Some websites may use "public" or "system" in doctypes, or completely
messed up casing such as PuBlIc (unlikely, but possible). This ensures
we don't care about the exact casing used.
This fixes https://gitlab.com/yorickpeterse/oga/issues/199
This ensures that Oga is able to tokenize input such as the following:
<script<script>foo</script>
Oga will now treat this as:
<script>foo</script>
This is based on libxml behaviour, which seems to differ a bit from
Chromium which treats the node as a text node. This however would
require complex look-ahead logic (as far as I can tell) that I really
don't want to implement in Oga.
Fixes#186
HTML identifiers containing colons should be treated in two ways:
* For element names the prefix (= the namespace prefix in case of XML)
should be ignored as HTML doesn't support/use namespaces.
* For attribute names a colon is a valid character, thus "foo:bar:baz"
should be treated as a single attribute name.
This fixes#142.
This allows for parsing of HTML such as:
<a href=lol("javascript")></a>
Here the "href" attribute would have its value set to:
lol("javascript")
Fixes#119
This allows the lexer to process input such as:
<a href=foo"></a>
For XML input the lexer still expects properly opened/closed attribute
values.
Fixes#109
Previous HTML such as this would be lexed incorrectly:
<div>
<ul>
<li>foo
</ul>
inside div
</div>
outside div
The lexer would see this as the following instead:
<div>
<ul>
<li>foo</li>
inside div
</ul>
outside div
</div>
This commit exposes the name of the closing tag to
XML::Lexer#on_element_end (omitted for self closing tags). This can be
used to automatically close nested tags that were left open, ensuring
the above HTML is lexer correctly.
The new setup ignores namespace prefixes as these are not used in HTML,
XML in turn won't even run the code to begin with since it doesn't allow
one to leave out closing tags.
This makes it easier to automatically insert preceding tokens when
starting a new element as we now have access to the name. Previously
on_element_start would be invoked first which doesn't receive an
argument.
Previously a single Ragel machine was used for processing HTML
script and style tags. This had the unfortunate side-effect that the
following was not parsed correctly (while being valid HTML):
<script>
var foo = "</style>";
</script>
The same applied to style tags:
<style>
/* </script> */
</style>
By using separate machines we can work around the above issue. The
downside is that this can produce multiple T_TEXT nodes, which have to
be stitched back together in the parser.
This adds lexing support for HTML/XML such as:
<foo bar="""></foo>
While technically invalid, some websites (e.g. yahoo.com) contain HTML
just like this.
The lexer handles this as following:
1. When we're in the "element_head" machine, do business as usual until
we bump into a "=".
2. Call (using Ragel's "fcall") the machine to use for processing the
attribute value (if any).
3. In this machine quoted strings are processed. The moment a string has
been processed the lexer jumps right back in to the "element_head"
machine. This ensures that any stray quotes are ignored instead of
being processed as extra attribute values (eventually leading to
parsing errors due to unbalanced quotes).
Similar to comments (ea8b4aa92f) and CDATA
tags (8acc7fc743) processing instructions
are now lexed in separate chunks _with_ proper support for streaming
input.
Related issue: #93
Instead of using a single token (T_CDATA) for a CDATA tag the lexer now
uses 3 tokens:
1. T_CDATA_START
2. T_CDATA_BODY
3. T_CDATA_END
The T_CDATA_BODY token can occur multiple times and is turned into a
single value in the XML parser. This is similar to the way strings are
lexed.
By changing the way CDATA tags are lexed Oga can now lex CDATA tags
containing newlines when using an IO as input. For example, this would
previously fail:
Oga.parse_xml(StringIO.new("<![CDATA[\nfoo]]>"))
Because IO input reads input per line the input for the lexer would be
as following:
"<![CDATA[\n"
"foo]]>"
Related issues: #93
This basically re-applies the technique used for HTML <script> tags. With this
extra addition I decided to rename/normalize a few things so it's easier to add
any extra tags in the future. One downside of this setup is that the following
will not be parsed by Oga:
<style>
</script>
</style>
The same applies to script tags containing a literal </style> tag. Since this
particular case is rather unlikely to occur I'm OK with not supporting it as it
_does_ simplify the lexer quite a bit.
