runtime(java): Improve the recognition of the "indent" method declarations (vim/vim#14659)
There is a flaw in the current implementation that has been
exacerbated around v5.2. It lies in the recognition of all
three indentation styles simultaneously: a tab, two space,
and eight space character(s). With it, it is not uncommon
to misidentify various constructs as method declarations
when they belong to two-space indented members and other
blocks of a type and are offset at eight space characters or
a tab from the start of the line.
For example,
------------------------------------------------------------
class Test
{
static String hello() { return "hello"; }
public static void main(String[] args)
{
try {
if (args.length > 0) {
// FIXME: eight spaces.
System.out.println(args[0]);
} else {
// FIXME: a tab.
System.out.println(hello());
}
} catch (Exception e) {
throw new Error(e);
}
}
}
------------------------------------------------------------
------------------------------------------------------------
:let g:java_highlight_functions = 'indent'
:doautocmd Syntax
------------------------------------------------------------
A better approach is to pick an only indentation style out
of all supported styles (so either two spaces _or_ eight
spaces _or_ a tab). Note that tabs and spaces can still be
mixed, only the leading tab or the leading run of spaces
matters for the recognition. And there is no reason to not
complement the set of valid styles with any number of spaces
from 1 to 8, inclusively.
Please proceed with the necessary change as follows:
- rename from "indent" to "indent2" for a 2-space run;
- rename from "indent" to "indent8" for an 8-space run;
- continue to have "indent" for a tab run;
- define an "indent" variable with a suffix number denoting
the preferred amount of indentation for any other run of
spaces [1-8].
As before, this alternative style of recognition of method
declarations still does not prescribe naming conventions and
still cannot recognise method declarations in nested types
that are conventionally indented.
The proposed changes also follow suit of "style" in stopping
the claiming of constructor and enum constant declarations.
c4d0c8c812
Co-authored-by: Aliaksei Budavei <32549825+zzzyxwvut@users.noreply.github.com>
runtime(java): Improve the recognition of the "style" method declarations
- Request the new regexp engine (v7.3.970) for [:upper:] and
[:lower:].
- Recognise declarations of in-line annotated methods.
- Recognise declarations of _strictfp_ methods.
- Establish partial order for method modifiers as shown in
the MethodModifier production; namely, _public_ and
friends should be written the leftmost, possibly followed
by _abstract_ or _default_, or possibly followed by other
modifiers.
- Stop looking for parameterisable primitive types (void<?>,
int<Object>, etc., are malformed).
- Stop looking for arrays of _void_.
- Acknowledge the prevailing convention for method names to
begin with a small letter and for class/interface names to
begin with a capital letter; and, therefore, desist from
claiming declarations of enum constants and constructors
with javaFuncDef.
Rationale:
+ Constructor is distinct from method:
* its (overloaded) name is not arbitrary;
* its return type is implicit;
* its _throws_ clause depends on indirect vagaries of
instance (variable) initialisers;
* its invocation makes other constructors of its type
hierarchy invoked one by one, concluding with the
primordial constructor;
* its explicit invocation, via _this_ or _super_, can
only appear as the first statement in a constructor
(not anymore, see JEP 447); else, its _super_ call
cannot appear in constructors of _record_ or _enum_;
and neither invocation is allowed for the primordial
constructor;
* it is not a member of its class, like initialisers,
and is never inherited;
* it is never _abstract_ or _native_.
+ Constructor declarations tend to be few in number and
merit visual recognition from method declarations.
+ Enum constants define a fixed set of type instances
and more resemble class variable initialisers.
Note that the code duplicated for @javaFuncParams is written
keeping in mind for g:java_highlight_functions a pending 3rd
variant, which would require none of the :syn-cluster added
groups.
closes: vim/vim#14620a4c085a3e6
Co-authored-by: Aliaksei Budavei <0x000c70@gmail.com>
runtime(java): Recognise non-ASCII identifiers (vim/vim#14543)
* runtime(java): Recognise non-ASCII identifiers
Also:
- Remove the already commented out and less general in its
definition javaFuncDef alternative.
- Stop recognising some bespoke {p,trace} debugging API.
Non-ASCII identifiers have been supported from the outset
of the Java language.
