is primitive. The gems team is
thinking about how to make this more flexible (e.g. get gems from this
source, unless its not available, then get them from that source).

The Rails team already uses this with much success. Beta gems are
available from http://gems.rubyonrails.com has the edge, the public
repository has the stable. If you have a site to spare, this is already
a much easier approach since you can push as many unstable gems, and as
often as you like, without cluttering a main repository.

 

[James Edward Gray II]

Please make a list of features of RubyGems not present in RPA or the
rpa-base client, do the inverse operation and take note.

Now have a look at the codebase of both projects.

Now you understand why not.

This is not meant as an attack. I'm asking an honest question,
hopefully of the RPA developers. Is the RPA still under active
development? It seems to have slipped off the public radar. (Or
perhaps it's just my radar.)

James Edward Gray II