Hi folks,
I was talking with Rolf on IRC, and I mentioned that a while ago I'd
brainstormed a list of features that would have been useful to my
workplace in particular.
Here are the list of things that would be appealing for Gitorious in
my workplace. These are just some rough ideas.
== Large f
Hi Gitorious devs,
I have a couple merge requests pending for libdolt and dolt, and it
would be great to see
these merged upstream. I've listed them in order of priority.
https://gitorious.org/gitorious/libdolt/merge_requests/8 - This should
allow Jenkins builds to pass again for libdolt.
https:
Hi folks,
There was some work in master recently that referenced issue #7. For example:
commit 74617cfcb0abaae7e34d5a10abac609d9e98d5f1
Author: Pawel Pierzchala
Date: Fri Mar 7 15:24:52 2014 +0100
Tweak sorting ui. Refs #7
However, this page is a 404 error for me:
https://gitorious.org/
Hi folks,
I have a couple merge requests pending, and it would be great to see
these merged upstream. I've listed them in order of ease-of-review.
Correct spelling of LDAPAuthenticationTest in comments:
https://gitorious.org/gitorious/mainline/merge_requests/256
Remove Ruby 1.8 support from dolt
On Wed, Jan 15, 2014 at 12:08 PM, Katie Poskaitis
wrote:
> Found the issue. Our passwords expire after a single use and it tries to use
> your password 3 times when logging in an registering the user.
Gitorious only binds once, though:
79 # If no bind user has been specified, bind directl
> # Map LDAP fields to database fields.
> # Default: displayname => fullname, mail => email
> attribute_mapping:
> cn: fullname
Try explicitly adding the "mail: email" mapping here, and let me know
how that works.
I had a similar problem with my installation a while back
On Mon, Jan 6, 2014 at 9:43 AM, Justin Smith wrote:
> I am trying to use the Web-Hooks with Jenkins-CI. Jenkins has a plugin to do
> this, see the source:
> https://github.com/jenkinsci/gitorious-plugin/blob/master/src/main/java/org/jenkinsci/plugins/gitorious/GitoriousStatus.java
>
> After not b
On Thu, Jan 2, 2014 at 5:47 AM, Paweł Pierzchała wrote:
> No worries, I've fixed the issue with tests, I merged all the other changes
> as well. I've released new versions of libdolt and makeup gems.
Cool, I was struggling with knowing what to do about the m[:suffix]
bit. Your solution works grea
On Thu, Jan 2, 2014 at 3:19 AM, Paweł Pierzchała wrote:
> Hi Ken,
> thank you for the contributions, I'll merge them today.
>
By the way, the github-markup version bump one needs more work before
it is safe to merge. Sorry about that!
https://gitorious.org/gitorious/makeup/merge_requests/2
- Ken
Hi folks,
I have a couple merge requests pending, and it would be great to see
these merged upstream. I've listed them below by ease-of-review.
Gem metadata updates (very trivial):
https://gitorious.org/gitorious/use_case/merge_requests/2
https://gitorious.org/gitorious/makeup/merge_requests/1
A
Hi folks,
I wanted to clarify how builds work in Jenkins, at
https://ci.gitorious.org/jenkins/
Each of the jobs are set up as "multi-configuration" projects. I set
things up this way so that we could test the builds on multiple
platforms.
For example, we started out on CentOS 6 with the standard
Looks like we'll want to fix it in the source code for "ui3".
https://gitorious.org/gitorious/ui3/merge_requests/2
On Mon, Dec 2, 2013 at 8:25 AM, Jonathan Wakely wrote:
> I probably should have just fixed it rather than posting it here.
>
> Merge request now at
> https://gitorious.org/gitorious
On Fri, Nov 29, 2013 at 10:03 AM, Alexandre Beauclair
wrote:
> I'm pretty much at a loss as to what I should do, as my configs are
> similar to the working install (adapted for the test install of course).
> Thanks for any input you can give me!