> An _identifier_ is an unlimited-length sequence of _Java
> letters_ and _Java digits_, the first of which must be a
> Java letter. An identifier cannot have the same spelling
> (Unicode character sequence) as a keyword . . . Boolean
> literal . . . or the null literal . . .
> . . . . . . . .
> Letters and digits may be drawn from the entire Unicode
> character set . . .
> . . . . . . . .
> A Java letter is a character for which the method
> Character.isJavaLetter . . . returns true. A Java
> letter-or-digit is a character for which the method
> Character.isJavaLetterOrDigit . . . returns true.
> . . . . . . . .
> The Java letters include . . . for historical reasons, the
> ASCII underscore (_) . . . and dollar sign ($) . . .
(Separate syntax tests will be written when particular parts
now touched will have been further improved.)
Reference:
https://javaalmanac.io/jdk/1.0/langspec.pdf [§3.8]
* Take on the maintenance of Java filetype and syntax files
4052474a1b
Co-authored-by: Aliaksei Budavei <32549825+zzzyxwvut@users.noreply.github.com>
runtime(java): Improve the matching of contextual keywords
- Recognise a _record_ contextual keyword.
- Recognise _non-sealed_, _sealed_, and _permits_ contextual
keywords.
- Admit _$_ to keyword characters.
- Group _abstract_, _final_, _default_, _(non-)sealed_
(apart from _(non-)sealed_, the incompossibility of these
modifiers calls for attention).
- Remove another _synchronized_ keyword redefinition.
I have also replaced a function with an expression. Before
patch 8.1.0515, it should have been declared :function! to
work with repeatable script sourcing; there is less to worry
about with an expression.
References:
https://openjdk.org/jeps/395 (Records)
https://openjdk.org/jeps/409 (Sealed Classes)
https://docs.oracle.com/javase/specs/jls/se21/html/jls-3.html#jls-3.8closes: vim/vim#144035ccdcc482e
Co-authored-by: Aliaksei Budavei <0x000c70@gmail.com>
runtime(java): Recognise the inline kind of the {@return} tag (vim/vim#14284)
Also:
- Refine comment matching (javaComment{Error\ and,Start}).
- Continue rewriting regexps (prefer atom grouping with
non-capturing parens; factor out common prefixes in
alternations).
- Allow for relative paths with the _file_ attribute of
{@snippet}.
- Anticipate HTML in the @see tags.
- Match the nullary method parens in javaDocSeeTagParam.
- Improve the boundary patterns for summary sentences of
documentation.
> This sentence ends at ... or at the first tag (as defined
> below).
There are Java documentation tags (@) and there are HTML
tags (<?>) (with Markdown looming large; see JEP 467). With
block tags, e.g. @param, @return, @see, we begin another
documentation "sentence" whether or not the author has
terminated the summary sentence with a period; with
.<!-- -->, we may follow abbreviations, enumerations,
initials, (but instead consider @literal or ) _within_
the summary sentence. On the other hand, inline tags, e.g.
@code, @link, @literal, should not terminate the summary
sentence.
References:
https://bugs.openjdk.org/browse/JDK-8075778https://www.oracle.com/technical-resources/articles/java/javadoc-tool.html#firstsentencehttps://docs.oracle.com/en/java/javase/21/docs/specs/javadoc/doc-comment-spec.html8e59a7ba88
Co-authored-by: Aliaksei Budavei <32549825+zzzyxwvut@users.noreply.github.com>
runtime(java): Recognise the {@snippet} documentation tag (vim/vim#14271)
Remember that ‘code fragments are typically Java source
code, but they may also be fragments of properties files,
source code in other languages, or plain text.’ Therefore,
with these changes, markup tags are highlighted in the Java
source files (as external snippets) and in the {@snippet}
tags.
Also:
- Improve matching of the multi-line {@code} documentation
tag with any contained balanced braces.
- Recognise the {@literal} documentation tag.
- Highlight stray blanks in comments.
Related to an enhancement proposal for PCRE-like callouts
discussed at https://github.com/vim/vim/issues/11217.
References:
https://openjdk.org/jeps/413https://docs.oracle.com/en/java/javase/21/docs/specs/javadoc/doc-comment-spec.html3e72bf10a0
Co-authored-by: Aliaksei Budavei <32549825+zzzyxwvut@users.noreply.github.com>
runtime(java): Recognise string templates (vim/vim#14150)
As this is encouraged in the referenced JEPs, "to visually
distinguish a string template from a string literal, and
a text block template from a text block", the default
colours for java\%[Debug]StrTempl are made distinct from
java\%[Debug]String.
