Skip to content

Commit

Permalink
Fix #222: Revamp the parser
Browse files Browse the repository at this point in the history 
Introduce a new "consume a WebVTT block" concept that is used
for parsing the header, cues, and discarding of bad cues.

This should be strictly editorial except that it also fixes #224.
It should be easier to add new block types like STYLE, and hopefully
easier to implement without having the algorithm use GOTO everywhere.
  • Loading branch information
@zcorpan
zcorpan committed Nov 6, 2015
1 parent bda95b7 commit 2cdb336
Show file tree
Hide file tree
Showing 2 changed files with 304 additions and 255 deletions.
300 changes: 174 additions & 126 deletions index.bs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -1777,7 +1777,7 @@ stream lacks this WebVTT file signature, then the parser aborts.</p>
to |input|, then the |position| pointer may, when so instructed by the algorithms, be moved past
the end of |input|.</p></li>

<li>Let |line| be a string variable. Unset the |already collected line| flag.</li>
<li>Let |line| be a string variable.</li>

<!-- SIGNATURE CHECK -->

Expand All @@ -1804,195 +1804,243 @@ stream lacks this WebVTT file signature, then the parser aborts.</p>
<li><p>The character indicated by |position| is a U+000A LINE FEED (LF) character. Advance
|position| to the next character in |input|.</p></li>

<li><p><i>Header</i>: <a spec=html>Collect a sequence of characters</a> that are <em>not</em>
U+000A LINE FEED (LF) characters. Let |line| be those characters, if any.</p></li>

<!-- METADATA HEADER PARSING -->
<li><p><i>Header</i>: If the character indicated by |position| is not a U+000A LINE FEED (LF)
character, then <a>collect a WebVTT block</a> with the <i>in header</i> flag set, and let |regions|
be the result. Otherwise, let |regions| be an empty <a>text track list of regions</a> and advance
|position| to the next character in |input|.</p></li>

<li><p>Let |regions| be a <a>text track list of regions</a>.</p></li>
<li><p><a spec=html>Collect a sequence of characters</a> that are U+000A LINE FEED (LF)
characters.</p></li>

<li>
<i>Metadata header loop</i>: If |line| is not the empty string, run the following
substeps:

<ol>

<li><p><i>Metadata header creation</i>: Let |metadata| be a new <a>WebVTT metadata
header</a>.</p></li>
<p><i>Cue loop</i>: While |position| doesn't point past the end of |input|:</p>

<li><p>Let |metadata|'s <a lt="WebVTT metadata header name">name</a> be the empty
string.</p></li>
<ol>

<li><p>Let |metadata|'s <a lt="WebVTT metadata header value">value</a> be the empty
string.</p></li>
<li><p><a>Collect a WebVTT block</a>, and let |block| be the returned value.</p></li>

<li><p>If |line| contains the character ":" (A U+003A COLON), then set <a lt="WebVTT metadata
header name">metadata's name</a> to the substring of |line| before the first ":" character and <a
lt="WebVTT metadata header value">metadata's value</a> to the substring after this
character.</p></li>
<li><p>If |block| is a <a>WebVTT cue</a>, add |block| to the <a>text track list of cues</a>
|output|.</p></li>

<li>
<p>If <a lt="WebVTT metadata header name">metadata's name</a> equals "Region":</p>
<li><p><a spec=html>Collect a sequence of characters</a> that are U+000A LINE FEED (LF)
characters.</p></li>

<ol>
<li><i>Region creation</i>: Let |region| be a new <a>WebVTT region</a>.</li>
<li>Let |region|'s <a lt="WebVTT region identifier">identifier</a> be the empty string.</li>
<li>Let |region|'s <a lt="WebVTT region width">width</a> be 100.</li>
<li>Let |region|'s <a lt="WebVTT region lines">lines</a> be 3.</li>
<li>Let |region|'s <a lt="WebVTT region anchor">anchor point</a> be (0,100).</li>
<li>Let |region|'s <a lt="WebVTT region viewport anchor">viewport anchor point</a> be
(0,100).</li>
<li>Let |region|'s <a lt="WebVTT region scroll">scroll value</a> be <a lt="WebVTT region scroll
none">NONE</a>.</li>
<li><a>Collect WebVTT region settings</a> from <a lt="WebVTT metadata header value">metadata's
value</a> using |region| for the results.</li>
<li><i>Region processing</i>: Construct a <a>WebVTT Region Object</a> from |region|.</li>
<li>Append |region| to the <a>text track list of regions</a> |regions|.</li>
</ol>
</li>
</ol>

