Forum OpenACS Development: Is the OpenACS Patch Tracker (part of Bug Tracker) still being used?

1: Is the OpenACS Patch Tracker (part of Bug Tracker) still being used?

Posted by Andrew Helsley on 12/10/14 11:37 PM

I have a couple of interrelated questions...:

Is the OpenACS Patch Tracker still being used to submit enhancements to OpenACS that seem too small to be TIPs?
... if not, maybe the wording at https://openacs.org/contribute/ should be modified to reflect the fact that patches are for bugs only. Furthermore, if the patch tracker is not used, it begs the question: should I submit suggested enhancements as a "Bug" with the "Bug Type" set to "Suggestion" and the "Severity" set to "6 - Enhancement"?
... if not, should I always just submit a TIP, no matter how simple the enhancement is? (and maybe reference the patch from the TIP if it is a big change?)

For example, I think there should be a proc for resolving a host name into an IP address without throwing an error when the name does not exist. This would be a new proc of 6 lines or so. As such, the TIP process seems to be overkill.

2: Re: Is the OpenACS Patch Tracker (part of Bug Tracker) still being used? (response to 1)

Posted by Gustaf Neumann on 12/11/14 09:31 AM

I would say, the preferred way is (2). Most of the patches are accompanied by issues, and as there are "suggestions" entry types i see no reason for having different channels.

For the suggestion of the host-node-map: sure, it is a useful extension to check the name lookup, at least for the near future. In general the host-node-map has several shortcomings and needs some rethinking. In particular it should be integrated with the virtual server capabilities of nsd, but that is a larger project.

just my 2c

3: Re: Is the OpenACS Patch Tracker (part of Bug Tracker) still being used? (response to 2)

Posted by Andrew Helsley on 12/16/14 11:01 PM

I think it would be even better if the host-node-map was integrated completely into the site-map or at least linked from there. The most notable actual short-comings that I've noticed are the inability to have more than 1 name associated with the root node and the inability to explicitly link the root node to the hostname from nssock and nsopenssl modules.