/*!
* Replaces up to ONE quoted string inside a string starting at idx_start.
*
* \param strline string to be processed
* \param idx_start index of line character to start search
* \param contd specifies the quote string is continued from the previous line
* \param CurrentQuoteEnd end quote character of the current status
*
* \return method status
*/
int CCodeCounter::ReplaceQuote(string &strline, size_t &idx_start, bool &contd, char &CurrentQuoteEnd)
{
size_t idx_end, idx_quote;
if (contd)
{
idx_start = 0;
if (strline[0] == CurrentQuoteEnd)
{
idx_start = 1;
contd = false;
return 1;
}
strline[0] = '$';
}
else
{
// handle two quote chars in some languages, both " and ' may be accepted
idx_start = FindQuote(strline, QuoteStart, idx_start, QuoteEscapeFront);
if (idx_start != string::npos)
{
idx_quote = QuoteStart.find_first_of(strline[idx_start]);
CurrentQuoteEnd = QuoteEnd[idx_quote];
}
else
{
idx_start = strline.length();
return 0;
}
}
idx_end = CUtil::FindCharAvoidEscape(strline, CurrentQuoteEnd, idx_start + 1, QuoteEscapeFront);
if (idx_end == string::npos)
{
idx_end = strline.length() - 1;
strline.replace(idx_start + 1, idx_end - idx_start, idx_end - idx_start, '$');
contd = true;
idx_start = idx_end + 1;
}
else
{
if ((QuoteEscapeRear) && (strline.length() > idx_end + 1) && (strline[idx_end+1] == QuoteEscapeRear))
{
strline[idx_end] = '$';
strline[idx_end+1] = '$';
}
else
{
contd = false;
strline.replace(idx_start + 1, idx_end - idx_start - 1, idx_end - idx_start - 1, '$');
idx_start = idx_end + 1;
}
}
return 1;
}
/*!
* Counts the number of comment lines, removes comments, and
* replaces quoted strings by special chars, e.g., $
* All arguments are modified by the method.
* Special processing for """ and ''' which can be multi-line literal
* or a multi-line comment if it stands alone.
*
* \param fmap list of processed file lines
* \param result counter results
* \param fmapBak list of original file lines (same as fmap except it contains unmodified quoted strings)
*
* \return method status
*/
int CPythonCounter::CountCommentsSLOC(filemap* fmap, results* result, filemap *fmapBak)
{
if (BlockCommentStart.empty() && LineCommentStart.empty())
return 0;
if (classtype == UNKNOWN || classtype == DATAFILE)
return 0;
bool contd = false;
bool contd_nextline;
int comment_type = 0;
/*
comment_type:
0 : not a comment
1 : line comment, whole line
2 : line comment, embedded
3 : block comment, undecided
4 : block comment, embedded
*/
size_t idx_start, idx_end, comment_start;
size_t quote_idx_start;
string curBlckCmtStart, curBlckCmtEnd, tmp;
string CurrentQuoteEnd = "";
bool quote_contd = false;
filemap::iterator itfmBak = fmapBak->begin();
quote_idx_start = 0;
for (filemap::iterator iter = fmap->begin(); iter != fmap->end(); iter++, itfmBak++)
{
contd_nextline = false;
quote_idx_start = 0;
idx_start = 0;
if (CUtil::CheckBlank(iter->line))
continue;
if (quote_contd)
{
// Replace quote until next character
ReplaceQuote(iter->line, quote_idx_start, quote_contd, CurrentQuoteEnd);
if (quote_contd)
continue;
}
if (contd)
comment_type = 3;
while (!contd_nextline && idx_start < iter->line.length())
{
// need to handle multiple quote chars in some languages, both " and ' may be accepted
quote_idx_start = FindQuote(iter->line, QuoteStart, quote_idx_start, QuoteEscapeFront);
comment_start = idx_start;
if (!contd)
{
FindCommentStart(iter->line, comment_start, comment_type, curBlckCmtStart, curBlckCmtEnd);
if (comment_start != string::npos && comment_type > 2)
{
// python: check whether this is a multi-line literal or a block comment
tmp = CUtil::TrimString(iter->line, -1);
if (iter->line.length() - tmp.length() != comment_start)
{
quote_idx_start = comment_start;
comment_start = string::npos;
}
}
}
if (comment_start == string::npos && quote_idx_start == string::npos)
break;
if (comment_start != string::npos)
idx_start = comment_start;
// if found quote before comment, e.g., "this is quote");//comment
if (quote_idx_start != string::npos && (comment_start == string::npos || quote_idx_start < comment_start))
{
ReplaceQuote(iter->line, quote_idx_start, quote_contd, CurrentQuoteEnd);
if (quote_idx_start > idx_start && quote_idx_start != iter->line.length())
{
// comment delimiter inside quote
idx_start = quote_idx_start;
continue;
}
}
else if (comment_start != string::npos)
{
