#include "support/lstrings.h"
#include "support/textutils.h"
+#include <algorithm>
#include <atomic>
#include <map>
#include <functional>
}
-// escape what needs escaping
-docstring xmlize(docstring const &str, XMLStream::EscapeSettings e)
-{
- return xml::escapeString(str, e);
-}
-
-
docstring cleanAttr(docstring const & str)
{
docstring newname;
{
docstring output = '<' + tag_;
if (!attr_.empty()) {
- docstring attributes = xml::xmlize(attr_, XMLStream::ESCAPE_NONE);
- attributes.erase(attributes.begin(), std::find_if(attributes.begin(), attributes.end(),
- [](int c) {return !std::isspace(c);}));
- if (!attributes.empty()) {
+ docstring attributes = xml::trimLeft(xml::escapeString(attr_, XMLStream::ESCAPE_NONE));
+ if (!attributes.empty())
output += ' ' + attributes;
- }
}
output += ">";
return output;
docstring CompTag::writeTag() const
{
- docstring output = '<' + from_utf8(tag_);
+ docstring output = '<' + tag_;
if (!attr_.empty()) {
// Erase the beginning of the attributes if it contains space characters: this function deals with that
// automatically.
- docstring attributes = xmlize(from_utf8(attr_), XMLStream::ESCAPE_NONE);
+ docstring attributes = escapeString(attr_, XMLStream::ESCAPE_NONE);
attributes.erase(attributes.begin(), std::find_if(attributes.begin(), attributes.end(),
[](int c) {return !std::isspace(c);}));
if (!attributes.empty()) {
} // namespace xml
-void XMLStream::writeError(std::string const &s) const
+void XMLStream::writeError(std::string const &s)
{
LYXERR0(s);
- os_ << from_utf8("<!-- Output Error: " + s + " -->\n");
+ *this << ESCAPE_NONE << from_utf8("<!-- Output Error: " + s + " -->");
+ *this << xml::CR();
}
-void XMLStream::writeError(docstring const &s) const
+void XMLStream::writeError(docstring const &s)
{
LYXERR0(s);
- os_ << from_utf8("<!-- Output Error: ") << s << from_utf8(" -->\n");
+ *this << ESCAPE_NONE << from_utf8("<!-- Output Error: ");
+ *this << s;
+ *this << ESCAPE_NONE << from_utf8(" -->");
+ *this << xml::CR();
+}
+
+
+XMLStream::TagPtr XMLStream::getLastStackTag()
+{
+ return tag_stack_.back();
}
if (**curtag != xml::parsep_tag)
os_ << (*curtag)->writeEndTag();
tag_stack_.pop_back();
- // this shouldn't happen, since then the font tags
- // weren't in any other tag.
-// LASSERT(!tag_stack_.empty(), return true);
if (tag_stack_.empty())
return true;
curtag = &tag_stack_.back();
XMLStream &XMLStream::operator<<(docstring const &d)
{
+ is_last_tag_cr_ = false;
clearTagDeque();
- os_ << xml::xmlize(d, escape_);
+ os_ << xml::escapeString(d, escape_);
escape_ = ESCAPE_ALL;
return *this;
}
XMLStream &XMLStream::operator<<(const char *s)
{
+ is_last_tag_cr_ = false;
clearTagDeque();
docstring const d = from_ascii(s);
- os_ << xml::xmlize(d, escape_);
+ os_ << xml::escapeString(d, escape_);
escape_ = ESCAPE_ALL;
return *this;
}
XMLStream &XMLStream::operator<<(char_type c)
{
+ is_last_tag_cr_ = false;
clearTagDeque();
os_ << xml::escapeChar(c, escape_);
escape_ = ESCAPE_ALL;
XMLStream &XMLStream::operator<<(char c)
{
+ is_last_tag_cr_ = false;
clearTagDeque();
os_ << xml::escapeChar(c, escape_);
escape_ = ESCAPE_ALL;
XMLStream &XMLStream::operator<<(int i)
{
+ is_last_tag_cr_ = false;
clearTagDeque();
os_ << i;
escape_ = ESCAPE_ALL;
XMLStream &XMLStream::operator<<(EscapeSettings e)
{
+ // Don't update is_last_tag_cr_ here, as this does not output anything.
escape_ = e;
return *this;
}
XMLStream &XMLStream::operator<<(xml::StartTag const &tag)
{
+ is_last_tag_cr_ = false;
if (tag.tag_.empty())
return *this;
pending_tags_.push_back(makeTagPtr(tag));
XMLStream &XMLStream::operator<<(xml::ParTag const &tag)
{
+ is_last_tag_cr_ = false;
if (tag.tag_.empty())
return *this;
pending_tags_.push_back(makeTagPtr(tag));
XMLStream &XMLStream::operator<<(xml::CompTag const &tag)
{
+ is_last_tag_cr_ = false;
if (tag.tag_.empty())
return *this;
clearTagDeque();
XMLStream &XMLStream::operator<<(xml::FontTag const &tag)
{
+ is_last_tag_cr_ = false;
if (tag.tag_.empty())
return *this;
pending_tags_.push_back(makeTagPtr(tag));
XMLStream &XMLStream::operator<<(xml::CR const &)
{
+ is_last_tag_cr_ = true;
clearTagDeque();
os_ << from_ascii("\n");
return *this;
// best to make things work.
