[gitorious] Re: Confused About Versioning and Upgrading
When I run git tag, I get: 20110110-1 loba-loba v2.0.0 Seems like the only version tag you guys have is v2.0.0 on git://gitorious.org/gitorious/mainline.git So you guys are not following the versioning scheme yet? How can there not be other versions but code changes? Thanks, Drew On Aug 1, 12:20 am, Marius Mårnes Mathiesen wrote: > On Sat, Jul 30, 2011 at 3:54 AM, Drew wrote: > > This is what I usually do. I have a branch called my_config that I > > have pushed my changes in there. Now every now and then I do a git > > pull and get the latest from git://gitorious.org/gitorious/mainline.git. > > Some times gitorious comes up with no complains and some times I have > > to either install new gems or run "rake db:migrate" > > The bundle install or rake migrate steps are two of the reasons why we went > for versioning Gitorious. A patch version should not require: > - rake db:migrate > - bundle install > > Minor versions could require you to do so. So a rule of thumb would be to > upgrade patch versions simply by using git pull, upgrading to a new minor > version you should always both migrate and install gems, since that may be > required. > > > But I did a "git pull" and that downloaded bunch of changes (which > > broke gitorious, but I fixed it by installing new gems) > > I ran "bundle exec rake changelog" again, but the output is the same. > > If you did a git pull while on the master branch, that will have been > updated. Our versioning scheme is based on using git tags to check out a > specific version. > > > So my question is, how does "bundle exec rake changelog" work? If the > > version hasn't changed, then what were all those changes that were > > downloaded? > > It will compare the available tags on the server to the current tag you're > running with. > > Cheers, > - Marius -- To post to this group, send email to gitorious@googlegroups.com To unsubscribe from this group, send email to gitorious+unsubscr...@googlegroups.com
[gitorious] Re: Help for newbie on upgrade
Also, Marius the message I recieve is the following fatal: v2.0 does not point to a commit On Aug 1, 9:37 am, Areva wrote: > Thank you for your response. > > I have created another server with the current gitorious V.2.0 and was > curious is there a mechanism to pull over the repositories and > database from the older version. The upgrade just won't happen. > > I am looking into your suggestion of the capistrano and a mirror of > the server. LOL, being new to some of this makes things interesting > and frustrating. > > Thanks again > > BIll > > On Jul 31, 11:20 pm, Marius Mårnes Mathiesen > > > > > > > > wrote: > > On Fri, Jul 29, 2011 at 7:19 PM, Areva wrote: > > > Apparently I really have no clue on the upgrade. The site will > > > not upgrade using git fetch as stated in the wiki for upgrade, says on > > > the merge does not point to commit. > > > Hmm, this doesn't seem right - which URL are you using and which commit is > > the warning about? > > > > So maintaining the code on the client machine I would like to push the > > > branch changes to the production WebServer, this is where I get all > > > jacked up. Can't seem to wrap my head around how to proceed with this > > > so that any change I make can be pushed up to the production > > > webserver. > > > What we're doing for gitorious.org is roughly: > > - we have a separate git repo/branch on a separate server that contains our > > deployment configuration (a Capistrano recipe, I'll get to that later). > > Whenever we get ready to deploy, we merge mainline's master into this branch > > and push the changes for that branch to our server. Doing something similar > > you could keep your customizations in a separate branch and merge/rebase > > master onto that. > > - we use Capistrano (https://github.com/capistrano/capistrano/wiki) to > > deploy to the server. Unless you have something you know is better, you > > should use Capistrano too. In our configuration, Capistrano will deploy to > > our servers from the separate branch I mentioned above. When you deploy > > using Capistrano, you don't need to worry about removing cached > > javascript/css files; simply point Capistrano to a branch and it will deploy > > from there. > > - if you use the Gitorious version tags > > (https://gitorious.org/gitorious/pages/Versioning), merge the given tag into > > your branch instead of master > > > - Marius -- To post to this group, send email to gitorious@googlegroups.com To unsubscribe from this group, send email to gitorious+unsubscr...@googlegroups.com
[gitorious] Re: Error cloning/pushing