// comment delimiter starts first
switch (comment_type)
{
case 1: // line comment, definitely whole line
iter->line = "";
itfmBak->line = "";
result->comment_lines++;
contd_nextline = true;
break;
case 2: // line comment, possibly embedded
iter->line = iter->line.substr(0, idx_start);
itfmBak->line = itfmBak->line.substr(0, idx_start);
// trim trailing space
iter->line = CUtil::TrimString(iter->line, 1);
itfmBak->line = CUtil::TrimString(itfmBak->line, 1);
if (iter->line.empty())
result->comment_lines++; // whole line
else
result->e_comm_lines++; // embedded
contd_nextline = true;
break;
case 3: // block comment
case 4:
if (contd)
idx_end = iter->line.find(curBlckCmtEnd);
else
idx_end = iter->line.find(curBlckCmtEnd, idx_start + curBlckCmtStart.length());
if (idx_end == string::npos)
{
if (comment_type == 3)
{
iter->line = "";
itfmBak->line = "";
result->comment_lines++;
}
else if (comment_type == 4)
{
iter->line = iter->line.substr(0, idx_start);
itfmBak->line = itfmBak->line.substr(0, idx_start);
// trim trailing space
iter->line = CUtil::TrimString(iter->line, 1);
itfmBak->line = CUtil::TrimString(itfmBak->line, 1);
if (iter->line.empty())
result->comment_lines++; // whole line
else
result->e_comm_lines++; // embedded
}
contd = true;
contd_nextline = true;
break;
}
else
{
contd = false;
iter->line.erase(idx_start, idx_end - idx_start + curBlckCmtEnd.length());
itfmBak->line.erase(idx_start, idx_end - idx_start + curBlckCmtEnd.length());
if (iter->line.empty())
result->comment_lines++;
else
{
// trim trailing space
iter->line = CUtil::TrimString(iter->line, 1);
itfmBak->line = CUtil::TrimString(itfmBak->line, 1);
if (iter->line.empty())
result->comment_lines++; // whole line
else
result->e_comm_lines++; // embedded
}
// quote chars found may be erased as it is inside comment
quote_idx_start = idx_start;
}
break;
default:
cout << "Error in CountCommentsSLOC()" << endl;
break;
}
}
}
}
return 1;
}
/*!
* Replaces up to ONE quoted string inside a string starting at idx_start.
* Uses a string instead of a character to allow processing multi-line
* literals """ and '''.
*
* \param strline string to be processed
* \param idx_start index of line character to start search
* \param contd specifies the quote string is continued from the previous line
* \param CurrentQuoteEnd end quote string of the current status
*
* \return method status
*/
int CPythonCounter::ReplaceQuote(string &strline, size_t &idx_start, bool &contd, string &CurrentQuoteEnd)
{
size_t idx_end, idx_quote;
if (contd)
{
// python: use string instead of character to check for """ and '''
idx_start = 0;
if (strline.length() >= CurrentQuoteEnd.length() &&
strline.substr(0, CurrentQuoteEnd.length()) == CurrentQuoteEnd)
{
idx_start = CurrentQuoteEnd.length();
contd = false;
return 1;
}
strline[0] = '$';
}
else
{
// handle two quote chars in some languages, both " and ' may be accepted
idx_start = FindQuote(strline, QuoteStart, idx_start, QuoteEscapeFront);
if (idx_start != string::npos)
{
idx_quote = QuoteStart.find_first_of(strline[idx_start]);
CurrentQuoteEnd = QuoteEnd[idx_quote];
// python: check for """ or '''
if (strline.length() >= idx_start + 3)
{
if (CurrentQuoteEnd == "\"")
{
if (strline.substr(idx_start, 3) == "\"\"\"")
CurrentQuoteEnd = "\"\"\"";
}
else if (CurrentQuoteEnd == "'")
{
if (strline.substr(idx_start, 3) == "'''")
CurrentQuoteEnd = "'''";
}
}
}
else
{
idx_start = strline.length();
return 0;
}
}
// python: handle """ and '''
if (CurrentQuoteEnd.length() == 3)
{
if (idx_start + 3 >= strline.length())
idx_end = string::npos;
else
{
idx_end = strline.find(CurrentQuoteEnd, idx_start + 3);
if (idx_end != string::npos)
idx_end += 2; // shift to last quote character
}
}
else
idx_end = CUtil::FindCharAvoidEscape(strline, CurrentQuoteEnd[0], idx_start + 1, QuoteEscapeFront);
if (idx_end == string::npos)
{
idx_end = strline.length() - 1;
strline.replace(idx_start + 1, idx_end - idx_start, idx_end - idx_start, '$');
contd = true;
idx_start = idx_end + 1;
}
else
{
if (CurrentQuoteEnd.length() != 3 && (QuoteEscapeRear) && (strline.length() > idx_end + 1) && (strline[idx_end+1] == QuoteEscapeRear))
{
strline[idx_end] = '$';
strline[idx_end+1] = '$';
}
else
{
contd = false;
strline.replace(idx_start + 1, idx_end - idx_start - 1, idx_end - idx_start - 1, '$');
idx_start = idx_end + 1;
}
}
return 1;
}
/*!
