libxml2's HTML parser adds <p> start tags in some situations. This
behavior, which doesn't follow any standard, was added in 2000, see
here: http://veillard.com/XML/messages/0655.html
Text nodes that only contain whitespace don't imply a <p> tag, but the
whitespace check cannot work reliably if we're parsing partial text data
which can happen with both pull and push parser.
The logic in `areBlanks` is hard to follow. The checks involving `CUR`
depend on the position of the input pointer and seem dubious. It's also
possible that the behavior changed inadvertently with a later commit.
As a result, it's hard to come up with good test cases.
We now process leading whitespace before creating implied tags. This is
more in line with HTML5 and should avoid at least some issues with
partial text data.
For example, parsing the string "<head> x" used to result in:
<html>
<head></head>
<body><p> x</p></body>
</html>
And now results in:
<html>
<head> </head>
<body><p>x</p></body>
</html>
Except for the implied <p> tag, this matches HTML5.
Fix a long-standing issue where QNames starting with a non-ASCII
character would be rejected. This became more visible after "streaming"
XPath evaluation was disabled since the latter handled non-ASCII names
correctly.
Fixes#818.
The Document Object Model (DOM) Level 3 Core Specification says:
> Adjacent CDATASection nodes are not merged by use of the normalize
> method of the Node interface.
Fixes#412.
The latest spec for what it essentially an XPath extension seems to be
this working draft from 2002:
https://www.w3.org/TR/xptr-xpointer/
The xpointer() scheme is listed as "being reviewed" in the XPointer
registry since at least 2006. libxml2 seems to be the only modern
software that tries to implement this spec, but the code has many bugs
and quality issues.
If you configure --with-legacy, old symbols are retained for ABI
compatibility.
This should only be done in xmlParseReference.
The handling of undeclared entities is still somewhat inconsistent. In
element content we create references even if entity substitution is
enabled. In attribute values undeclared entities are always ignored.
Also decode entities in namespace URIs if entity substitution wasn't
requested. This should fix some corner cases when comparing namespace
URIs. The Namespaces in XML 1.0 spec says:
> In a namespace declaration, the URI reference is the normalized value
> of the attribute, so replacement of XML character and entity
> references has already been done before any comparison.
Make the serialization code escape special characters in namespace URIs
like in attribute values. This fixes serialization if entities were
substituted when parsing.
Fixes https://gitlab.gnome.org/GNOME/libxslt/-/issues/106
Fixes a regression from commit e0dd330b, resulting in duplicate
attributes in the predefined XML namespace not being detected or
extraneous default attributes being passed.
Fixes#704.
Don't use a separate function to handle "complex" attributes. Validate
UTF-8 byte sequences without decoding. This should improve performance
considerably when parsing multi-byte UTF-8 sequences.
Use a string buffer to avoid unnecessary allocations and copying when
expanding entities.
Normalize attribute values in a single pass while expanding entities.
Be more lenient in recovery mode.
If no entity substitution was requested, validate entities without
expanding. Fixes#596.
Also fixes#655.
Use in ctxt->input->entity instead of ctxt->inputNr to determine whether
we are inside a parameter entity.
Stop using ctxt->external to check whether we're in an external DTD.
This is signaled by ctxt->inSubset == 2.
These were added recently in ea695ac0 and 8074b881 but were a total mess
of symbolic links and apparently mixed up files.
Symbolic links don't work on Windows.
Try to salvage one of the tests.
Pass RelaxNG structured error handler to XML parser.
Handle malloc failure from xmlRaiseError.
Remove argument from memory error handler.
Use xmlRaiseMemoryError.
Don't use xmlGenericError.
Remove TODO macro.
This regressed in commit e0dd330b.
Also fixes a long-standing issue where namespaces from default
attributes weren't added if they match an existing namespace.
Fixes#643.
Partial revert of cb927e85 fixing CRLFs not incrementing the line
number.
This requires to rework xmlParseQNameHashed. The original implementation
prompted the change to xmlCurrentChar which really shouldn't modify the
'cur' pointer as side effect. But the NEXTL macro relies on this
behavior.
Ultimately, we should reintroduce the change to xmlCurrentChar and fix
the NEXTL macro. This will lead to single CRs incrementing the line
number as well which seems more consistent.
Fixes#628.
Functions like xmlCleanSpecialAttr scan a hash table and possibly delete
entries in the callback. xmlHashScanFull must detect such deletions and
rescan the entry.
This regressed when rewriting the hash table code in 4a513d56.
Fixes#626.
