CLI2: Integrated Lexer::decomposePair

This commit is contained in:
Paul Beckingham
2015-07-08 09:48:15 -04:00
parent e6c4f48a48
commit 5a21945a0d

View File

@@ -1031,163 +1031,147 @@ void CLI2::desugarFilterAttributes ()
if (a._lextype == Lexer::Type::pair && if (a._lextype == Lexer::Type::pair &&
a.hasTag ("FILTER")) a.hasTag ("FILTER"))
{ {
changes = true; std::string raw = a.attribute ("raw");
std::string name;
auto raw = a.attribute ("raw"); std::string mod;
auto dot = raw.find ('.'); std::string sep;
auto colon = raw.find (':'); std::string value;
if (colon == std::string::npos) if (Lexer::decomposePair (raw, name, mod, sep, value))
colon = raw.find ('=');
std::string name = "";
std::string mod = "";
std::string value = "";
// If the dot appears after the colon, then it is part of the value, and
// should be ignored.
if (dot != std::string::npos &&
dot < colon)
{ {
name = raw.substr (0, dot); if (value == "")
mod = raw.substr (dot + 1, colon - dot - 1); value = "''";
}
else
{
name = raw.substr (0, colon);
}
value = raw.substr (colon + 1); bool found = false;
if (value == "") std::string canonical;
value = "''"; if (canonicalize (canonical, "pseudo", name))
{
A2 lhs (raw, Lexer::Type::identifier);
lhs.attribute ("canonical", canonical);
lhs.attribute ("value", value);
lhs.tag ("PSEUDO");
reconstructed.push_back (lhs);
found = true;
}
else if (canonicalize (canonical, "attribute", name) ||
canonicalize (canonical, "uda", name))
{
// TODO The "!" modifier is being dropped.
bool found = false; A2 lhs (name, Lexer::Type::dom);
std::string canonical; lhs.tag ("FILTER");
if (canonicalize (canonical, "pseudo", name)) lhs.attribute ("canonical", canonical);
{ lhs.attribute ("modifier", mod);
A2 lhs (raw, Lexer::Type::identifier);
lhs.attribute ("canonical", canonical); A2 op ("", Lexer::Type::op);
lhs.attribute ("value", value); op.tag ("FILTER");
lhs.tag ("PSEUDO");
reconstructed.push_back (lhs); A2 rhs ("", Lexer::Type::string);
found = true; rhs.tag ("FILTER");
}
else // Special case for '<name>:<value>'.
{ if (mod == "")
{
op.attribute ("raw", "=");
rhs.attribute ("raw", value);
}
else if (mod == "before" || mod == "under" || mod == "below")
{
op.attribute ("raw", "<");
rhs.attribute ("raw", value);
}
else if (mod == "after" || mod == "over" || mod == "above")
{
op.attribute ("raw", ">");
rhs.attribute ("raw", value);
}
else if (mod == "none")
{
op.attribute ("raw", "==");
rhs.attribute ("raw", "''");
}
else if (mod == "any")
{
op.attribute ("raw", "!==");
rhs.attribute ("raw", "''");
}
else if (mod == "is" || mod == "equals")
{
op.attribute ("raw", "==");
rhs.attribute ("raw", value);
}
else if (mod == "isnt" || mod == "not")
{
op.attribute ("raw", "!==");
rhs.attribute ("raw", value);
}
else if (mod == "has" || mod == "contains")
{
op.attribute ("raw", "~");
rhs.attribute ("raw", value);
}
else if (mod == "hasnt")
{
op.attribute ("raw", "!~");
rhs.attribute ("raw", value);
}
else if (mod == "startswith" || mod == "left")
{
op.attribute ("raw", "~");
rhs.attribute ("raw", "^" + value);
}
else if (mod == "endswith" || mod == "right")
{
op.attribute ("raw", "~");
rhs.attribute ("raw", value + "$");
}
else if (mod == "word")
{
op.attribute ("raw", "~");
#if defined (DARWIN)
rhs.attribute ("raw", value);
#elif defined (SOLARIS)
rhs.attribute ("raw", "\\<" + value + "\\>");