* Processes physical and logical lines according to language specific rules.
* NOTE: all the blank lines +
* whole line comments
* should have been blanked from filemap by previous processing
* before reaching this function
*
* \param fmap list of processed file lines
* \param result counter results
* \param fmapBak list of original file lines (same as fmap except it contains unmodified quoted strings)
*
* \return method status
*/
int CTagCounter::LanguageSpecificProcess(filemap* fmap, results* result, filemap* fmapBak)
{
unsigned int cnt = 0;
filemap::iterator fit, fitBak;
string line, lineBak;
size_t lineNumber = 0;
string exclude = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789_$";
string strLSLOC, strLSLOCBak;
char prev_char = 0;
bool ok, data_continue = false;
unsigned int phys_exec_lines = 0;
unsigned int phys_data_lines = 0;
unsigned int temp_lines = 0;
size_t idx_start, quote_idx_start;
char CurrentQuoteEnd = 0;
bool quote_contd = false;
QuoteStart = ">";
QuoteEnd = "<";
for (fit = fmap->begin(), fitBak = fmapBak->begin(); fit != fmap->end(); fit++, fitBak++)
{
if (!CUtil::CheckBlank(fit->line))
{
// replace "quotes" - string between close and open tags
// must be processed after comments since comments start/end with same character as tag
quote_idx_start = 0;
idx_start = 0;
if (quote_contd)
{
// replace quote until next character
ReplaceQuote(fit->line, quote_idx_start, quote_contd, CurrentQuoteEnd);
}
if (!quote_contd)
{
while (idx_start < fit->line.length())
{
quote_idx_start = FindQuote(fit->line, QuoteStart, quote_idx_start, QuoteEscapeFront);
if (quote_idx_start == string::npos)
break;
ReplaceQuote(fit->line, quote_idx_start, quote_contd, CurrentQuoteEnd);
if (quote_idx_start > idx_start)
{
// comment delimiter inside quote
idx_start = quote_idx_start;
continue;
}
}
}
line = fit->line;
lineBak = fitBak->line;
lineNumber = fit->lineNumber;
LSLOC(result, line, lineNumber, lineBak, strLSLOC, strLSLOCBak, prev_char,
data_continue, temp_lines, phys_exec_lines, phys_data_lines);
if (print_cmplx)
{
cnt = 0;
CountTagTally(line, exec_name_list, cnt, 1, exclude, "", "", &result->exec_name_count, false);
}
result->exec_lines[PHY] += phys_exec_lines;
phys_exec_lines = 0;
result->data_lines[PHY] += phys_data_lines;
phys_data_lines = 0;
}
}
QuoteStart = "";
QuoteEnd = "";
// capture closing tag
if (strLSLOC.length() > 0)
{
bool trunc_flag = false;
if (strLSLOCBak.length() == this->lsloc_truncate)
trunc_flag = true;
ok = result->addSLOC(strLSLOCBak, lineNumber, trunc_flag);
cnt = 0;
if (data_name_list.size() > 0)
CUtil::CountTally(strLSLOC, data_name_list, cnt, 1, exclude, "", "", &result->data_name_count);
if (data_continue || cnt > 0)
{
if (ok)
result->data_lines[LOG]++;
result->data_lines[PHY]++;
}
else
{
if (ok)
result->exec_lines[LOG]++;
// since physical data lines are recorded at next LSLOC, check if first line was a data line
if (data_name_list.size() > 0)
{
fit = fmap->begin();
cnt = 0;
CUtil::CountTally(fit->line, data_name_list, cnt, 1, exclude, "", "", NULL);
if (cnt > 0)
result->exec_lines[PHY]++;
}
}
}
return 0;
}
/*!
* Counts the number of comment lines, removes comments, and
* replaces quoted strings by special chars, e.g., $
* All arguments are modified by the method.
