[PATCH] D55257: Inline handling of DependentSizedArrayType

[Aaron Ballman via Phabricator via cfe-commits]

aaron.ballman added a comment.

In D55257#1318376 <https://reviews.llvm.org/D55257#1318376>, @steveire wrote:

> In D55257#1318328 <https://reviews.llvm.org/D55257#1318328>, @aaron.ballman wrote:
>
> > > It is necessary to perform all printing before any traversal to child nodes.
> >
> > This piqued my interest -- is `VisitFunctionDecl()` then incorrect because it streams output, then dumps parameter children, then dumps more output, then dumps override children? Or do you mean "don't interleave `VisitFoo()` calls with streaming output"?
>
>
> Can you relate your question to https://reviews.llvm.org/D55083 ?


Ah, I was looking at code before having fetched those changes, so perhaps my example is poor. Mostly, I'm wondering what you meant by "traversal to child nodes" -- do you mean:

1. it's bad to output to the stream, then dumpChild(), then output to the stream again
2. it's bad to output to the stream, then VisitFoo(), then output to the stream again
3. both #1 and #2
4. neither #1 nor #2

(as in: when I'm doing a code review a few months from now, what should I be watching out for in this scenario?)


Repository:
  rC Clang

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D55257/new/

https://reviews.llvm.org/D55257