</li>

<!-- FIXME: right now ignores all WebVTT metadata headers that don't specify regions. -->
<li><p><i>End</i>: The file has ended. Abort these steps. The <a>WebVTT parser</a> has finished.
The file was successfully processed.</p></li>

<li><p>If |position| is past the end of |input|, then jump to the step labeled <i>end</i>.</p></li>
</ol>

<li><p>The character indicated by |position| is a U+000A LINE FEED (LF) character. Advance
|position| to the next character in |input|.</p></li>
<p>When the algorithm above says to <dfn>collect a WebVTT block</dfn>, optionally with a flag <i>in
header</i> set, the user agent must run the following steps:</p>

<li><p>If |line| contains the three-character substring "<code>--></code>" (U+002D HYPHEN-MINUS,
U+002D HYPHEN-MINUS, U+003E GREATER-THAN SIGN), then set the |already collected line| flag and jump
to the step labeled <i>cue loop</i>.</p></li>
<ol algorithm="collect a WebVTT block">

<li><p>If |line| is not the empty string, then jump back to the step labeled
<i>header</i>.</p></li>
<li><p>Let |input|, |position| and |regions| be the same variables as those of the same name in the
algorithm that invoked these steps.</p></li>

<li><p><i>Cue loop</i>: If the |already collected line| flag is set, then jump to the step labeled
<i>cue creation</i>.</p></li>
<li><p>Let |line count| be &minus;1.</p></li>

<li><p><a spec=html>Collect a sequence of characters</a> that are U+000A LINE FEED (LF)
characters.</p></li>
<li><p>Let |previous position| be |position|.</p></li>

<li><p><a spec=html>Collect a sequence of characters</a> that are <em>not</em> U+000A LINE FEED
(LF) characters. Let |line| be those characters, if any.</p></li>
<li><p>Let |line| be the empty string.</p></li>

<li><p>Let |buffer| be the empty string.</p></li>

<li><p>Let |seen EOF| be false.</p></li>

<li><p>If |line| is the empty string, then jump to the step labeled <i>end</i>. (In such a case,
|position| is also forcibly past the end of |input|<!-- since we've just collected newlines, so we
have none of those, and we've failed to collect anything that's not a newline, so we have none of
that either, meaning we have nothing. -->.)</p></li>
<li><p>Let |seen arrow| be false.</p></li>

<li><p>Let |cue| be null.</p></li>

<li><p>If <i>in header</i> is set, let |regions| be a <a>text track list of regions</a>.</p></li>

<li>
<p><i>Cue creation</i>: Let |cue| be a new <a>WebVTT cue</a> and initialize it as follows:</p>

<p><i>Loop</i>: Run these substeps in a loop:</p>

<ol>
<li><p>Let |cue|'s <a>text track cue identifier</a> be the empty string.</p></li>

<li><p>Let |cue|'s <a>text track cue pause-on-exit flag</a> be false.</p></li>
<li><p><a spec=html>Collect a sequence of characters</a> that are <em>not</em> U+000A LINE FEED
(LF) characters. Let |line| be those characters, if any.</p></li>

<li><p>Let |cue|'s <a>WebVTT cue region</a> be null.</p></li>
<li><p>Increment |line count| by 1.</p></li>

<li><p>Let |cue|'s <a>WebVTT cue writing direction</a> be <a lt="WebVTT cue horizontal writing
direction">horizontal</a>.</p></li>
<li><p>If |position| is past the end of |input|, let |seen EOF| be true. Otherwise, the character
indicated by |position| is a U+000A LINE FEED (LF) character; advance |position| to the next
character in |input|.</p></li>

<li><p>Let |cue|'s <a>WebVTT cue snap-to-lines flag</a> be true.</p></li>
<li>

<li><p>Let |cue|'s <a>WebVTT cue line</a> be <a lt="WebVTT cue line automatic">auto</a>.</p></li>
<p>If |line| contains the three-character substring "<code>--></code>" (U+002D HYPHEN-MINUS,
U+002D HYPHEN-MINUS, U+003E GREATER-THAN SIGN), then run these substeps:</p>

<li><p>Let |cue|'s <a>WebVTT cue line alignment</a> be <a lt="WebVTT cue line start
alignment">start alignment</a>.</p></li>
<ol>

<li><p>Let |cue|'s <a>WebVTT cue position</a> be <a lt="WebVTT cue automatic
position">auto</a>.</p></li>
<li>

<li><p>Let |cue|'s <a>WebVTT cue position alignment</a> be <a lt="WebVTT cue position automatic
alignment">auto</a>.</p></li>
<p>If <i>in header</i> is not set and at least one of the following conditions are true:</p>