According to §3.2 Lexical Translations (JLS, c. 1996 or any
more recent version), line terminators, white space, and
comments are discarded before tokens are accepted. Since
a template expression comprises a template processor, a dot,
and a template, it may be visually appealing to break up
its head across a few lines whenever its tail already spans
multiple lines. Curiously, no allowance for it is made in
the distributed tests for OpenJDK 21; the proposed regexp
patterns take in consideration a line terminator and white
space after a dot.
References:
https://openjdk.org/jeps/430 (Preview)
https://openjdk.org/jeps/459 (Second Preview)
https://openjdk.org/jeps/465a2c65809da
Co-authored-by: Aliaksei Budavei <32549825+zzzyxwvut@users.noreply.github.com>
runtime(java): Recognise text blocks (vim/vim#14128)
Also, accept as valid the space escape sequence `\s`.
Also, consistently use the claimed `javaDebug` prefix for
syntax group definitions kept under `g:java_highlight_debug`.
Since `javaStringError` is commented out for its generality,
let's comment out `javaDebugStringError`, its copy, as well.
References:
https://openjdk.org/jeps/378https://docs.oracle.com/javase/specs/jls/se17/html/jls-3.html#jls-3.10.7Closesvim/vim#10910.
b3030b653b
Co-authored-by: Aliaksei Budavei <32549825+zzzyxwvut@users.noreply.github.com>
runtime(java): Recognise _when_ clauses in _switch_ blocks
Also:
- distinguish _yield_ when used as a contextual keyword from
when used qualified as a method or a method reference (as
can be seen in testdir/input/java_switch.java, variables
and method declarations named _yield_ will be recognised
as the namesake keyword--consider picking other names for
variables, and defining g:java_highlight_functions to have
method names painted; since _yield_ statements can have
trailing parens, they must be recognised as statements,
for only qualified _yield_ method calls are supported);
- recognise grouped _default_ _case_ labels;
- describe primitive types for _case_ labels (JLS, §14.11,
§3.10.1);
- recognise some non-ASCII identifiers (see javaLambdaDef,
javaUserLabel) (further improvement for better recognition
of identifiers will be arranged in a separate PR).
Because the arrow '->' is used in two kinds of expressions,
lambda (abstractions) and _switch_, necessary changes were
made for the recognition of either (and further improvement
touching lambda expressions will be separately arranged).
Because 'default' is used for instance method declarations
in interfaces and in _switch_ labels, necessary changes were
made for the recognition of either (and further improvement
touching method declarations will be separately arranged).
Finally, it deemed appropriate to put 'yield' in the syntax
group of javaOperator rather than javaStatement, for its
member 'var' is also another contextual keyword (e.g., this
is valid syntax: "var var = var(test.var);").
References:
https://openjdk.org/jeps/361 (Switch Expressions)
https://openjdk.org/jeps/440 (Record Patterns)
https://openjdk.org/jeps/441 (Pattern Matching for switch)
Also, add a Java specific filetype plugin for the syntax
test, so that no soft-wrapping of long indented lines occur.
Otherwise the syntax scripts would miss a few lines during
scrolling and verification of the screen dumps.
closes: vim/vim#141059ecf02cd5f
Co-authored-by: Aliaksei Budavei <0x000c70@gmail.com>
Vim runtime files based on 7.4.384 / hg changeset 7090d7f160f7
Excluding:
Amiga icons (*.info, icons/)
doc/hangulin.txt
tutor/
spell/
lang/ (only used for menu translations)
macros/maze/, macros/hanoi/, macros/life/, macros/urm/
These were used to test vi compatibility.
termcap
"Demonstration of a termcap file (for the Amiga and Archimedes)"
Helped-by: Rich Wareham <rjw57@cam.ac.uk>
Helped-by: John <john.schmidt.h@gmail.com>
Helped-by: Yann <yann@yann-salaun.com>
Helped-by: Christophe Badoit <c.badoit@lesiteimmo.com>
Helped-by: drasill <github@tof2k.com>
Helped-by: Tae Sandoval Murgan <taecilla@gmail.com>
Helped-by: Lowe Thiderman <lowe.thiderman@gmail.com>