// It can be used for simple syntax parsers, like lyxrc,
// texclass and others to come. [asierra30/03/96]
//
-// (C) 1996 Lyx Team.
+// Copyright 1996 Lyx Team.
#include <config.h>
+#include <algorithm>
#include <cstdlib>
#ifdef __GNUG__
#include "lyxlex.h"
#include "debug.h"
#include "support/filetools.h"
+#include "support/lyxalgo.h"
+
+using std::ios;
+using std::lower_bound;
+using std::sort;
+
+
+// namespace {
+struct compare_tags {
+ // used by lower_bound
+ inline
+ int operator()(keyword_item const & a, char const * const tag) const {
+ return compare_no_case(a.tag, tag) < 0;
+ }
+ // used by sorted and sort
+ inline
+ int operator()(keyword_item const & a, keyword_item const & b) const {
+ return compare_no_case(a.tag, b.tag) < 0;
+ }
+};
+// } // end of anon namespace
+
LyXLex::LyXLex(keyword_item * tab, int num)
- : table(tab), no_items(num)
+ : is(&fb__), table(tab), no_items(num)
{
- file = 0;
- owns_file = false;
- status = 0;
+ status = 0;
pushed = 0;
+ // Check if the table is sorted and if not, sort it.
+ if (table && !sorted(table, table + no_items, compare_tags())) {
+ lyxerr << "The table passed to LyXLex is not sorted!!\n"
+ << "Tell the developers to fix it!" << endl;
+ // We sort it anyway to avoid problems.
+ lyxerr << "\nUnsorted:\n";
+ printTable(lyxerr);
+
+ sort(table, table + no_items,
+ compare_tags());
+ lyxerr << "\nSorted:\n";
+ printTable(lyxerr);
+ }
}
pushed = tmppu;
table = tab;
no_items = num;
+ // Check if the table is sorted and if not, sort it.
+ if (table && !sorted(table, table + no_items, compare_tags())) {
+ lyxerr << "The table passed to LyXLex is not sorted!!\n"
+ << "Tell the developers to fix it!" << endl;
+ // We sort it anyway to avoid problems.
+ lyxerr << "\nUnsorted:\n";
+ printTable(lyxerr);
+
+ sort(table, table + no_items, compare_tags());
+ lyxerr << "\nSorted:\n";
+ printTable(lyxerr);
+ }
}
}
-void LyXLex::printTable()
+void LyXLex::printTable(ostream & os)
{
- lyxerr << "\nNumber of tags: " << no_items << endl;
+ os << "\nNumber of tags: " << no_items << '\n';
for(int i= 0; i < no_items; ++i)
- lyxerr << "table[" << i
- << "]: tag: `" << table[i].tag
- << "' code:" << table[i].code << endl;
- lyxerr << endl;
+ os << "table[" << i
+ << "]: tag: `" << table[i].tag
+ << "' code:" << table[i].code << '\n';
+ os.flush();
}
bool LyXLex::setFile(string const & filename)
{
- if (file)
- lyxerr << "Error in LyXLex::setFile: file already set." <<endl;
- file = fopen(filename.c_str(), "r");
+ if (fb__.is_open())
+ lyxerr << "Error in LyXLex::setFile: "
+ "file or stream already set." << endl;
+ fb__.open(filename.c_str(), ios::in);
+ is.rdbuf(&fb__);
name = filename;
- owns_file = true;
lineno = 0;
- return (file ? true : false);
+ return fb__.is_open() && is.good();
}
-void LyXLex::setFile(FILE * f)
+void LyXLex::setStream(istream & i)
{
- if (file)
- lyxerr << "Error in LyXLex::setFile: file already set."
- << endl;
- file = f;
- owns_file = false;
- lineno = 0; // this is bogus if the file already has been read from
+ if (fb__.is_open() || is.rdbuf()->in_avail())
+ lyxerr << "Error in LyXLex::setStream: "
+ "file or stream already set." << endl;
+ is.rdbuf(i.rdbuf());
+ lineno = 0;
}
int LyXLex::lex()
{
//NOTE: possible bug.
