In preparation for switching this to a plain string in
PHP-Parser 5, deprecate direct access to the property and
provide an API that will work on both versions.
(cherry picked from commit c9e5a13d68486e9fd75f9be1b4639644e54e7f4f)
Normalize CRLF to LF in getReformattedText(). That was, if the
comment is pretty-printed, we will use proper LF newlines, rather
than inserting indentation between the CR and LF.
At the same time, this also makes it easier to emit actual CRLF
newlines with a custom pretty printer.
Fixes#599.
Node visitation is now properly nested. The call sequence will
now be
$visitor1->enterNode($n);
$visitor2->enterNode($n);
$visitor2->leaveNode($n);
$visitor1->leaveNode($n);
rather than
$visitor1->enterNode($n);
$visitor2->enterNode($n);
$visitor1->leaveNode($n);
$visitor2->leaveNode($n);
Fixes#899.
For types the use of a string is ambiguous -- it could be either
an Identifier or a Name. Don't guess.
Retain the implicit promotion to Identifier in places where only
Identifier is legal, e.g. various symbol names.
By making flags on the Param builder configurable by providing make(Public|Protected|Private) methods we can promote parameters to properties from the constructor
In formatting-preserving pretty printing, treat DEL as a label
character based on the target PHP version (the default of 7.1
implying it isn't one).
This avoids failure to round-trip an unchanged input.
We shouldn't ltrim when printing a whole file, that way we will
not just fail to preserve formatting, but actually change semantics
by dropping meaningful whitespace.
We may need parentheses around an argument-less yield to distinguish
(yield) - $a from yield -$a and similar.
Treat argument-less yield like a prefix operator. This will print
parentheses a bit more often than is really required, because the
arity ambiguity only exists for certain operators. This seems like
a reasonably good approximation through, as it only affects
argument-less yield on the LHS of infix operators.
Doc strings have a trailing \n and these will get interpreted as
\r\n and removed from the string contents.
For nowdoc, fall back to single quote if there's a trailing \r.
For heredoc, escape all isolated \r -- unlike \n and \r\n this is
really a special character, because this is no longer relevant as
an actual newline character.
Since PHP 7.0 yield is a proper expression, so print it with
proper precedence. If the pretty printer is configured for
older version (non-default), then always print parentheses.
There is an interesting interaction here, in that => is resolved
in favor of yield if the yield occurs as part of an array (or
outer yield). This kind of bypasses the entire precedence hierarchy
and *only* affects yield expressions. For the sake of simplicity
this is modeled via normal LHS precedence, because this will only
add unnecessary parentheses to a handful of low precedence unary
operators. If desired, a special marker for this purpose could be
added though.
The pretty printer is supposed to produce a minimal-parentheses
printing for expressions. However, for prefix operators, we were
failing to do so in ways that are practically meaningful, e.g.
$a = yield from $b produced redundant parentheses around the
yield from.
For prefix operators, the precedence of the direct parent operator
is not actually relevant: What matters is the precedence of any
infix operator whose LHS it appears on. For example, $a . include $b
does not require parentheses, but (include $a) . $b does.
To handle this, keep separate track of the parent operator
precedence, and a special LHS precedence which is used for unary
operator printing only. Because the LHS precedence may not come
from the direct parent, we have to pass it down the call stack.
And if we do that, we may as well do the same for the parent
precedence.
This also fixes an integration test failure with php-src, because
arrow function precedence was previously not respected (and would
be ugly to respect without this change).
When detecting whether the string contains the end label, allow
leading whitespace in front of it. This is legal since the
introduction of flexible doc strings.
With the introduction of flexible doc strings, the ending label
is no longer required to be followed by a semicolon or newline.
We need to prevent doc string printing if the label is followed
by any non-label character.
This makes pretty printing round trip to another Float literal,
rather than a constant lookup. The 1e1000 form in particular is
chosen because that seems to be the typical form used in various
tests.