*
* \param fmap list of processed file lines
* \param result counter results
* \param fmapBak list of original file lines (same as fmap except it contains unmodified quoted strings)
*
* \return method status
*/
int CFortranCounter::CountCommentsSLOC(filemap* fmap, results* result, filemap *fmapBak)
{
if (LineCommentStart.empty())
return 0;
if (classtype == UNKNOWN || classtype == DATAFILE)
return 0;
bool contd_nextline;
int comment_type = 0;
/*
comment_type:
0 : not comment
1 : line comment, whole line
2 : line comment, embedded
*/
size_t i, idx_start, comment_start;
size_t quote_idx_start;
string curBlckCmtStart, curBlckCmtEnd, prevLine;
char CurrentQuoteEnd = 0;
bool quote_contd = false, found, foundSpc;
filemap::iterator itfmBak = fmapBak->begin();
quote_idx_start = 0;
prevLine = "";
for (filemap::iterator iter = fmap->begin(); iter != fmap->end(); iter++, itfmBak++)
{
contd_nextline = false;
quote_idx_start = 0;
idx_start = 0;
if (CUtil::CheckBlank(iter->line))
continue;
if (quote_contd)
{
// replace quote until next character
ReplaceQuote(iter->line, quote_idx_start, quote_contd, CurrentQuoteEnd);
prevLine = itfmBak->line;
if (quote_contd)
continue;
}
while (!contd_nextline && idx_start < iter->line.length())
{
quote_idx_start = FindQuote(iter->line, QuoteStart, quote_idx_start, QuoteEscapeFront);
comment_start = idx_start;
// check for comment delimiters 'C', 'c' in col 1 (works for most cases)
found = false;
if ((iter->line[0] == 'C' || iter->line[0] == 'c') &&
(prevLine.length() < 1 || prevLine[prevLine.length() - 1] != '&'))
{
// check for reserved 'c' words
for (vector<string>::iterator viter = c_keywords.begin(); viter != c_keywords.end(); viter++)
{
if (CUtil::FindKeyword(iter->line, *viter, 0, TO_END_OF_STRING, false) == 0)
{
found = true;
break;
}
}
if (!found)
{
// check for function or assignment (check for 'c__()' or 'c__ =')
foundSpc = false;
for (i = 1; i < iter->line.length(); i++)
{
if (iter->line[i] == '(')
{
found = true;
break;
}
else if (iter->line[i] == '=')
{
if (i >= iter->line.length() - 1 || iter->line[i + 1] != '=')
found = true;
break;
}
else if (iter->line[i] == ' ')
foundSpc = true;
else if (foundSpc)
break;
}
}
found = !found;
}
// check for comment delimiters '*', '!' in col 1
if (found || ((iter->line[0] == '*' || iter->line[0] == '!') &&
(prevLine.length() < 1 || prevLine[prevLine.length() - 1] != '&')))
{
comment_start = 0;
comment_type = 1;
}
// commented out to favor Fortran 90+ (in Fortran 77 any character in column 6 indicates continuation, not comment)
// else if (iter->line.length() > 6 && iter->line[5] == '!' && CUtil::CheckBlank(iter->line.substr(0, 5)))
// comment_start = string::npos;
else
{
FindCommentStart(iter->line, comment_start, comment_type, curBlckCmtStart, curBlckCmtEnd);
if (comment_start != string::npos)
{
// check for characters before comment
for (i = 0; i < comment_start; i++)
{
if (iter->line[i] != ' ')
{
comment_type = 2;
break;
}
}
}
}
if (comment_start == string::npos && quote_idx_start == string::npos)
{
prevLine = itfmBak->line;
break;
}
if (comment_start != string::npos)
idx_start = comment_start;
// if found quote before comment
if (quote_idx_start != string::npos && (comment_start == string::npos || quote_idx_start < comment_start))
{
ReplaceQuote(iter->line, quote_idx_start, quote_contd, CurrentQuoteEnd);
if (quote_idx_start > idx_start)
{
if (quote_contd)
{
if (itfmBak->line[itfmBak->line.length() - 1] == '&')
{
iter->line[iter->line.length() - 1] = '&';
if (itfmBak->line.length() > 2 && itfmBak->line[itfmBak->line.length() - 2] == ' ')
iter->line[iter->line.length() - 2] = ' ';
}
}
idx_start = quote_idx_start;
prevLine = itfmBak->line;
continue; // comment delimiter inside quote
}
}
else if (idx_start != string::npos)
{
// comment delimiter starts first
switch(comment_type)
{
case 1: // line comment, definitely whole line
case 3:
prevLine = "";
iter->line = "";
itfmBak->line = "";
result->comment_lines++;
contd_nextline = true;
break;
case 2: // line comment, possibly embedded
case 4:
result->e_comm_lines++;
prevLine = "";
iter->line = iter->line.substr(0, idx_start);
itfmBak->line = itfmBak->line.substr(0, idx_start);
// trim trailing space
iter->line = CUtil::TrimString(iter->line, 1);
itfmBak->line = CUtil::TrimString(itfmBak->line, 1);
if (iter->line.empty())
result->comment_lines++; // whole line
else
result->e_comm_lines++; // embedded
contd_nextline = true;
break;
default:
cout << "Error in CountCommentsSLOC()";
break;
}
}
}
}
return 1;
}