The goal is to allow users to attach doc strings and
`private/protected` to constructor declarations.
TODO: reject non applicable modifiers such as `unsafe` and `partial`.
cc @Kha
@Kha I elaborated it as a definition. It works because we can now
reference Parser declarations in `syntax` command.
This change allowed us to replace `p.getArg 0` with `p` in the
`Websever` demo.
@Kha Note that I had to change the pattern. After I replaced,
```
syntax text : child
```
with
```
syntax Prelim.text : child
```
Thus, I wrote
```
`(child|$t:text)
```
as
```
`(child|$t)
```
It works for this example, but it may be a problem in general.
@Kha We can now delete the command
```
@[termParser] def childStxQuot : Parser :=
checkPrec maxPrec >> (node `Lean.Parser.Term.stxQuot $ symbol "`(child|" >> categoryParser `child 0 >> symbol ")")
```
from the web demo.
We only keep trying other macros if the macro `m` produced
`unsupportedSyntax`. If the macro produced an error, we keep the
error.
@Kha This should fix the web demo.
@Kha Note that I had to write the weird pattern
```
match_syntax stx with
| `(notation:$prec $items* => $rhs) => expandNotationAux stx prec items rhs
| `(notation $noprec* $items* => $rhs) => expandNotationAux stx none items rhs
| _ => Macro.throwUnsupported
```
with the weird `$noprec*` to match the case where the optional
precedence is not provided. I realized this is not a bug, but
I guess most users will be puzzled by this behavior. If we had a kind
for `notationItem`, I would be able to write
```
`(notation $items:notationItems* => $rhs)
```
@Kha I implemented the solution 1 I described at Zulip.
I also tried to document the issue, and made sure test
`choiceMacroRules` fail if the bug is reintroduced.
@Kha It is not clear to me why I had to change the following line
```
-syntax term ">>>" term : foo
+syntax term:1 ">>>":1 term : foo
```
The test breaks without it.