- if (next() && status == LEX_TOKEN)
- return search_kw(buff);
- else
- return status;
+ if (next() && status == LEX_TOKEN)
+ return search_kw(buff);
+ else
+ return status;
}
int LyXLex::GetInteger() const
{
- if (buff[0] > ' ')
- return atoi(buff);
- else {
- printError("Bad integer `$$Token'");
- return -1;
- }
+ if (buff[0] > ' ')
+ return atoi(buff);
+ else {
+ printError("Bad integer `$$Token'");
+ return -1;
+ }
}
bool LyXLex::GetBool() const
{
- if (compare(buff, "true") == 0)
- return true;
- else if (compare(buff, "false") != 0)
- printError("Bad boolean `$$Token'. Use \"false\" or \"true\"");
- return false;
+ if (compare(buff, "true") == 0)
+ return true;
+ else if (compare(buff, "false") != 0)
+ printError("Bad boolean `$$Token'. Use \"false\" or \"true\"");
+ return false;
}
bool LyXLex::EatLine()
{
int i = 0;
- int c = '\0'; // getc() returns an int
-
- while (!feof(file) && c!= '\n' && i!= (LEX_MAX_BUFF-1)) {
- c = getc(file);
+ unsigned char c = '\0';
+ char cc = 0;
+ while(is && c != '\n' && i != (LEX_MAX_BUFF - 1)) {
+ is.get(cc);
+ c = cc;
+ lyxerr[Debug::LYXLEX] << "LyXLex::EatLine read char: `"
+ << c << "'" << endl;
if (c != '\r')
buff[i++] = c;
}
- if (i == (LEX_MAX_BUFF-1) && c != '\n') {
- printError("Line too long");
+ if (i == (LEX_MAX_BUFF - 1) && c != '\n') {
+ printError("Line too long");
c = '\n'; // Pretend we had an end of line
--lineno; // but don't increase line counter (netto effect)
++i; // and preserve last character read.
int LyXLex::search_kw(char const * const tag) const
{
- int m, k = 0 , l = 0, r = no_items;
-
- while (l < r) {
- m = (l + r) / 2;
-
- if (lyxerr.debugging(Debug::PARSER)) {
- lyxerr << "LyXLex::search_kw: elem " << m
- << " tag " << table[m].tag
- << " search tag " << tag
- << endl;
- }
-
- if (table[m].tag)
- k = compare_no_case(table[m].tag, tag);
- if (k == 0)
- return table[m].code;
- else
- if (k < 0) l = m + 1; else r = m;
- }
+ keyword_item * res =
+ lower_bound(table, table + no_items, tag, compare_tags());
+ if (res != table + no_items && !compare_no_case(res->tag, tag))
+ return res->code;
return LEX_UNDEF;
}
bool LyXLex::next(bool esc)
{
if (!esc) {
- int c; // getc() returns an int
-
+ unsigned char c = 0; // getc() returns an int
+ char cc = 0;
status = 0;
- while (!feof(file) && !status) {
- c = getc(file);
+ while (is && !status) {
+ is.get(cc);
+ c = cc;
if (c == '#') {
// Read rest of line (fast :-)
- fgets(buff, sizeof(buff), file);
+ is.getline(buff, sizeof(buff));
+ lyxerr[Debug::LYXLEX] << "Comment read: `" << c
+ << buff << "'" << endl;
++lineno;
continue;
}
if (c == '\"') {
int i = -1;
do {
- c = getc(file);
+ is.get(cc);
+ c = cc;
if (c != '\r')
buff[++i] = c;
- } while (c!= '\"' && c!= '\n' && !feof(file) &&
- i!= (LEX_MAX_BUFF-2));
+ } while (c != '\"' && c != '\n' && is &&
+ i != (LEX_MAX_BUFF - 2));
- if (i == (LEX_MAX_BUFF-2)) {
+ if (i == (LEX_MAX_BUFF - 2)) {
printError("Line too long");
c = '\"'; // Pretend we got a "
++i;
}
- if (c!= '\"') {
+ if (c != '\"') {
printError("Missing quote");
if (c == '\n')
++lineno;
if (c == ',')
continue; /* Skip ','s */
-
- if (c > ' ' && !feof(file)) {
+
+ // using relational operators with chars other
+ // than == and != is not safe. And if it is done
+ // the type _have_ to be unsigned. It usually a
+ // lot better to use the functions from cctype
+ if (c > ' ' && is) {
int i = 0;
do {
buff[i++] = c;
- c = getc(file);
- } while (c > ' ' && c != ',' && !feof(file) &&
- (i != LEX_MAX_BUFF-1) );
- if (i == LEX_MAX_BUFF-1) {
+ is.get(cc);
+ c = cc;
+ } while (c > ' ' && c != ',' && is
+ && (i != LEX_MAX_BUFF - 1) );
+ if (i == LEX_MAX_BUFF - 1) {
printError("Line too long");
}
buff[i] = '\0';