Here's the code that's failing, in
lib/gitorious/au
On Wed, Nov 27, 2013 at 10:09 AM, Ken Dreyer wrote:
> By the way, looking at your build failure a bit more, it looks like
> the root problem is that OpenSSL in RHEL 6 does not support elliptic
> curves, and chruby unconditionally wants to build against an OpenSSL
> with EC enab
Hi David,
You mentioned in the subject: "how to use existing ruby"
I really wish that Gitorious was going to work on RHEL 6's standard
Ruby. The tricky part is that RHEL 6 ships with Ruby 1.8.7, and there
have been a massive amount of changes between that version of Ruby and
something more modern
Hi folks,
ui3's .gitmodules file references "samsam" on GitHub. This stuck out
to me because the Git URL for the submodule happens to an SSH URL, so
I thought we should change it to a HTTPS URL for consistency.
However, I'm now thinking that samsam might have been added to
.gitmodules inadvertent
rious 4, requires 1.9+ and
> recommends ruby 2.0.
Hi Marcin,
I'm at the place where I recognize that 1.8.7 compatibility is
entirely a lost cause, so these merge requests were simply to allow us
to transition Jenkins over to Ruby 1.9.3. That's what I meant in my
earlier email:
O
Hi folks,
I'm trying to figure out if we can drop the binding_of_caller
dependency from Gitorious.
binding_of_caller entered our Gemfile with this commit:
commit 42de02566fdb8c6aa0616305b549b185ec0e9e0d
Author: Marius Mathiesen
Date: Tue Jul 2 13:04:45 2013 +0200
BetterErrors
The co
Hi folks,
I was looking over the dependency chain for mainline, and I see that
the proxymachine gem underpins the bin/git-proxy script. A couple
questions:
1) In general, should Gitorious administrators be using bin/git-proxy
instead of script/git-daemon?
2) The proxymachine gem has not had any
Hi folks,
Gitorious depends on ruby-yadis, but this gem has not been updated on
rubygems.org in five years. The links in the documentation (eg
yadis.org, or http://www.openidenabled.com/yadis/yadis-notes/) are
broken or redirects. I'm thinking this project may be dead upstream.
Looking through th
On Sat, Oct 5, 2013 at 4:24 AM, Ken Dreyer wrote:
> Since Gitorious mainline itself is now more heavily tied to Ruby 1.9
> as a minimum dependency, I've given up on maintaining Ruby 1.8 support
> in capillary_rb and libdolt as well. I've removed the Ruby 1.8
> buildslav
On Tue, Oct 8, 2013 at 12:14 AM, Christian Johansen wrote:
>> https://gitorious.org/capillary/capillary_rb/merge_requests/8
>
> Thanks Ken!
>
> Merged. If you find more stuff like this (unnecessary gems) feel free to
> let me know.
Awesome! Thank you.
- Ken
--
--
To post to this group, send e
Hi folks,
In packaging the capillary Gem for Fedora, I'm running the test suite
during the RPM's %check function. The capillary test suite requires
the mini_shoulda gem, so I would need to package this gem for Fedora.
The problem is that mini_shoulda is deprecated, so there's not much
point in pac
With the recent UTF-8 fixes in Gitorious, there have been a couple
patches to capillary_rb and libdolt that have brought in newer String
functions like encode() or force_encoding() unconditionally. Since
these functions are not available on Ruby 1.8, this means that these
gems now only run on Ruby
On Wed, Aug 21, 2013 at 1:25 AM, Marius Mårnes Mathiesen
wrote:
> Ken Dreyer writes:
>
>> Hi folks,
>>
>> I finally added the Ruby 1.9.3 buildslave to the gitorious project in
>> Jenkins. The test suite is currently failing because activesupport is
>> trying
Hi folks,
In aa360a8, Marcin mentioned a bug in ActsAsTaggableOn:
commit aa360a8ed80d3d7e946c0b02a6e08be546ab25bb
Author: Marcin Kulik
Date: Wed Sep 18 09:47:11 2013 +0200
Disable remove_unused_tags of ActsAsTaggableOn gem
This feature of the gem has a bug and there is no upstream f
Hi folks,
I wonder if someone with admin access to the Ohloh Gitorious settings
could alter it to point at the "next" branch? Currently it's pointing
at "master", which doesn't get much activity. I think "next" would be
much more representative of the work that is happening in the project.
