mirror of
https://gerrit.wikimedia.org/r/mediawiki/extensions/VisualEditor
synced 2024-12-01 09:26:37 +00:00
Minor improvement to italic/bold, documentation on failed modularization of
static parser functions.
This commit is contained in:
Gabriel Wicke
parent
8def550629
commit
694b998f24
|
|
@ -1,11 +1,47 @@
|
|||
/* Produces output more or less compatible with FakeParser; plug it into FP's output and see */
|
||||
{
|
||||
/* Fixme: use static functions to separate module! Unfortunately, this
|
||||
* does not work:
|
||||
* var tu = require('./mediawiki.tokenizer.utils.js');
|
||||
* console.log(tu.flatten([]));
|
||||
* Using exports in the module gets a bit further, but accesses to
|
||||
* tu.flatten in productions still fail. Thus, I just moved the functions
|
||||
* here until a solution is found:
|
||||
*/
|
||||
|
||||
/* Static utilities */
|
||||
|
||||
// Flatten a list of lists.
|
||||
var flatten = function ( e ) {
|
||||
var es = [];
|
||||
// flatten sub-arrays
|
||||
for(var i = 0, length = e.length; i < length; i++) {
|
||||
var ei = e[i];
|
||||
if ($.isArray(ei))
|
||||
es = es.concat(flatten(ei));
|
||||
else
|
||||
es.push(ei);
|
||||
};
|
||||
return es;
|
||||
};
|
||||
|
||||
// Remove escaped quotes from attributes etc
|
||||
var unquote = function (quotec, text) {
|
||||
return text.replace('\\' + quotec, quotec);
|
||||
};
|
||||
|
||||
|
||||
// Debug print with global switch
|
||||
var dp = function ( msg ) {
|
||||
if ( false ) {
|
||||
console.log(msg);
|
||||
}
|
||||
};
|
||||
|
||||
var pp = function ( s ) { return JSON.stringify(s, null, 2); }
|
||||
|
||||
/* End static utilities */
|
||||
|
||||
/*
|
||||
* Flags for specific parse environments (inside tables, links etc). Flags
|
||||
* trigger syntactic stops in the inline_breaks production, which
|
||||
|
|
@ -27,89 +63,9 @@
|
|||
syntaxFlags[flag]--;
|
||||
};
|
||||
|
||||
|
||||
var pp = function ( s ) { return JSON.stringify(s, null, 2); }
|
||||
|
||||
// Convert list prefixes to a list of WikiDom list styles
|
||||
var bulletsToTypes = function (bullets) {
|
||||
var bTypes = [];
|
||||
var blen = bullets.length;
|
||||
for (var i = 0; i < bullets.length; i++) {
|
||||
switch (bullets[i]) {
|
||||
case '*':
|
||||
bTypes.push('bullet'); break;
|
||||
case '#':
|
||||
bTypes.push('number'); break;
|
||||
case ';':
|
||||
bTypes.push('term'); break;
|
||||
case ':':
|
||||
bTypes.push('description'); break;
|
||||
}
|
||||
}
|
||||
return bTypes;
|
||||
};
|
||||
|
||||
/*var extractInline = function ( node ) {
|
||||
return { text: extractText(node, 0) };
|
||||
};
|
||||
|
||||
|
||||
// return [text [annotations]]
|
||||
var extractText = function ( node, offset ) {
|
||||
dp("extract: " + pp(node));
|
||||
if (typeof node === 'string') {
|
||||
return [node, []];
|
||||
} else if ($.isArray(node)) {
|
||||
var texts = [],
|
||||
annotations = [];
|
||||
for (var i = 0, length = node.length; i < length; i++) {
|
||||
var res = extractText(node[i], offset);
|
||||
texts.push(res[0]);
|
||||
annotations.concat(res[1]);
|
||||
offset += res[0].length;
|
||||
}
|
||||
return [texts.join(''), annotations];
|
||||
} else if ( 'text' in node ) {
|
||||
var res = extractText(node, offset);
|
||||
if ('annotations' in node) {
|
||||
return [res[0], node.annotations.concat(res[1])];
|
||||
} else {
|
||||
return res;
|
||||
}
|
||||
} else if ( 'content' in node ) {
|
||||
return extractText(node.content, offset);
|
||||
} else if ( 'children' in node ) {
|
||||
var texts = [];
|
||||
for (var i = 0, length = node.children.length; i < length; i++) {
|
||||
texts.push(extractText(node.children[i]));
|
||||
}
|
||||
return texts.join('');
|
||||
} else {
|
||||
throw ("extract failed: " + pp(node));
|
||||
}
|
||||
};
|
||||
*/
|
||||
|
||||
// Start position of top-level block
|
||||
// Could also provide positions for lower-level blocks using a stack.
|
||||
var blockStart = 0;
|
||||
|
||||
var unquote = function (quotec, text) {
|
||||
return text.replace('\\' + quotec, quotec);
|
||||
};
|
||||
|
||||
var flatten = function ( e ) {
|
||||
var es = [];
|
||||
// flatten sub-arrays
|
||||
for(var i = 0, length = e.length; i < length; i++) {
|
||||
var ei = e[i];
|
||||
if ($.isArray(ei))
|
||||
es = es.concat(flatten(ei));
|
||||
else
|
||||
es.push(ei);
|
||||
};
|
||||
return es;
|
||||
};
|
||||
}
|
||||
|
||||
start
|
||||
|
|
@ -455,26 +411,33 @@ link_text
|
|||
|
||||
link_end = "]]"
|
||||
|
||||
/* This implementation of bold and italic is very basic so far, and misses the
|
||||
* finer points of doQuotes in the parser. A rough plan to get closer:
|
||||
* - '''' -> ' '''
|
||||
* - last ''''' in a row of ' is used
|
||||
* - if *both* italics and bolds are unbalanced, check for prefix
|
||||
* - convert single-letter or multi-letter non-space prefixed tick back to
|
||||
* text
|
||||
*/
|
||||
bold
|
||||
= bold_marker
|
||||
& { dp('benter:' + pos); return setFlag('bold'); }
|
||||
c:inlineline
|
||||
bold_marker {
|
||||
(bold_marker / &newline) {
|
||||
clearFlag('bold');
|
||||
return [{ type: 'TAG', name: 'b' }]
|
||||
.concat(c, [{type: 'ENDTAG', name: 'b'}]);
|
||||
}
|
||||
/ bold_marker { clearFlag('bold'); return null }
|
||||
|
||||
bold_marker
|
||||
= "'''"
|
||||
bold_marker = "'''"
|
||||
|
||||
|
||||
italic
|
||||
= italic_marker
|
||||
& { dp('ienter:' + pos); return setFlag('italic'); }
|
||||
c:inlineline
|
||||
italic_marker {
|
||||
(italic_marker / &newline) {
|
||||
clearFlag('italic');
|
||||
dp('ileave:' + pos);
|
||||
return [{ type: 'TAG', name: 'i' }]
|
||||
|
|
@ -482,8 +445,7 @@ italic
|
|||
}
|
||||
/ italic_marker { clearFlag('italic'); return null }
|
||||
|
||||
italic_marker
|
||||
= "''"
|
||||
italic_marker = "''"
|
||||
|
||||
/* Will need to check anything xmlish agains known/allowed HTML tags and
|
||||
* registered extensions, otherwise fail the match. Should ref be treated as a
|
||||
|
|
|
|||
Loading…
Reference in a new issue