[cfe-dev] RFC: Easier AST Matching by Default

[David Rector via cfe-dev]

Good idea — revert to default for now, spend some time reflecting on the ideas.  Unless anyone objects, let’s do that.

I am knee deep in another project right now, but if by the time I finish it in a few months if nobody else has tried this out, I’ll have a look and implement it.  (I don’t in fact actually use ASTMatchers, I only got into this thread to give you some support b/c I admired your dedication to helping beginners — and still do!)  

If somebody else wants to try it in the meantime, go for it!  It seems straightforward but there are definitely some details to consider — e.g. you’d definitely want to adjust TraverseStmt/TraverseType so that, after you delve into the children an implicit node via getAs, you would then skip any implicit nodes you’ve already covered — otherwise you’d get duplicates and/or be doing a lot of needless re-traversal.  (And don’t forget, while you’re under the hood, to fix the DataRecursionQueue Stmt-traversal order problem in raised by Billy way back when...).

Good luck all,

Dave

> On Jun 30, 2020, at 6:18 PM, Stephen Kelly <steveire at gmail.com> wrote:
> 
> 
> On 30/06/2020 22:48, David Rector wrote:
>> Ultimately Stephen, you put in the work, and are entitled to an up or down vote — that’s fair.  Thank you for that work, and good luck to you and others,
> 
> 
> Sorry, there may be misunderstanding here. Rather than bowing out, I think it would be good if you made a proof of concept.
> 
> I think, but I'm not certain, that there is more mood to investigate in the direction you wish to proceed.
> 
> My suggestion is to revert the default to give time for you to flesh out the design.
> 
> Thanks,
> 
> Stephen.
> 
>