https:
Hi folks,
I finally added the Ruby 1.9.3 buildslave to the gitorious project in
Jenkins. The test suite is currently failing because activesupport is
trying to search for "tmp/cache", and that directory is not present in
a clean Git checkout. I.e:
Errno::ENOENT: No such file or directory - /pa
On Mon, Aug 5, 2013 at 4:59 PM, Ken Dreyer wrote:
> Hi folks,
>
> Unfortunately the recent ThinkingSphinx bump in
> 0e162d083d4f005355b1320b2dad91688cb0c615 pushed us to a version of
> ThinkingSphinx that does not work on Ruby 1.8.7.
By the way, in order to avoid an being the co
Hi folks,
Unfortunately the recent ThinkingSphinx bump in
0e162d083d4f005355b1320b2dad91688cb0c615 pushed us to a version of
ThinkingSphinx that does not work on Ruby 1.8.7.
I looked a bit into what it would take to fix it, and the short
version is that it was over my head :)
ThinkingSphinx has
Might I suggest a different IRC /topic for the #gitorious channel on
Freenode? I think some slight tweaks would eliminate repeat questions
in IRC, and it might encourage users to get more involved. In
particular I noticed bugs get reported in IRC during off-hours.
Current #gitorious topic:
http:/
On Fri, Jun 28, 2013 at 8:10 AM, Christian Johansen wrote:
>>> There is also a problem related to the way that execute() gets called
>>> on anonymous classes on Ruby 1.8.7. I tried to wrap my mind around the
>>> use_case gem so I could understand how to fix it, but it is beyond me.
>>> I've filed
have time. The new Jenkins setup
makes it easier for me to watch for this stuff too.
Thanks for considering this!
- Ken
[1] https://gitorious.org/gitorious/mainline/merge_requests/239
[2] http://people.redhat.com/bkabrda/ruby193-rhel-6/
On Wed, Jul 3, 2013 at 1:08 AM, Marius Mårnes Mathiesen
Hi folks,
It looks like the "better_errors" gem that went into the next branch
is incompatible with Ruby 1.8.7:
https://ci.gitorious.org/jenkins/job/mainline/13/label=centos-6-x86_64/console
The specific commit was
42de02566fdb8c6aa0616305b549b185ec0e9e0d
Is there any way to use a 1.8.7-compati
I recently investigated some test suite failures on the "next" branch.
For Fedora 19 (Ruby 2.0.0), there's a really simple adjustment that is
needed. The details are at
https://gitorious.org/gitorious/mainline/merge_requests/235
For CentOS 6 (Ruby 1.8.7), the failures are more complex. The
"refut
Did you restart the appropriate services after changing the
configuration files that Marius pointed out? You'll likely need to
restart unicorn and nginx, at a minimum.
- Ken
On Wed, Jun 26, 2013 at 3:06 PM, wrote:
> How do I disable SSL with this VM?
>
> I tried to disable it in gitorious.yum,
Hi Christian,
On Wed, May 15, 2013 at 11:44 PM, Christian Johansen
wrote:
> Well, the old test simply stubs the method being used. In your updated
> example that's not going to turn out pretty. I suggest giving FakeWeb a
> look: https://github.com/chrisk/fakeweb
>
> Let me know if you need some h
Hi folks,
I've added preliminary support for HTTPS web hooks:
https://gitorious.org/~ktdreyer/gitorious/ktdreyers-mainline/commits/https-webhooks
I'm a newbie to Ruby testing so please be kind :-) ... I'm not sure
how to update the unit tests to match this change. See my "FIXME"
lines in test/un
Hi Frank,
Check out the "custom_username_label" in gitorious.yml. You can set
the login label string to whatever you wish.
- Ken
On Tue, May 7, 2013 at 6:34 AM, wrote:
> Hi, I am running Gitorious inside our company using LDAP Authentication.
>
> Gitorious is running fine, but there is one th
Hi Steven,
I think this is the way it's supposed to work. In Git terms, when I do
a "push", I'm not implying that I am the author of any commits that
I'm pushing. I may have gotten all the commits from somewhere else
entirely.