#else
rhs.attribute ("raw", "\\b" + value + "\\b");
#endif
}
else if (mod == "noword")
{
op.attribute ("raw", "!~");
#if defined (DARWIN)
rhs.attribute ("raw", value);
#elif defined (SOLARIS)
rhs.attribute ("raw", "\\<" + value + "\\>");
#else
rhs.attribute ("raw", "\\b" + value + "\\b");
#endif
}
else
throw format (STRING_PARSER_UNKNOWN_ATTMOD, mod);
reconstructed.push_back (lhs);
reconstructed.push_back (op);
reconstructed.push_back (rhs);
found = true;
}
// If the name does not canonicalize to either an attribute or a UDA // If the name does not canonicalize to either an attribute or a UDA
// then it is not a recognized Lexer::Type::pair, so downgrade it to // then it is not a recognized Lexer::Type::pair, so downgrade it to
// Lexer::Type::word. // Lexer::Type::word.
if (! canonicalize (canonical, "attribute", name) && else
! canonicalize (canonical, "uda", name))
{ {
a._lextype = Lexer::Type::word; a._lextype = Lexer::Type::word;
continue;
} }
// TODO The "!" modifier is being dropped. if (found)
changes = true;
A2 lhs (name, Lexer::Type::dom);
lhs.tag ("FILTER");
lhs.attribute ("canonical", canonical);
lhs.attribute ("modifier", mod);
A2 op ("", Lexer::Type::op);
op.tag ("FILTER");
A2 rhs ("", Lexer::Type::string);
rhs.tag ("FILTER");
// Special case for '<name>:<value>'.
if (mod == "")
{
op.attribute ("raw", "=");
rhs.attribute ("raw", value);
}
else if (mod == "before" || mod == "under" || mod == "below")
{
op.attribute ("raw", "<");
rhs.attribute ("raw", value);
}
else if (mod == "after" || mod == "over" || mod == "above")
{
op.attribute ("raw", ">");
rhs.attribute ("raw", value);
}
else if (mod == "none")
{
op.attribute ("raw", "==");
rhs.attribute ("raw", "''");
}
else if (mod == "any")
{
op.attribute ("raw", "!==");
rhs.attribute ("raw", "''");
}
else if (mod == "is" || mod == "equals")
{
op.attribute ("raw", "==");
rhs.attribute ("raw", value);
}
else if (mod == "isnt" || mod == "not")
{
op.attribute ("raw", "!==");
rhs.attribute ("raw", value);
}
else if (mod == "has" || mod == "contains")
{
op.attribute ("raw", "~");
rhs.attribute ("raw", value);
}
else if (mod == "hasnt")
{
op.attribute ("raw", "!~");
rhs.attribute ("raw", value);
}
else if (mod == "startswith" || mod == "left")
{
op.attribute ("raw", "~");
rhs.attribute ("raw", "^" + value);
}
else if (mod == "endswith" || mod == "right")
{
op.attribute ("raw", "~");
rhs.attribute ("raw", value + "$");
}
else if (mod == "word")
{
op.attribute ("raw", "~");
#if defined (DARWIN)
rhs.attribute ("raw", value);
#elif defined (SOLARIS)
rhs.attribute ("raw", "\\<" + value + "\\>");
#else
rhs.attribute ("raw", "\\b" + value + "\\b");
#endif
}
else if (mod == "noword")
{
op.attribute ("raw", "!~");
#if defined (DARWIN)
rhs.attribute ("raw", value);
#elif defined (SOLARIS)
rhs.attribute ("raw", "\\<" + value + "\\>");
#else
rhs.attribute ("raw", "\\b" + value + "\\b");
#endif
}
else else
throw format (STRING_PARSER_UNKNOWN_ATTMOD, mod); reconstructed.push_back (a);
reconstructed.push_back (lhs);
reconstructed.push_back (op);
reconstructed.push_back (rhs);
found = true;
} }
// Failed to decompose.
if (found)
changes = true;
else else
reconstructed.push_back (a); reconstructed.push_back (a);
} }
// Not a FILTER pair.
else else
reconstructed.push_back (a); reconstructed.push_back (a);
} }