<li><p>Let |cue|'s <a>WebVTT cue size</a> be 100.</p></li>
<ul>

<li><p>Let |cue|'s <a>WebVTT cue text alignment</a> be <a lt="WebVTT cue middle alignment">middle
alignment</a>.</p></li>
<li><p>|line count| is zero</p></li>

<li><p>Let |cue|'s <a>text track cue text</a> be the empty string.</p></li>
</ol>
<li><p>|line count| is 1 and |seen arrow| is false</p></li>

</li>
</ul>

<li><p>If |line| contains the three-character substring "<code>--></code>" (U+002D HYPHEN-MINUS,
U+002D HYPHEN-MINUS, U+003E GREATER-THAN SIGN), then jump to the step labeled <i>timings</i>
below.</p></li>
<p>...then run these substeps:</p>

<li><p>Let |cue|'s <a>text track cue identifier</a> be |line|.</p></li>
<ol>

<li><p>If |position| is past the end of |input|, then discard |cue| and jump to the step labeled
<i>end</i>.</p></li>
<li><p>Let |seen arrow| be true.</p></li>

<li><p>If the character indicated by |position| is a U+000A LINE FEED (LF) character, advance
|position| to the next character in |input|.</p></li>
<li><p>Let |previous position| be |position|.</p></li>

<li><p><a spec=html>Collect a sequence of characters</a> that are <em>not</em> U+000A LINE FEED
(LF) characters. Let |line| be those characters, if any.</p></li>
<li>

<li><p>If |line| is the empty string, then discard |cue| and jump to the step labeled <i>cue
loop</i>.</p></li>
<p><i>Cue creation</i>: Let |cue| be a new <a>WebVTT cue</a> and initialize it as
follows:</p>

<li><p><i>Timings</i>: Unset the |already collected line| flag.</p></li>
<ol>

<li><p><a>Collect WebVTT cue timings and settings</a> from |line| using |regions| for |cue|. If
that fails, jump to the step labeled <i>bad cue</i>.</p></li>
<li><p>Let |cue|'s <a>text track cue identifier</a> be the empty string.</p></li>

<li><p>Let |cue text| be the empty string.</p></li>
<li><p>Let |cue|'s <a>text track cue pause-on-exit flag</a> be false.</p></li>

<li><p><i>Cue text loop</i>: If |position| is past the end of |input|, then jump to the step
labeled <i>cue text processing</i>.</p></li>
<li><p>Let |cue|'s <a>WebVTT cue region</a> be null.</p></li>

<li><p>If the character indicated by |position| is a U+000A LINE FEED (LF) character, advance
|position| to the next character in |input|.</p></li>
<li><p>Let |cue|'s <a>WebVTT cue writing direction</a> be <a lt="WebVTT cue horizontal
writing direction">horizontal</a>.</p></li>

<li><p><a spec=html>Collect a sequence of characters</a> that are <em>not</em> U+000A LINE FEED
(LF) characters. Let |line| be those characters, if any.</p></li>
<li><p>Let |cue|'s <a>WebVTT cue snap-to-lines flag</a> be true.</p></li>

<li><p>If |line| is the empty string, then jump to the step labeled <i>cue text
processing</i>.</p></li>
<li><p>Let |cue|'s <a>WebVTT cue line</a> be <a lt="WebVTT cue line
automatic">auto</a>.</p></li>

<li><p>If |line| contains the three-character substring "<code>--></code>" (U+002D HYPHEN-MINUS,
U+002D HYPHEN-MINUS, U+003E GREATER-THAN SIGN), then set the |already collected line| flag and jump
to the step labeled <i>cue text processing</i>.</p></li>
<li><p>Let |cue|'s <a>WebVTT cue line alignment</a> be <a lt="WebVTT cue line start
alignment">start alignment</a>.</p></li>

<li><p>If |cue text| is not empty, append a U+000A LINE FEED (LF) character to |cue text|.</p></li>
<li><p>Let |cue|'s <a>WebVTT cue position</a> be <a lt="WebVTT cue automatic
position">auto</a>.</p></li>

<li><p>Let |cue text| be the concatenation of |cue text| and |line|.</p></li>
<li><p>Let |cue|'s <a>WebVTT cue position alignment</a> be <a lt="WebVTT cue position
automatic alignment">auto</a>.</p></li>

<li><p>Return to the step labeled <i>cue text loop</i>.</p></li>
<li><p>Let |cue|'s <a>WebVTT cue size</a> be 100.</p></li>

<li><p><i>Cue text processing</i>: Let the <a>text track cue text</a> of |cue| be |cue text|, and
let the <a>rules for extracting the chapter title</a> be the <a>WebVTT rules for extracting the
chapter title</a>.</p></li>
<li><p>Let |cue|'s <a>WebVTT cue text alignment</a> be <a lt="WebVTT cue middle
alignment">middle alignment</a>.</p></li>