For example, when I clone the gitorious mainline repository into my
ow
On Tue, Mar 19, 2013 at 1:57 AM, Christian Johansen wrote:
> Being able to run multiple Gitorious instances on a box would be a
> nice-to-have, but in my opinion not critical, and probably not well
> suited for production use unless each instance has relatively few
> users/projects or sees sparse
Hi Bobbie,
You may be able to get more information out of the production.log
file. Here's some things to check:
On Mon, Mar 11, 2013 at 2:37 PM, Bobby wrote:
> When trying to authenticate through the web interface, I have tried the
> following options below and keep getting the error message "Em
primarily for splitting out the "git" user,
so that I can have a unique "git" UID inside each chroot, with its own
home dir, SSH keys, etc. I don't know of any other way to do that,
short of full OS virtualization.
- Ken
> On Mar 18, 2013, at 12:24, Ken Dreyer wr
On Mon, Mar 18, 2013 at 12:46 PM, Johannes Ernst wrote:
> It may not be so easy to run more than one installation of Gitorious on the
> same server.
I've recently been looking into this exact problem, after I found out
to my disappointment that my server hardware doesn't support
virtualization.
Hi Joshua,
It looks like some email to the list got delayed, and I think we might
have talked about this in IRC, but I'll reply here too for the record.
>From your pastebin I see you have this line:
KrbMethodK5Passwd on
mod_auth_kerb has two modes of operation: Negotiate authentication and
Bas
I recently hit a snag in my packaging efforts for Fedora. The problem
is that Gitorious wants to run its test suite against a live MySQL
server. Unfortunately there's no reliable way to start up and shut
down mysqld inside the build chroots that Fedora uses.
Here's a mailing list thread where I ra
On Mon, Feb 11, 2013 at 12:58 PM, Andrew Kerr wrote:
> I do have to wonder if it might be better to just roll up an RPM and be done
> with it. Gitorious is already structured in a way that would lend its self
> well to this, I believe. Or skip the dependencies and just have the a bunch
> of good
On Mon, Feb 4, 2013 at 6:23 AM, Marius Mathiesen
wrote:
> Any other things we should consider? Other thoughts?
Hi Marius,
I think this is a great move for the project.
What do you think about time frame commitments? When I ran into a
security issue a while back[1], I wasn't sure how long it wou
On Thu, Jan 31, 2013 at 3:06 AM, Christian Johansen wrote:
> As it happens, I am one of two principal Buster.JS developers.
Yes, thank you for your work on this!
> I have now replaced this module with "bane", which provides the
> functionality needed for Capillary (the JavaScript git graph modul
I'm slowly working on the RPM packaging for Gitorious and its
dependencies. I had a question about the buster git submodule.
(Apologies if this is a newbie question. I'm trying to learn the
history of buster.js as I go along.)
The Gitorious project pulls from
https://gitorious.org/buster/buster-co
On Thu, Jan 24, 2013 at 6:35 AM, anapsix wrote:
> Would a pre-receive hook get file sizes from client when push is
> initiated? If so, calculating available space (quota allowance) and
> rejecting push based on that would be trivial..
