Basically, if you could point me at the problem I'd be willing to work on it. Are you (the Roxygen team) satisfied with your basic approach, and then adding support for setRefClass is just a logical extension of the existing parser? Otherwise, what are the next steps? <div>
<br></div><div>Thanks again!<br clear="all"><div>--</div><font color="#666666">Robert Tirrell | </font><span style="color:rgb(102, 102, 102)"><a href="mailto:rpt@stanford.edu" target="_blank">rpt@stanford.edu</a> | (607) 437-6532</span><div>
<div><font color="#CCCCCC"><font color="#333333"><font color="#666666">Program in Biomedical Informatics | </font></font></font><font color="#CCCCCC"><font color="#333333"><font color="#666666">Butte Lab | </font></font></font><span style="color:rgb(102, 102, 102)">Stanford University</span></div>
</div><br>
<br><br><div class="gmail_quote">On Thu, Mar 17, 2011 at 12:55, Rob Tirrell <span dir="ltr"><<a href="mailto:rpt@stanford.edu">rpt@stanford.edu</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div dir="ltr">Yeah, I think that would be the goal - to descend into each class and read tags, as if they were naked functions plus some context derived from the enclosing call to setRefClass. I don't know how this would work with the inline documentation reference classes allow. <div>
As far as that commit message, I don't exactly recall what prompted it. I initially had some misundertsandings about how Rd2 works.<br clear="all"><div>-- </div><div class="im"><font color="#666666">Robert Tirrell | </font><span style="color:rgb(102, 102, 102)"><a href="mailto:rpt@stanford.edu" target="_blank">rpt@stanford.edu</a> | <a href="tel:%28607%29%20437-6532" value="+16074376532" target="_blank">(607) 437-6532</a></span><div>
<div><font color="#CCCCCC"><font color="#333333"><font color="#666666">Program in Biomedical Informatics | </font></font></font><font color="#CCCCCC"><font color="#333333"><font color="#666666">Butte Lab | </font></font></font><span style="color:rgb(102, 102, 102)">Stanford University</span></div>
</div><br>
<br><br></div><div><div></div><div class="h5"><div class="gmail_quote">On Thu, Mar 17, 2011 at 11:54, Peter Danenberg <span dir="ltr"><<a href="mailto:pcd@roxygen.org" target="_blank">pcd@roxygen.org</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Quoth Rob Tirrell on Sweetmorn, the 3rd of Discord:<br>
<div>> As a quick followup: I've been working on an database abstraction<br>
> package (<a href="https://github.com/rtirrell/databasr" target="_blank">https://github.com/rtirrell/databasr</a>) that makes extensive<br>
> use of reference classes.<br>
<br>
</div>Thanks, Rob; I now have a better intuition of how refclasses might<br>
work.<br>
<br>
Using JoinClause as an example, you'd like Roxygen to descend into the<br>
class definition; documenting e.g. the individual methods?<br>
<br>
#' Represents a JOIN clause.<br>
JoinClause <- setRefClass('JoinClause',<br>
<br>
[...]<br>
<br>
methods = list(<br>
<br>
[...]<br>
<br>
#' Add children to this JOIN clause.<br>
#'<br>
#' This sets the type based on the last element we add.<br>
#' @param ... arguments to add as children.<br>
#' @return \code{.self}<br>
add_children = function(...) {<br>
<br>
[...]<br>
<br>
}<br>
<br>
[...]))<br>
<br>
Also, one of your last commit-messages was "damn roxygen;" do you<br>
remember what the specific frustration was?<br>
</blockquote></div><br></div></div></div></div>
</blockquote></div><br></div>