We heard you loud and clear that you wanted to integrate GitHub with your existing services without having to setup a custom post-receive service. Now, when you go to your repository’s edit screen, you’ll see the following:
We knocked out the first two services that we’re most familiar with, Lighthouse and Campfire, but expect to see more coming very soon. Feel free to request a service (or +1 the ticket if your service is already listed) so we know which are the most help to you guys.
I assume the Lighthouse hook is used to close a ticket. Is there a specific format that that commit message needs to follow?
Okay, I set this up and don’t have a clue as to what this does… ;-)
yes, yes, yes! i was just mumbling to myself how much i dislike the post-commit solution! you guys rock!
Lighthouse API Basically, use is like this in commit messages-
Prototype new feature [#15]
New feature is implemented and tested [#15 tagged:committed responsible:rick milestone:“Launch” state:resolved]
@robby when you push to github, we pipe all of the commit info info to any of the services you setup. If you use the special syntax that jlindley linked to, Lighthouse will do cool things with it.
So, this will send a message to campfire when you commit?
@b3b0p yes, but if you’re on an ssl account, you’ll need to wait a bit longer until we push out support for it.
Woo hoo. I see the messages on my campfire. But my Lighthouse ain’t getting it yet. For “project id” I’m using Lighthouse’s “token name” per these instructions: http://www.lighthouseapp.com/help/using-beacons-and-api-tokens
Is that right? Or should I be using part of the email identifier used to email tickets into Lighthouse e.g. ticket+name.99-abc32xyz@lighthouseapp.com
for the record, your account(subdomain) is the first part of your lighthouse URL:
http://jtoe.lighthouseapp.com
Every commit says it’s me, even when it’s collaborators on my github project…fine for now, but def something I want to see fixed/updated for sure
Very cool feature. I know its new, but wanted to add a +1 for more customization of the integration level – meaning some control of how ‘chatty’ it is. For example, posting all the changesets to lighthouse for me is total overkill and actually a deterent – thats what github is for, lighthouse is for tickets, not repo tracking. Clients (for example) don’t need to see commits, and its better that they don’t.
Anyhow, great feature, glad to see github going that direction!
I couldn’t figure this out very well at first until I found this page: http://lighthouseapp.com/help/setting-up-subversion-integration
Also – this helps: http://hoth.entp.com/2008/4/11/github-and-lighthouse-sitting-in-a-tree
Be aware… Service integration is now on the SERVICES tab. (I couldn’t find it at first. :P) Mitch
I don’t understand how this is realistically usable if you can’t distinguish between collaborates on a project… As @imagetic points out, any commits appear to come from the Lighthouse project administrator. Guess I will have to resort back to manual commit hooks.
Some of the converted reports may not have report links since these are explicitly removed during the report conversion process. The reason these links are removed traces back to a fundamental way in which SQL Reporting Services handles report links search engine optimization.
An example would make this issue clearer. Let us say Report A links to Report B via parameters Param1 and Param2. Also let us say that Param1 and Param2 which are parameters for Report B2 take a valid set of values (basically these parameters offer a list of possible values for the users to select by running a SQL query). Quite often it so happens that Report A has blank values for these parameters in some of the rows and when you drill down from these rows to Report B, SQL Reporting Services throws an exception that the values for these parameters are not valid. This is because it matches the value passed from Report A to the list of plausible values by running the SQL Query behind that report parameter for Report B and finds that a blank value is not in the list of valid values. The legacy ASP-based reporting point handled this gracefully and would allow passing blank/null parameter values to the linked report on internet phone service. But SQL Reporting Services enforces a tighter restriction on what gets passed into the linked report to avoid any possibilities of SQL injection attacks.
A good example where this issue manifests is with reports that deal with status messages. Internally many of the reports define status messages to be a list of valid values (like 0=Information, 1=Warning, 2=Error etc.). If you look at any parent report that provides a high level view of say something like patch deployment status you may see that this report links to a drill-down for a machine specific report for patch deployment. Now it is quite likely that at some point there is no status message available for a particular machine (may be because the patch deployment was delayed for some reason) and that row shows up as blank in the parent report but if you drill down this report SQL Reporting Services will throw an exception that the blank value for status message ID is not acceptable.
Workaround: Run the linked report directly by passing the appropriate set of report parameter values.
I know its new, but wanted to add a +1 for more customization of the integration level – meaning some control of how ‘chatty’ it is. For example, posting all the changesets to lighthouse for me is total overkill and actually a deterent.
Furnished apartments Dubai || villas in Dubai || buy property Abu Dhabi || Broker Management
Love the post, im off to buy the best toaster now :-)