Yes, I was wondering the same thing. On the other hand, it lo
On Thu, Jan 24, 2013 at 4:14 AM, Marius Mårnes Mathiesen
wrote:
> Another approach could be to use eg. btrfs and set up subvolumes for new
> projects in the repository root, assuming you're running with un-sharded
> paths. Let's say a user creates a project "gitorious" and the first
> repository "
Hi folks,
I just had a developer fill up my Gitorious partition with a very
large git repo. The problem was compounded by the fact that I
neglected to put /var/gitorious on its own partition... whoops. I
remedied that in hindsight :)
I'm trying to brainstorm the best way to prevent one repo or pr
On Fri, Jan 18, 2013 at 4:38 PM, wrote:
> /home/git/gitorious/lib/tasks/../gitorious.rb
> module Gitorious
> <<< HEAD
> VERSION = "2.3.2"
> ===
> VERSION = "2.4.1"
v2.4.1
> end
>
The <<< and >>> characters makes me wonder if you're installation is
stuck in the middle of a me
On Mon, Jan 14, 2013 at 3:24 AM, Jaco Theron wrote:
> Hi Marius,
>
> I installed sphinx RPM like this:
>
> wget
> ftp://mirror.as24220.net/pub/epel/6/x86_64/sphinx-0.9.9-1.el6.x86_64.rpm
> rpm -Uvh sphinx-0.9.9-1.el6.x86_64.rpm
>
Hi Jaco,
You should also be sure to install the "epel-release"
I noticed that even when is_gitorious_dot_org is false, there is still
a link to "/gitorious" here:
app/views/site/index.html.erb
<% if GitoriousConfig['public_mode'] || logged_in? -%>
<% pull_box(nil, :class => "expose") do -%>
Gitorious provides open source infrastructure
fo
On Thu, Jan 10, 2013 at 11:34 AM, Jaco Theron wrote:
> Encountered a problem with installing sphinx. (CentOS 6)
>
> sudo yum -y install sphinx
>
> The above command does not find sphinx.
Does the sphinx package in EPEL work for you?
- Ken
--
To post to this group, send email to gitorious@googl
On Tue, Dec 11, 2012 at 1:32 AM, Marius Mårnes Mathiesen
wrote:
> However, there's an if clause around that call: `unless
> committerships.any?{|c|c.committer == another_owner}` which means that
> if the new owner already has access to the repository, a new
> committership is not created. Could th
Context: I've got an intranet Gitorious instance set up, with a couple
self-managed teams.
I just had some users lose access to some repos when they transferred
the repos to a team. Even though these users were members of the group
which was the "Owner", the users still had no admin rights to the
On Thu, Dec 6, 2012 at 2:01 PM, Walter Boring wrote:
> Well, I resorted to the only option I could at this point and hacked the
> code myself.
Can you let us know what version of the code you are running? It
sounds like this could be a bug in that version.
- Ken
--
To post to this group, send
On Mon, Nov 19, 2012 at 11:19 AM, Walter Boring wrote:
> This is the user record just prior to calling user.save
> # "walter.bor...@hp.com", crypted_password: nil, salt: nil, created_at: nil,
Unfortunately "@" signs are not currently supported in usernames. You
should set "hp.com" within the dist
On Wed, Nov 14, 2012 at 2:34 PM, Semyazz wrote:
> I'd like to have Gitorious --- repository server and Jenkins --- build
> server.
You might want to read this thread from the Jenkins Gitorious plugin author [1]
> Also I'd like to have some kind of intermediate state where experienced
> developer
On Mon, Oct 15, 2012 at 5:35 AM, Marius Mårnes Mathiesen
wrote:
> What do you think?
Sounds straightforward to me. Thanks for outlining this.
My one suggestion is to make this policy crystal clear to new
developers. Whenever I write a patch for a project I want to know
"what branch do I base thi
On Tue, Oct 2, 2012 at 12:21 PM, Chris Holden wrote:
> Thomas,
>
> THank you, I will do the work since it's my need, just wanted to confirm I
> was on the right path. :)
Looks like some of the $operatingsystem conditionals could use
$osfamily instead, and work across the whole RHEL family.
- Ken
On Mon, Sep 24, 2012 at 3:37 AM, Marius Mårnes Mathiesen
wrote:
> Although I wasn't around at the time, I would think it either had to do with
> a higher probabilty for uniqueness with a three char username or the risk of
> brute force attacks on shorter usernames?
Thank you. Do you think this is
On Fri, Sep 21, 2012 at 3:34 PM, Jacob Blinn wrote:
> We are getting this error whenever we navigate to the HTTPS side to login.
> Nothing on the server has actually changed so I can't tell why suddenly we
> get this error. It seems to only happen in Chrome, but Firefox will load and
> allow login
On Fri, Sep 21, 2012 at 6:52 AM, Peter Kjellerstedt
wrote:
> Couldn’t you just change the validation in app/models/user.rb that validates
> the length of the login to be between 3 and 40 characters to allow 2 to 40
> instead? That is what we have done here…
My concern with this approach is that i
On Tue, Sep 11, 2012 at 12:01 PM, Federico Don wrote:
> I can see on http://gitorious.com/feature_list "code review"
>
> what is the meaning of this point?
