meisl

Boiled-down version: of course I checked :) Space cost: it *is* linear in the input size. But the constant factor can make it impractical: it is the nr of rules...

Added variants 2 and 3 to [the example](https://github.com/meisl/pegjs/blob/master/tools/benchmark/nesting/nesting-profile.pegjs). As mentioned above, I realized that v0 (with `*`) *is not equivalent* to the others, e.g. `x(x)(x)`. That's bad, even though the...

Hi & Thanks, @Mingun Your `group` rule is what I was hinting at above. And your grammar does accept mentioned `x(x)(x)` - with only *one* more rule, not *two* as...

@Mingun I forgot: Could you explain what you meant by > `add = call ("+" call)* // note that use ("+" add) is useless here` ?

Hey @Mingun, big thanks for spotting `v0`/common prefix, that was my mistake: it should NOT have the choice in `add` and `call`. Got too busy with copy/paste, and somehow did...

Re `x(x)(x)`: let me recap, starting with another citation from my https://github.com/pegjs/pegjs/issues/623#issuecomment-523127403 above (emphasis added): > * v0 accepts `x(x)(x)` while the others do not > * BUT: all four...

@Mingun, you are aware that * the actual non-equivalence of v0 and v1 (at that time), was discovered [very early on in this thread](https://github.com/pegjs/pegjs/issues/623#issuecomment-522691966) * right in the OP `*`...

First off, @Mingun: I really appreciate this discussion. Thanks! :smiley: #### Re "OP": By that I mean Opening Post - the [very first comment](https://github.com/pegjs/pegjs/issues/623#issue-480874026) with wich the issue was opened....

I have not forgotten about the rest: precedence, what's "expressed" in a PEG. But for now: only re your last post, @Mingun: --- ### Refactoring worked for `add` You are...

I didn't get the [call tree thing](https://github.com/pegjs/pegjs/issues/623#issuecomment-522667361) quite right. That is, for inputs `x`, `(x)`, `((x))`, ... For *n* being the number of pairs of parentheses in such an input,...