[#386100] Numeric#coerce docs are disaster — 7stud -- <bbxx789_05ss@...>

num.coerce(numeric) =E2=86=92 array

14 messages 2011/08/02

[#386114] Documentation Improvement Proposal — Chris White <cwprogram@...>

=3D Issues =3D

24 messages 2011/08/02
[#386115] Re: Documentation Improvement Proposal — Steve Klabnik <steve@...> 2011/08/02

I reeeeeealy dislike user comments on documentation. It's one of the

[#386117] Re: Documentation Improvement Proposal — Phillip Gawlowski <cmdjackryan@...> 2011/08/02

On Tue, Aug 2, 2011 at 7:39 PM, Steve Klabnik <steve@steveklabnik.com> wrot=

[#386118] Re: Documentation Improvement Proposal — Steve Klabnik <steve@...> 2011/08/02

> What's wrong with stealing WikiPedia's procedures? The model works

[#386119] Re: Documentation Improvement Proposal — Chris White <cwprogram@...> 2011/08/02

On Aug 2, 2011, at 11:00 AM, Steve Klabnik wrote:

[#386123] Re: Documentation Improvement Proposal — Steve Klabnik <steve@...> 2011/08/02

Apologies, I've just responded to everyone in-line.

[#386231] Brainstorming ideas how to improve Ruby's documentation — Marc Heiler <shevegen@...>

The title is misleading...

42 messages 2011/08/05
[#386233] Re: Brainstorming ideas how to improve Ruby's documentation — "Fred L." <f.linard@...> 2011/08/05

Hello,

[#386235] Re: Brainstorming ideas how to improve Ruby's documentation — Alexander Litvinovsky <alexander.litvinovsky@...> 2011/08/05

What are you talking about? Ruby has a nice docs, railsapi.com for example.

[#386297] Help out with the next version of ruby-lang.org — Magnus Holm <judofyr@...>

https://github.com/rubylang/ruby-lang.org

11 messages 2011/08/07

[#386341] Exceptional Ruby and Metaprogramming Ruby has anyone picked these up? — Kevin <darkintent@...>

I'm thinking of picking up these two books and was wondering if anyone

11 messages 2011/08/09

[#386378] ruby installation — "Momodou J." <modou75alieu@...>

how to implement this in windows :

16 messages 2011/08/09

[#386401] *WHY* does this not work? — serialhex <serialhex@...>

ok, so code:

23 messages 2011/08/09
[#386403] Re: *WHY* does this not work? — "Darryl L. Pierce" <mcpierce@...> 2011/08/09

On Wed, Aug 10, 2011 at 03:52:59AM +0900, serialhex wrote:

[#386404] Re: *WHY* does this not work? — serialhex <serialhex@...> 2011/08/09

On Tue, Aug 9, 2011 at 3:05 PM, Darryl L. Pierce <mcpierce@gmail.com> wrote:

[#386480] Odd regexp behavior — Glen Holcomb <damnbigman@...>

I'm running 1.9.2-p180

16 messages 2011/08/10

[#386506] Distributing Ruby program as a standalone executable (exe) for windows — Michelle Pace <michelle@...>

Hello there,

10 messages 2011/08/11

[#386539] Online tutor for Ruby — T J Pereira <tj5155@...>

I am finding it difficult to apply the RUBY program. Its because i have

18 messages 2011/08/12
[#386541] Re: Online tutor for Ruby — Phillip Gawlowski <cmdjackryan@...> 2011/08/12

On Fri, Aug 12, 2011 at 6:00 AM, T J Pereira <tj5155@tm.net.my> wrote:

[#386637] class inheritance and class constants — Iñaki Baz Castillo <ibc@...>

------------------------

16 messages 2011/08/14

[#386784] Green Shoes v1.0 released — ashbb <ashbbb@...>

Hello, everyone.

15 messages 2011/08/18
[#392062] Re: Green Shoes v1.0 released — Barry Yu <yubarry@...> 2012/01/09

why do I get this error?