>
> Merge requests and code review
>
> The Gitorious web interface aids code review and git workflow.
>
> Where do I find this option? is It
On Tue, Sep 11, 2012 at 11:48 AM, Federico Don wrote:
> Hello,
>
> Gitorious Private has integrated code review? Where do I find this option?
I think that means "users can comment on commits and merge requests"?
- Ken
--
To post to this group, send email to gitorious@googlegroups.com
To unsubs
On Wed, Aug 15, 2012 at 9:18 AM, Dominick Rivard
wrote:
> Thanks for pointing out the instructions. I followed them, the only thing I
> added to fix was a conflict in the gitorious.sample.yml.
Ok, sounds like the author needs to rebase their merge request onto
the latest version of the code in ma
On Thu, Aug 16, 2012 at 7:04 AM, wrote:
> and then Marius Mårnes Mathiesen
> suggests to add github's API to gitorious.
I can speak to this one, since it was my original suggestion and
Marius was answering - I was just tossing out an idea for API
construction in the future version of Gitoriou
Hello,
I'd like to get some feedback on a configuration idea.
The gitorious.yml file contains a "cookie_secret" session key. Other
than this one parameter, gitorious.yml doesn't contain any
cryptographic material. I think it would be advantageous to move this
one parameter out into its own file,
On Fri, Aug 10, 2012 at 11:57 AM, Dominick Rivard
wrote:
> I would be happy to help out. I just need to find out how to get that
> specific patch from git.
Sure, visit this link:
https://gitorious.org/gitorious/mainline/merge_requests/2247
Then click the "How to apply this merge request to your
Capillary devs,
I see capillary 1.0.1 has been released. Would you mind pushing the
tag to the capillary repo on Gitorious?
(If there is a better avenue to request this sort of thing, please let me know.)
- Ken
--
To post to this group, send email to gitorious@googlegroups.com
To unsubscribe f
On Thu, Aug 9, 2012 at 9:45 PM, Dominick Rivard
wrote:
> Is there a way to disable it through a
> configuration?
The functionality does not yet exist in mainline gitorious, but there
is a patch in the merge request queue to do this [1]. If the patch
works for you, you could offer some positive fe
On Fri, Aug 3, 2012 at 3:15 AM, Marius Mårnes Mathiesen
wrote:
> While we're at the packaging topic: I saw some posts on a Fedora mailing
> list that you're working on packaging Gitorious for Fedora, that would be
> really great! Any news? Need help?
Nice, thanks for asking. I've broken out all t
On Wed, Aug 1, 2012 at 4:17 AM, dpwrussell wrote:
> I see that there is the concept of versioning for some time now in
> Gitorious, but all the installation instructions I have seen involve a
> checkout from mainline (If versioning is active I'd have expected a checkout
> of a certain version bran
On Fri, Jul 27, 2012 at 2:50 PM, Federico Don wrote:
> I need the opinion of the group, you must implement code reviews with gerrit
> in gitorious? is this really necessary?
>
> I hope your opinion!
My preference would be to simply use merge requests for this purpose
because it is simpler. Gitori
Would there be any problems with copying / reimplementing GitHub's API
in Gitorious?
There are lots of client libraries for various languages that use
GitHub's V3 API, and it would be neat if we could "drop in" Gitorious
as the backend.
- Ken
--
To post to this group, send email to gitorious@go
Sebastian Noack and I were discussing the Gitorious::Authentication
plugins, and Sebastian brought up an interesting point.
Sebastian noticed that the Kerberos plugin only replaces dots ('.') in
usernames, but there are more invalid characters. In his SSL
Authentication plugin, he uses a regex to
Federico Don
>>>
>>> Hello,
>>>
>>> I have this configuration, thanks to this update could authenticate with
>>> the test_ldap_connection .but I can not authenticate via web
>>>
>>> I need help!!
