Bugs 17918, 18143, 18145 - debugcall() issues

[bastistician]

https://bugs.r-project.org/show_bug.cgi?id=17918

• Maybe we can get func simply by evaluating call[[1L]] in parent.frame(2L)?
• The current patch proposal also adds an environment arg to the methods::isGeneric(funsym) call, which actually seems to be one option to fix Bug 18143 (another one is to use the fdef argument of isGeneric()).
• Bug 18145 highlights a problem with func <- getS3method(genname, class(arg)), which does not reliably find the S3 method that will eventually be called. Furthermore, class(arg) is bad as getS3method() expects a single class name.

These related issues are best worked on in one go.

[gmbecker]

As I'm responsible for that I'm happy to collaborate on a fix in either direction (i.e., with me working in collab with R-core, or me mentoring/overseeing work by others in prep for submission to R-core).

I will be on hand for Sessions 1 and 3

[gmbecker]

I have started working on fixes to at least some of these issues and will continue

[gmbecker]

I have filed (one part) of the isGeneric bug we unearthed during our exploration of this here: https://bugs.r-project.org/show_bug.cgi?id=18369

[gmbecker]

I've filed the narrower, separable warning bug here: https://bugs.r-project.org/show_bug.cgi?id=18370 with a patch

[bastistician]

@gmbecker: Thank you for your inputs on these issues. I'll have a look later this week.

[gmbecker]

I've submitted a patch for https://bugs.r-project.org/show_bug.cgi?id=18145 which also fixes https://bugs.r-project.org/show_bug.cgi?id=18143 in the process.