Fixes#80
When lexing input in HTML mode the lexer has to treat _all_ content of a
<script> tag as plain text. This ensures that the lexer can process input such
as "x <y" and "// <foo>" correctly.
Fixes#70.
For JRuby this has little to no benefits as it uses strings for method names.
However, both MRI and Rubinius will perform a Symbol lookup whenever rb_intern()
is called. By doing this once for all callback names and caching the resulting
VALUE objects the lexer timings can be reduced by about 25%. In case of the
benchmark benchmark/xml/lexer/string_average_bench.rb this means it runs in
around 500ms instead of 700ms.
When lexing multi-line strings everything used to work fine as long as the input
were to be read as a whole. However, when using an IO instance all hell would
break loose. Due to the lexer reading IO instances on a per line basis,
sometimes Ragel would end up setting "ts" to NULL. For example, the following
input would break the lexer:
<foo class="\nbar" />
Due to the input being read per line, the following data would be sent to the
lexer:
<foo class="\n
bar" />
This would result in different (or NULL) pointers being used for building a
string, in turn resulting in memory allocation errors.
To work around this the string lexing setup has been broken into separate
machines for single and double quoted strings. The tokens used have also been
changed so that instead of just "T_STRING" there are now the following tokens:
* T_STRING_SQUOTE
* T_STRING_DQUOTE
* T_STRING_BODY
A string can have multiple T_STRING_BODY tokens (= multi-line strings, only the
case for IO inputs). These strings are stitched back together by the parser.
This fixes#58.
Instead of storing "act" and "cs" as an instance variable they (along with some
other variables) are now stored in a struct. This struct is attached to a lexer
instance using the (crappy) Data_Get_Struct/Data_Wrap_Struct API.
Haven't bumped into any problems just yet. However, in theory all sorts of evil
could happen here. Which is part of the problem of C: so much shit is undefined
behaviour that you can take a single step and fall in 15 holes at the same time.
In theory, because nobody bothered to actually specify it properly.
Instead of relying on String#count for counting newlines in text nodes, Oga now
does this in C/Java. String#count isn't exactly the fastest way of counting
characters. Performance was measured using
benchmark/xml/lexer/string_average_bench.rb. Before this patch the results were
as following:
MRI: 0.529s
Rbx: 4.965s
JRuby: 0.622s
After this patch:
MRI: 0.424s
Rbx: 1.942s
JRuby: 0.665s => numbers vary a bit, seem roughly the same as before
The commands used for benchmarking:
$ rake clean # to make sure that C exts aren't shared between MRI/Rbx
$ rake generate
$ rake fixtures
$ ruby benchmark/xml/lexer/string_average_bench.rb
The big difference for Rbx is probably due to the implementation of String#count
not being super fast. Some changes were made
(https://github.com/rubinius/rubinius/pull/3133) to the method, but this hasn't
been released yet.
JRuby seems to perform in a similar way, so either it was already optimizing
things for me or I suck at writing well performing Java code.
This fixes#51.
This ensures that Oga can lex the following properly:
<input value="" />
Previously Ragel would stop upon finding the empty string. This was caused due
to the string rules being declared as following:
string_dquote = (dquote ^dquote+ dquote);
string_squote = (squote ^squote+ squote);
These rules only match strings _with_ content, not without. Since Ragel stops
consuming input the moment it finds unhandled data this resulted in incorrect
tokens being emitted.
Thanks to some heavy rubberducking with @whitequark the lexer is now a little
bit better at lexing T_TEXT nodes. For example, previously the following could
not be lexed properly:
"foo < bar"
There might still be some tweaking to do but we're getting there.
Using create_makefile('liboga/liboga') will compile liboga.so into
path-to-gem/lib/liboga/ and therefore require_relative in oga.rb will fail.
Therefore the right parameter for create_makefile is 'liboga' ->
path-to-gem/lib/liboga.so
Yorick Peterse
Yorick Peterse
Laurence Lee
Andrei Botalov
Benjamin Klotz