XMLStream &XMLStream::operator<<(xml::EndTag const &etag)
{
+ is_last_tag_cr_ = false;
+
if (etag.tag_.empty())
return *this;
+ "' when other tags were pending. Last pending tag is `"
+ to_utf8(pending_tags_.back()->writeTag())
+ "'. Tag discarded.");
- pending_tags_.erase(dit);
+ if (!pending_tags_.empty())
+ pending_tags_.erase(dit);
return *this;
}
}
string estr = "Closing tag `" + to_utf8(etag.tag_)
+ "' when other tags are pending. Discarded pending tags:\n";
for (dit = pending_tags_.begin(); dit != den; ++dit)
- estr += to_utf8(xml::xmlize((*dit)->writeTag(), XMLStream::ESCAPE_ALL)) + "\n";
+ estr += to_utf8(xml::escapeString((*dit)->writeTag(), XMLStream::ESCAPE_ALL)) + "\n";
writeError(estr);
// clear the pending tags...
pending_tags_.clear();
}
+bool xml::isNotOnlySpace(docstring const & str)
+{
+ for (auto const & c: str) {
+ if (c != ' ' && c != '\t' && c != '\n' && c != '\v' && c != '\f' && c != '\r')
+ return true;
+ }
+ return false;
+}
+
+
+docstring xml::trimLeft(docstring const & str)
+{
+ size_t i = 0;
+ for (auto const & c: str) {
+ if (c != ' ' && c != '\t' && c != '\n' && c != '\v' && c != '\f' && c != '\r')
+ return str.substr(i, docstring::npos);
+ i++;
+ }
+ return str;
+}
+
+
docstring xml::cleanID(docstring const & orig)
{
- // The standard xml:id only allows letters,
- // digits, '-' and '.' in a name.
- // This routine replaces illegal characters by '-' or '.'
- // and adds a number for uniqueness if need be.
- docstring const allowed = from_ascii(".-_");
+ // The standard xml:id only allows letters, digits, '-' and '.' in a name.
+ // This routine replaces illegal characters by '-' or '.' and adds a number for uniqueness if need be.
// Use a cache of already mangled names: the alterations may merge several IDs as one. This ensures that the IDs
// are not mixed up in the document.
+ // This code could be improved: it uses Qt outside the GUI part. Any TLS implementation could do the trick.
typedef map<docstring, docstring> MangledMap;
static QThreadStorage<MangledMap> tMangledNames;
static QThreadStorage<int> tMangleID;
- MangledMap & mangledNames = tMangledNames.localData();
-
// If the name is already known, just return it.
- MangledMap::const_iterator const known = mangledNames.find(orig);
+ MangledMap & mangledNames = tMangledNames.localData();
+ auto const known = mangledNames.find(orig);
if (known != mangledNames.end())
return known->second;
// Start creating the mangled name by iterating over the characters.
docstring content;
- docstring::const_iterator it = orig.begin();
- docstring::const_iterator end = orig.end();
+ auto it = orig.cbegin();
+ auto end = orig.cend();
// Make sure it starts with a letter.
- if (!isAlphaASCII(*it) && allowed.find(*it) >= allowed.size())
+ if (!isAlphaASCII(*it))
content += "x";
- // Do the mangling.
+ // Parse the ID character by character and change what needs to.
bool mangle = false; // Indicates whether the ID had to be changed, i.e. if ID no more ensured to be unique.
for (; it != end; ++it) {
char_type c = *it;
- if (isAlphaASCII(c) || isDigitASCII(c) || c == '-' || c == '.'
- || allowed.find(c) < allowed.size())
+ if (isAlphaASCII(c) || isDigitASCII(c) || c == '-' || c == '.' || c == '_') {
content += c;
- else if (c == '_' || c == ' ') {
- mangle = true;
- content += "-";
- }
- else if (c == ':' || c == ',' || c == ';' || c == '!') {
+ } else if (c == ':' || c == ',' || c == ';' || c == '!') {
mangle = true;
content += ".";
- }
- else {
+ } else { // Other invalid characters, such as ' '.
mangle = true;
content += "-";
}
}
- if (mangle) {
+ // If there had to be a change, check if ID unicity is still guaranteed.
+ // This avoids having a clash if satisfying XML requirements for ID makes two IDs identical, like "a:b" and "a!b",
+ // as both of them would be transformed as "a.b". With this procedure, one will become "a.b" and the other "a.b-1".
+ if (mangle && mangledNames.find(content) != mangledNames.end()) {
int & mangleID = tMangleID.localData();
- content += "-" + convert<docstring>(mangleID++);
+ if (mangleID > 0)
+ content += "-" + convert<docstring>(mangleID);
+ mangleID += 1;
}
+ // Save the new ID to avoid recomputing it afterwards and to ensure stability over the document.
mangledNames[orig] = content;
-
return content;
}
projects / lyx.git / blobdiff