[racket] Fix for dherman/json

Robby Findler wrote at 05/02/2011 06:19 AM:
> Are there bug reporting systems that you or Neil (or anyone, really) like?
>   

I've used a lot of issue-tracking systems, but don't recall a 
particularly likable one.  They range from passable to infuriating.

The main annoyance with the PLaneT Trac thing is that it's an 
*additional* thing I need to deal with.  I was already getting bug 
reports and enhancement requests emailed to me directly, in addition to 
the issues that I identified myself... and now I have to keep track of 
an third place where issues are noted, and then hassle with its 
interface when handling those issues.  That's making the process 
heavier-weight for me, not lighter-weight.

The mixed feelings I'm having are that, as a package *author*, I don't 
want to have to use Trac, *but*, in an open source environment like 
PLaneT, as a package *user*, I can see value in getting known issues out 
into the open.

Despite feelings mixed, I'm leaning towards thinking that package 
authors/maintainers to be able to disable PLaneT Trac for their 
packages, on a per-package basis, with the understanding that the 
author/maintainer is then supposed to provide alternative means for 
reporting bugs.  The default would be for Trac to be enabled.

If people want to know the known issues in my packages, most of the 
issues are noted in the source code or documentation.  That's not 
optimal for them, compared to (in theory) being able to go to a 
well-known Web site and see all the issues, but I'll be somewhat less 
slow to respond to emailed bug reports.

One more comment: I estimate that I get a few times more issues from 
Racket users through email than through PLaneT Trac.  I know a lot of 
people don't like going to a Web site and filling out bug submission 
forms, and perhaps that's why I get more email than Trac.  This also 
means that known issues about my packages generally aren't in Trac anyway.

-- 
http://www.neilvandyke.org/