Conversation
|
||
files: | ||
|
||
"${g.nhw_crontabs}/${user}" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This should be modified to use paths.crontabs instead of g.nhw_crontabs.
Hi Neil - thanks for the submission! I think the paths bundle is an excellent idea. I have made a couple of comments on some lines. |
I like the paths idea too. Maybe its just me and my brain wanting to limit scope, does anyone else find it easier to read a commit that is just a single sketch? |
That's the idea in general - one sketch per pull request. In this case I think it made sense to bundle them because one uses the other. --Diego On Feb 28, 2012, at 7:39 PM, Nick Andersonreply@reply.github.com wrote:
|
Made changes. |
On 02/28/2012 08:02 PM, Neil H Watson wrote:
Did you push your changes to your fork? I still see the g.nhw_crontabs Nick Anderson nick@cmdln.org |
@@ -0,0 +1,34 @@ | |||
bundle agent cronjobs(user, jobs) { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should this one perhaps be named cronjobs_add since it just appends if not found?
neil what do you think about the bundle name change from cronjobs to cronjobs_add since it appends if not found. I think that would leave logical namespace for something like cronjobs_delete and other patterns for dealing with cronjobs, perhaps cron_d style related. |
I was thinking sketchname of cronjobs like you had but the only bundle you submitted in that sketch was also named cronjobs. I thought maybe just rename that bundle and we can add more cronjob related bundles to the sketch am curious what others think though before you keep making changes on my whims :) |
I agree - I think the sketch should be named |
@@ -0,0 +1,3 @@ | |||
author: Neil H Watson, neil@watson-wilson.ca | |||
ostype: Any UNIX |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Are these metadata values for ostype, tested, and cfengine_version supposed to be valid cfengine classes?
Which ones are required?
Should there be a comma between the name and email address?
I ask because I am thinking ahead to what the not yet finished tooling is going to want to expect and it would be nice to have a tool that could be used to validate a submission
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
In this case, ostype
should be unix
, which is a hard class defined by CFEngine on any Unix-style operating system. The author
should preferably be Name <email@address.com>
.
Indeed the idea is that we'll have a tool that can validate and submit sketches, and also one to install them.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Which if any fields are required?
I have one other large sketch to add. Is it better to wait for this pull to complete and then start a new one? |
Your in the same spot I was, your clone is progressing at a faster pace than upstream. I don't know what the official stance is, but it is probably best to get your stuff on a feature branch. I see one more thing that needs to be addressed in the submission, Then you should update your clone from upstream, and create a feature branch off of it where you can add your new sketch and submit a pull request for that. I just had to work through doing that yesterday so if you need help let me know. |
Hi Neil, The best is to start a new branch for each sketch, then you can submit multiple pull requests simultaneously. --Diego On Feb 29, 2012, at 8:21 PM, Neil H Watson reply@reply.github.com wrote:
|
On 02/29/2012 08:57 PM, Diego Zamboni wrote:
If you branch before your changes are merged and do a pull request on Nick Anderson nick@cmdln.org |
I think these two sketches look good, anyone see any issues with merging? |
@@ -0,0 +1,3 @@ | |||
author: Neil H Watson <neil@watson-wilson.ca> | |||
ostype: Linux, Solaris 10, AIX | |||
tested: Debian 6 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
oops, this should be a valid class name probably debian_6 ?
That's why you never develop on your master branch - that way every branch starts off with the "clean" state of the master branch, without any other changes, and they can be merged independently. On Feb 29, 2012, at 8:58 PM, Nick Anderson wrote:
|
crontabs and common paths bundle.
bundles.cf: rm_rf and rm_rf_depth bundles: add docs and port to 3.5
Remove references to cf-report
Added more content instructing how to write promises
Crontabs is self explanatory. The paths bundle I'd like to be a community
effort to catalog the paths to commonly called commands.