status = LEX_TOKEN;
}
- if (c == '\r' && !feof(file)) {
+ if (c == '\r' && is) {
// The Windows support has lead to the
// possibility of "\r\n" at the end of
// a line. This will stop LyX choking
// when it expected to find a '\n'
- c = getc(file);
+ is.get(cc);
+ c = cc;
}
if (c == '\n')
}
if (status) return true;
- status = (feof(file)) ? LEX_FEOF: LEX_UNDEF;
+ status = is.eof() ? LEX_FEOF: LEX_UNDEF;
buff[0] = '\0';
return false;
} else {
- int c; // getc() returns an int
+ unsigned char c = 0; // getc() returns an int
+ char cc = 0;
status = 0;
- while (!feof(file) && !status) {
- c = getc(file);
+ while (is && !status) {
+ is.get(cc);
+ c = cc;
// skip ','s
if (c == ',') continue;
do {
if (c == '\\') {
// escape the next char
- c = getc(file);
+ is.get(cc);
+ c = cc;
}
buff[i++] = c;
- c = getc(file);
- } while (c > ' ' && c != ',' && !feof(file) &&
- (i != LEX_MAX_BUFF-1) );
- if (i == LEX_MAX_BUFF-1) {
+ is.get(cc);
+ c = cc;
+ } while (c > ' ' && c != ',' && is
+ && (i != LEX_MAX_BUFF - 1) );
+ if (i == LEX_MAX_BUFF - 1) {
printError("Line too long");
}
buff[i] = '\0';
if (c == '#') {
// Read rest of line (fast :-)
- fgets(buff, sizeof(buff), file);
+ is.getline(buff, sizeof(buff));
+ lyxerr[Debug::LYXLEX] << "Comment read: `" << c
+ << buff << "'" << endl;
++lineno;
continue;
}
bool escaped = false;
do {
escaped = false;
- c = getc(file);
+ is.get(cc);
+ c = cc;
if (c == '\r') continue;
if (c == '\\') {
// escape the next char
- c = getc(file);
+ is.get(cc);
+ c = cc;
escaped = true;
}
buff[++i] = c;
if (!escaped && c == '\"') break;
- } while (c!= '\n' && !feof(file) &&
- i!= (LEX_MAX_BUFF-2));
+ } while (c != '\n' && is &&
+ i != (LEX_MAX_BUFF - 2));
- if (i == (LEX_MAX_BUFF-2)) {
+ if (i == (LEX_MAX_BUFF - 2)) {
printError("Line too long");
c = '\"'; // Pretend we got a "
++i;
}
- if (c!= '\"') {
+ if (c != '\"') {
printError("Missing quote");
if (c == '\n')
++lineno;
break;
}
- if (c > ' ' && !feof(file)) {
+ if (c > ' ' && is) {
int i = 0;
do {
if (c == '\\') {
// escape the next char
- c = getc(file);
+ is.get(cc);
+ c = cc;
//escaped = true;
}
buff[i++] = c;
- c = getc(file);
- } while (c > ' ' && c != ',' && !feof(file) &&
- (i != LEX_MAX_BUFF-1) );
+ is.get(cc);
+ c = cc;
+ } while (c > ' ' && c != ',' && is
+ && (i != LEX_MAX_BUFF-1) );
if (i == LEX_MAX_BUFF-1) {
printError("Line too long");
}
buff[i] = '\0';
status = LEX_TOKEN;
}
-
// new line
if (c == '\n')
++lineno;
if (status) return true;
- status = (feof(file)) ? LEX_FEOF: LEX_UNDEF;
+ status = is.eof() ? LEX_FEOF : LEX_UNDEF;
buff[0] = '\0';
- return false;
+ return false;
}
}
bool LyXLex::nextToken()
{
status = 0;
- while (!feof(file) && !status) {
- int c = getc(file); // getc() returns an int
-
- if (c >= ' ' && !feof(file)) {
+ while (is && !status) {
+ unsigned char c = 0;
+ char cc = 0;
+ is.get(cc);
+ c = cc;
+ if (c >= ' ' && is) {
int i = 0;
if (c == '\\') { // first char == '\\'
do {
buff[i++] = c;
- c = getc(file);
- } while (c > ' ' && c != '\\' && !feof(file) &&
- i != (LEX_MAX_BUFF-1));
+ is.get(cc);
+ c = cc;
+ } while (c > ' ' && c != '\\' && is
+ && i != (LEX_MAX_BUFF-1));
} else {
do {
buff[i++] = c;
- c = getc(file);
- } while (c >= ' ' && c != '\\' && !feof(file)
+ is.get(cc);
+ c = cc;
+ } while (c >= ' ' && c != '\\' && is
&& i != (LEX_MAX_BUFF-1));
}
- if (i == (LEX_MAX_BUFF-1)) {
+ if (i == (LEX_MAX_BUFF - 1)) {
printError("Line too long");
}
- if (c == '\\') ungetc(c, file); // put it back
+ if (c == '\\') is.putback(c); // put it back
buff[i] = '\0';
status = LEX_TOKEN;
}
}
if (status) return true;
- status = (feof(file)) ? LEX_FEOF: LEX_UNDEF;
+ status = is.eof() ? LEX_FEOF: LEX_UNDEF;
buff[0] = '\0';
return false;
}