I've found that adding a symbolic link to the gitorious script in /usr/ bin causes a different error message '/usr/bin/env: ruby: No such file or directory' I'm not sure why it wouldn't find the link in /usr/local/bin though. It's added to the default path at startup. On Jul 27, 2:24 pm, Harry wrote: > The error was that I had also installed a gitorious package (arch > linux) which overwrote the 'gitorious' script. I uninstalled the > package, but now I get the error 'bash: gitorious: command not found'. > The script is in the path (I've tried both a symlink in /usr/local/bin > and directly adding the script directory to the path). I found another > post where the poster resolved the issue by setting permissions to > allow the script to be executed, but the permissions seem to be fine > on my system. The gitorious_auth.log file contains > I, [2011-07-27 14:13:29#1381] INFO -- : Connection from "10.1.27.48 > 55324 22" (): > I, [2011-07-27 14:13:29#1381] INFO -- : Need SSH_ORIGINAL_COMMAND > > On Jul 26, 6:59 pm, Robert Kirchgessner > wrote: > > > > > You need to check the gitorious_auth log. I have a feeling this is > > because either your rails app user doesn't have read writes on the > > repository directory or your .hooks are screwed up. You can run the > > fixup-hooks.rb file in the scripts directory to fix that. > > > Need to check the logs tho to really know whats going on. > > > On Jul 26, 12:52 pm, Harry wrote: > > > > Whenever I try to clone from the git repository that I created on a > > > local install of Gitorious or push to it I receive the following > > > error: > > > > == Gitorious: > > > == > > > Temporary error. Please try again shortly > > > > > > > I tried 'ssh git@localhost' and received the error: > > > > PTY allocation request failed on channel 0 > > > > but I assume that's Gitorious preventing shell access. I can't find > > > anything funky in the Gitorious or the ssh logs except > > > gitorious_auth.log, which I've yet to locate. -- To post to this group, send email to gitorious@googlegroups.com To unsubscribe from this group, send email to gitorious+unsubscr...@googlegroups.com
[gitorious] Re: Help for newbie on upgrade
Thank you for your response. I have created another server with the current gitorious V.2.0 and was curious is there a mechanism to pull over the repositories and database from the older version. The upgrade just won't happen. I am looking into your suggestion of the capistrano and a mirror of the server. LOL, being new to some of this makes things interesting and frustrating. Thanks again BIll On Jul 31, 11:20 pm, Marius Mårnes Mathiesen wrote: > On Fri, Jul 29, 2011 at 7:19 PM, Areva wrote: > > Apparently I really have no clue on the upgrade. The site will > > not upgrade using git fetch as stated in the wiki for upgrade, says on > > the merge does not point to commit. > > Hmm, this doesn't seem right - which URL are you using and which commit is > the warning about? > > > So maintaining the code on the client machine I would like to push the > > branch changes to the production WebServer, this is where I get all > > jacked up. Can't seem to wrap my head around how to proceed with this > > so that any change I make can be pushed up to the production > > webserver. > > What we're doing for gitorious.org is roughly: > - we have a separate git repo/branch on a separate server that contains our > deployment configuration (a Capistrano recipe, I'll get to that later). > Whenever we get ready to deploy, we merge mainline's master into this branch > and push the changes for that branch to our server. Doing something similar > you could keep your customizations in a separate branch and merge/rebase > master onto that. > - we use Capistrano (https://github.com/capistrano/capistrano/wiki) to > deploy to the server. Unless you have something you know is better, you > should use Capistrano too. In our configuration, Capistrano will deploy to > our servers from the separate branch I mentioned above. When you deploy > using Capistrano, you don't need to worry about removing cached > javascript/css files; simply point Capistrano to a branch and it will deploy > from there. > - if you use the Gitorious version tags > (https://gitorious.org/gitorious/pages/Versioning), merge the given tag into > your branch instead of master > > - Marius -- To post to this group, send email to gitorious@googlegroups.com To unsubscribe from this group, send email to gitorious+unsubscr...@googlegroups.com
Re: [gitorious] Confused About Versioning and Upgrading
On Sat, Jul 30, 2011 at 3:54 AM, Drew wrote: > This is what I usually do. I have a branch called my_config that I > have pushed my changes in there. Now every now and then I do a git > pull and get the latest from git://gitorious.org/gitorious/mainline.git. > Some times gitorious comes up with no complains and some times I have > to either install new gems or run "rake db:migrate" > The bundle install or rake migrate steps are two of the reasons why we went for versioning Gitorious. A patch version should not require: - rake db:migrate - bundle install Minor versions could require you to do so. So a rule of thumb would be to upgrade patch versions simply by using git pull, upgrading to a new minor version you should always both migrate and install gems, since that may be required. > But I did a "git pull" and that downloaded bunch of changes (which > broke gitorious, but I fixed it by installing new gems) > I ran "bundle exec rake changelog" again, but the output is the same. > If you did a git pull while on the master branch, that will have been updated. Our versioning scheme is based on using git tags to check out a specific version. > So my question is, how does "bundle exec rake changelog" work? If the > version hasn't changed, then what were all those changes that were > downloaded? > It will compare the available tags on the server to the current tag you're running with. Cheers, - Marius -- To post to this group, send email to gitorious@googlegroups.com To unsubscribe from this group, send email to gitorious+unsubscr...@googlegroups.com