Please review a change that will permit the use of a previously illegal
construction, to allow <newline> <spaces> `@` inside inline tags. This
will allow "<pre>{@code..." constructions that contain annotations, such
as the following:
/**
* This is a method.
* <pre>{@code
* @Override
* void m() { }
* }</pre>
*/
Previously, the text was first parsed into the main body and subsequent
block tags, and only then were those analyzed for inline tags. That
meant that the example just given was invalid, for having an incomplete
inline tag between `<pre>` and an apparent block tag named `@Override`.
With the change, `@` at the beginning of a line inside an inline tag
will not be treated as the beginning of a block tag, and so the comment
will parse as might be expected.
The change to the code is as simple as deleting the code that detected
block tags inside inline tags.
There are some minor compatibility effects.
1. Some comments that were previously invalid may become valid. This is
the desired effect.
2. Some comments that previously invalid may now be parsed differently.
In particular, in the case of a genuinely missing '}', the parse will
now swallow up any block tags that might follow. For example, consider
the following:
/**
* This is a method.
* @param p1 this has an unbalanced {@code description
* @param p2 this is the second parameter
*/
void m(int p1, int p2) { }
As a result of the change, the description for parameter p2 will be
swallowed up as part of the invalid description for p1. This will only
be visible in error messages and clients of the API that analyze
erroneous comments.
The tests are updated to accommodate the change. A specific test for the
`<pre>{@code...}</pre>` construction is added. The biggest change is to
the test code that "predicts" the output of the AST pretty printer,
which is now updated to better handle the new behavior of `{@code}` and
`{@literal}` tags.
-- Jon
JBS: https://bugs.openjdk.java.net/browse/JDK-8241780
Webrev: http://cr.openjdk.java.net/~jjg/8241780/webrev.00/index.html