>>>
On Mon, Jun 11, 2012 at 9:05 AM, Federico Don wrote:
> bind_username: USER_BIND
> bind_password: PASSWORD_BIND
As I understand it, the problem is that
https://gitorious.org/gitorious/mainline/merge_requests/181 has not
been updated with the changes that happened in commit 570f36e.
(B
On Mon, Jun 11, 2012 at 8:41 AM, Federico Don wrote:
> A NoMethodError occurred in sessions#create:
>
> undefined method `username_filter' for
> #
> [RAILS_ROOT]/lib/gitorious/authentication/ldap_authentication.rb:85:in
> `authenticate'
The username_filter function was added a month ago to mast
On Tue, Jun 5, 2012 at 6:23 AM, Thomas Chemineau
wrote:
> I ask myself if Gitorious was able to authenticate user by using HTTP header
> to find username. Could it be possible ?
Hi Thomas,
I've submitted a plugin for Kerberos authentication, which is very
similar to what you describe. You could
On Thu, May 24, 2012 at 7:13 AM, Ken Dreyer wrote:
> On Thu, May 24, 2012 at 3:57 AM, Marius Mårnes Mathiesen
> wrote:
>> That's an elegant solution. Would you mind sending a merge request for that?
>
> Sure thing, submitted at
> https://gitorious.org/gitorious/m
On Thu, May 24, 2012 at 3:57 AM, Marius Mårnes Mathiesen
wrote:
> That's an elegant solution. Would you mind sending a merge request for that?
Sure thing, submitted at
https://gitorious.org/gitorious/mainline/merge_requests/205
- Ken
--
To post to this group, send email to gitorious@googlegrou
On Thu, May 3, 2012 at 11:26 AM, Marius Mårnes Mathiesen
wrote:
> On Wed, May 2, 2012 at 5:33 AM, Ken Dreyer wrote:
> I think this looks really good, and I'd love it if you'd submit a merge
> request for it.
>
> Thanks - this is really appreciated!
Thanks for the enco
On Thu, May 17, 2012 at 9:46 AM, Ken Dreyer wrote:
> I think it would be better to register the LDAP or Kerberos users with
> cryptographically random passwords. What form do you recommend?
Digging a bit more, it looks like the Crowd plugin suffers from the
same "default password in t
I was doing some more tidying up for my KerberosAuthentication module,
and I have a question about user account creation. Specifically, what
is the best policy for storing a "dummy" password for a user?
The LDAPAuthentication class will set a default static password of
"left_blank" in auto_registe
On Mon, Apr 23, 2012 at 9:20 AM, Ken Dreyer wrote:
> Both patches are attached, rebased against current master (430eb2bda).
> Please let me know if these look ok :)
Gitorious devs,
If my approach here is ok, please let me know and I'll submit a merge
request to get this into mainl
Hello,
I'm wondering if ruby-hmac should be removed from Gitorious's Gemfile
and Gemfile.lock? I don't see any code that depends on ruby-hmac, and
in addition the upstream author recommends using the OpenSSL
functionality instead.
- Ken
--
To post to this group, send email to gitorious@googlegr
On Tue, Apr 17, 2012 at 2:32 AM, Marius Mårnes Mathiesen
wrote:
> Great work! I know next to nothing about Kerberos, but it's been on my todo
> list for a while to investigate how to support it in Gitorious.
Thanks! I appreciate your review and your feedback.
I implemented your suggestion for mo
On Wed, Apr 11, 2012 at 8:10 PM, Ken Dreyer wrote:
> I recently hacked Kerberos authentication onto the Gitorious web interface.
I ended up implementing this in a less hackish way, with a proper
Gitorious::Authentication::KerberosAuthentication module. The core
code changes are not as daunt
I recently hacked Kerberos authentication onto the Gitorious web interface.
In conjunction with Apache's mod_auth_kerb, the Rails app pulls the
username out of request.env['REMOTE_USER'] and uses that in a manner
similar to the current code's OpenID handling: If the user doesn't yet
exist in th
94 matches
Mail list logo