[#386796] Searching in a directory — Yu Yu <htwoo@...>

Hello,

21 messages 2011/08/18

[#386893] Gritty Details of super() — luke gruber <luke.gru@...>

Hey guys,

18 messages 2011/08/21

[#386900] Possble bug in Ruby parser (Fixnum#times within "case" statement) — Iñaki Baz Castillo <ibc@...>

Hi, I cannot find an explanation for the following issue so I think it's a =

15 messages 2011/08/21
[#386901] Re: Possble bug in Ruby parser (Fixnum#times within "case" statement) — Ryan Davis <ryand-ruby@...> 2011/08/21

[#386903] Re: Possble bug in Ruby parser (Fixnum#times within "case" statement) — Iñaki Baz Castillo <ibc@...> 2011/08/21

2011/8/22 Ryan Davis <ryand-ruby@zenspider.com>:

[#386920] New to Ruby some problems — jack jones <shehio_22@...>

I am new to Ruby, My mother tongue is C++ .. I have too many problems I

21 messages 2011/08/22

[#386949] Want to get involved with this doc stuff? I'm making it even easier — Steve Klabnik <steve@...>

Hey guys-

9 messages 2011/08/22

[#387058] How the access the values of this result — QAS WM <qaiserwali@...>

I am getting the following as a result of a script I run.

11 messages 2011/08/26

[#387070] overloading methods question please? — jack jones <shehio_22@...>

def do_something(a as Array)

11 messages 2011/08/26

[#387138] String#split resets regex captures variables (Ruby 1.8.7) — Olivier Lance <bestiol@...>

Hi,

10 messages 2011/08/29

[#387196] SAMSUNG to produce "Ruby on Rails in Silicon" System on a Chip — Ilias Lazaridis <ilias@...>

(public draft)

9 messages 2011/08/31

[#387197] Prepend a character to a string in ruby — ruby rails <rubyonrails4me@...>

Hi,

10 messages 2011/08/31

[#387212] GUI programming — Samuel Mensah <sasogeek@...>

Is ruby GUI programming something that will come along as I study ruby

19 messages 2011/08/31
[#387230] Re: GUI programming — Alexey Petrushin <axyd80@...> 2011/08/31

I believe right now it's better to stay with console, there's no Ruby

How do we fix gem permissions?

From: David Masover <ninja@...>
Date: 2011-08-12 00:44:40 UTC
List: ruby-talk #386530
I just recently tried a system rvm installation, as I want to run some Ruby 
scripts as isolated users who really don't need their own rvm or gems.

Here's the problem:

Gem permissions are retarded these days.

It seems like a gem is a tarball which wraps another couple of gzipped 
tarballs (really?), one of which is unpacked to the gem's final destination. 
Then some things are compiled if necessary, but the permissions are pretty 
much what the tarball sets.

That's useful when you want to have a binary file in bin, say, and you can 
just make it executable with chmod and the tarball will carry that through to 
the end-user.

It's problematic when I see this:

$ find $GEM_HOME/gems/x11-0.0.1a2 ! -perm /0044 | wc -l
130

Unless I'm missing something, this tells me that there are 130 files which the 
author of the 'x11' gem thinks that only the gem admin should be able to 
_read_.

To turn it around:

$ find $GEM_HOME/gems/x11-0.0.1a2 -perm /0044 | wc -l
21

In other words, there are only 21 files which can be read by a single user 
other than the person who installs gems. I cannot imagine this is what was 
intended, and my usual recipe for fixing this is:

$ find $GEM_HOME/gems/x11-0.0.1a2 -type f -exec chmod 644 '{}' \;
$ find $GEM_HOME/gems/x11-0.0.1a2 -type d -exec chmod 755 '{}' \;

...then fix anything that needs execute permission.

Since most of us have gone the lazy rvm-in-homedir way, this doesn't show up 
during gem development, but I imagine this violates the principle of least 
surprise for a lot of people -- you build a gem which works fine on your 
machine, and on all the dev machines, until someone tries to deploy it to a 
proper staging/test system with system gems, or until a developer wants system 
gems. Then, suddenly, you get stuff like this:

> require 'retarded'
LoadError: no such file to load -- retarded
...

I'm not making that up -- the gem is actually called 'retarded', and it's only 
a single file which my non-root user cannot load because someone set some 
screwy file permissions at some point.

Assuming I'm not the only person who cares about being able to install gems 
system-wide, or having gem permissions be sane on other machines, is there a 
good way to fix this? It seems to me that the obvious solution is to have 
files be mode 644, directories 755, unless explicitly overridden, maybe with 
some common shortcuts for making scripts executable -- and we already handle 
scripts specially anyway. In fact, I can't think of a single reason a gem ever 
needs anything to have permissions other than this.

Probably the least disruptive fix would be to change the gem building process, 
rather than the gem format. Tar is a perfectly fine format for storing file 
permissions, it's just that we shouldn't be getting them originally from the 
filesystem. It'd also be nice to have commandline options to override the 
tarball's permissions with what we can guess, in a similar way.

In This Thread

Prev Next