<li><p>Add |cue| to the <a>text track list of cues</a> |output|.</p></li>
<li><p>Let |cue|'s <a>text track cue text</a> be the empty string.</p></li>

<li><p>Jump to the step labeled <i>cue loop</i>.</p></li>
</ol>

<li><p><i>Bad cue</i>: Discard |cue|.</p></li>
</li>

<li><p><i>Bad cue loop</i>: If |position| is past the end of |input|, then jump to the step labeled
<i>end</i>.</p></li>
<li><p><a>Collect WebVTT cue timings and settings</a> from |line| using |regions| for |cue|.
If that fails, let |cue| be null. Otherwise, let |buffer| be the empty string.</p></li>

<li><p>If the character indicated by |position| is a U+000A LINE FEED (LF) character, advance
|position| to the next character in |input|.</p></li>
</ol>

<li><p><a spec=html>Collect a sequence of characters</a> that are <em>not</em> U+000A LINE FEED
(LF) characters. Let |line| be those characters, if any.</p></li>
<p>Otherwise, let |position| be |previous position| and break out of <i>loop</i>.</p>

<li><p>If |line| contains the three-character substring "<code>--></code>" (U+002D HYPHEN-MINUS,
U+002D HYPHEN-MINUS, U+003E GREATER-THAN SIGN), then set the |already collected line| flag and jump
to the step labeled <i>cue loop</i>.</p></li>
</li>

<li><p>If |line| is the empty string, then jump to the step labeled <i>cue loop</i>.</p></li>
</ol>

<li><p>Otherwise, jump to the step labeled <i>bad cue loop</i>.</p></li>
</li>

<li><p><i>End</i>: The file has ended. Abort these steps. The <a>WebVTT parser</a> has finished.
The file was successfully processed.</p></li>
<li><p>Otherwise, if |line| is the empty string, break out of <i>loop</i>.</p></li>

<li>

<p>Otherwise, run these substeps:</p>

<ol>

<!-- <li> <p>If |line count| is 1, run these substeps:</p> <ol> parse new block types here
</ol> </li> -->

<li>

<p>If <i>in header</i> is set, run these substeps:</p>

<ol>

<li>

<p>If |line| starts with the substring "<code>Region:</code>" (U+0052 LATIN CAPITAL LETTER R
character, U+0065 LATIN SMALL LETTER E character, U+0067 LATIN SMALL LETTER G character,
U+0069 LATIN SMALL LETTER I character, U+006F LATIN SMALL LETTER O character, U+006E LATIN
SMALL LETTER N character, U+003A COLON character (:)), run these substeps:</p>

<ol>

<li><p><i>Region creation</i>: Let |region| be a new <a>WebVTT region</a>.</p></li>

<li>Let |region|'s <a lt="WebVTT region identifier">identifier</a> be the empty
string.</li>

<li>Let |region|'s <a lt="WebVTT region width">width</a> be 100.</li>

<li>Let |region|'s <a lt="WebVTT region lines">lines</a> be 3.</li>

<li>Let |region|'s <a lt="WebVTT region anchor">anchor point</a> be (0,100).</li>

<li>Let |region|'s <a lt="WebVTT region viewport anchor">viewport anchor point</a> be
(0,100).</li>

<li>Let |region|'s <a lt="WebVTT region scroll">scroll value</a> be <a lt="WebVTT region
scroll none">NONE</a>.</li>

<li><p>Let |region value| be the substring of |line| after the first U+003A COLON character
(:).</p></li>

<li><a>Collect WebVTT region settings</a> from |region value| using |region| for the
results.</li>

<li><i>Region processing</i>: Construct a <a>WebVTT Region Object</a> from |region|.</li>

<li>Append |region| to the <a>text track list of regions</a> |regions|.</li>

</ol>

</li>

</ol>

</li>

<li><p>If |buffer| is not the empty string, append a U+000A LINE FEED (LF) character to
|buffer|.</p></li>

<li><p>Append |line| to |buffer|.</p></li>

<li><p>Let |previous position| be |position|.</p></li>

</ol>

</li>

<li><p>If |seen EOF| is true, break out of <i>loop</i>.</p></li>

</ol>

</li>

<li><p>If |cue| is not null, let the <a>text track cue text</a> of |cue| be |buffer|, and return
|cue|.</p></li>

<!-- return new block types here -->

<li><p>Otherwise, if <i>in header</i> is set, return |regions|.</p></li>

<li><p>Otherwise, return null.</p></li>

</ol>

Expand Down

0 comments on commit 2cdb336

Please sign in to comment.