int unc_text::compare(const unc_text& ref1, const unc_text& ref2, int len)
{
int idx, len1, len2;
len1 = ref1.size();
len2 = ref2.size();
if (len > 0)
{
for (idx = 0; (idx < len1) && (idx < len2) && (idx < len); idx++)
{
if (ref1.m_chars[idx] != ref2.m_chars[idx])
{
return(ref1.m_chars[idx] - ref2.m_chars[idx]);
}
}
if (idx == len)
{
return 0;
}
return(len1 - len2);
}
for (idx = 0; (idx < len1) && (idx < len2); idx++)
{
if (ref1.m_chars[idx] != ref2.m_chars[idx])
{
return(ref1.m_chars[idx] - ref2.m_chars[idx]);
}
}
return (len1 - len2);
}
bool unc_text::startswith(const unc_text &text, size_t idx) const
{
size_t si = 0;
const auto orig_idx = idx;
for ( ; idx < size() && si < text.size(); idx++, si++)
{
if (text.m_chars[si] != m_chars[idx])
{
return(false);
}
}
return(idx != orig_idx && (si == text.size()));
}
static int next_up(const unc_text& text, int idx, unc_text& tag)
{
int offs = 0;
while ((idx < text.size()) && unc_isspace(text[idx]))
{
idx++;
offs++;
}
if (text.startswith(tag, idx))
{
return(offs);
}
return(-1);
}
bool unc_text::startswith(const unc_text& text, int idx) const
{
bool match = false;
int si = 0;
while ((idx < size()) && (si < text.size()))
{
if (text.m_chars[si] != m_chars[idx])
{
return false;
}
idx++;
si++;
match = true;
}
return(match && (si == text.size()));
}
static void add_text(const unc_text& text)
{
for (int idx = 0; idx < text.size(); idx++)
{
add_char(text[idx]);
}
}
int unc_text::compare(const unc_text &ref1, const unc_text &ref2, size_t len, bool tcare)
{
const size_t len1 = ref1.size();
const size_t len2 = ref2.size();
const auto max_idx = std::min({ len, len1, len2 });
size_t idx = 0;
for ( ; idx < max_idx; idx++)
{
// exactly the same character ?
if (ref1.m_chars[idx] == ref2.m_chars[idx])
{
continue;
}
int diff; // Issue #2091
if (tcare)
{
diff = ref1.m_chars[idx] - ref2.m_chars[idx];
}
else
{
diff = unc_tolower(ref1.m_chars[idx]) - unc_tolower(ref2.m_chars[idx]);
}
if (diff == 0)
{
/*
* if we're comparing the same character but in different case
* we want to favor lower case before upper case (e.g. a before A)
* so the order is the reverse of ASCII order (we negate).
*/
return(-(ref1.m_chars[idx] - ref2.m_chars[idx]));
}
// return the case-insensitive diff to sort alphabetically
return(diff);
}
if (idx == len)
{
return(0);
}
// underflow save: return(len1 - len2);
return((len1 > len2) ? (len1 - len2) : -static_cast<int>(len2 - len1));
} // unc_text::compare
/**
* Checks for and updates the lead chars.
*
* @param line the comment line
* @return 0=not present, >0=number of chars that are part of the lead
*/
static int cmt_parse_lead(const unc_text& line, int is_last)
{
int len = 0;
while ((len < 32) && (len < line.size()))
{
if ((len > 0) && (line[len] == '/'))
{
/* ignore combined comments */
int tmp = len + 1;
while ((tmp < line.size()) && unc_isspace(line[tmp]))
{
tmp++;
}
if ((tmp < line.size()) && (line[tmp] == '/'))
{
return 1;
}
break;
}
else if (strchr("*|\\#+", line[len]) == NULL)
{
break;
}
len++;
}
if (len > 30)
{
return 1;
}
if ((len > 0) && ((len >= line.size()) || unc_isspace(line[len])))
{
return len;
}
if ((len == 1) && (line[0] == '*'))
{
return len;
}
if (is_last && (len > 0))
{
return len;
}
return 0;
}
/**
* Count the number of characters to the end of the next chunk of text.
* If it exceeds the limit, return true.
*/
static bool next_word_exceeds_limit(const unc_text& text, int idx)
{
int length = 0;
/* Count any whitespace */
while ((idx < text.size()) && unc_isspace(text[idx]))
{
idx++;
length++;
}
/* Count non-whitespace */
while ((idx < text.size()) && !unc_isspace(text[idx]))
{
idx++;
length++;
}
return((cpd.column + length - 1) > cpd.settings[UO_cmt_width].n);
}
/**
* This renders the #if condition to a string buffer.
*/
static void generate_if_conditional_as_text(unc_text& dst, chunk_t *ifdef)
{
chunk_t *pc;
int column = -1;
dst.clear();
for (pc = ifdef; pc != NULL; pc = chunk_get_next(pc))
{
if (column == -1)
{
column = pc->column;
}
if ((pc->type == CT_NEWLINE) ||
(pc->type == CT_COMMENT_MULTI) ||
(pc->type == CT_COMMENT_CPP))
{
break;
}
else if (pc->type == CT_NL_CONT)
{
dst += ' ';
column = -1;
}
else if ((pc->type == CT_COMMENT) ||
(pc->type == CT_COMMENT_EMBED))
{
}
else // if (pc->type == CT_JUNK) || else
{
int spacing;
for (spacing = pc->column - column; spacing > 0; spacing--)
{
dst += ' ';
column++;
}
dst.append(pc->str);
column += pc->len();
}
}
}
void unc_text::append(const unc_text &ref)
{
if (ref.size() == 0)
{
return;
}
m_logtext.pop_back();
m_logtext.insert(std::end(m_logtext),
std::begin(ref.m_logtext), std::end(ref.m_logtext));
m_chars.insert(m_chars.end(), ref.m_chars.begin(), ref.m_chars.end());
}
int unc_text::replace(const char *oldtext, const unc_text& newtext)
{
int fidx = find(oldtext);
int olen = strlen(oldtext);
int rcnt = 0;
while (fidx >= 0)
{
rcnt++;
erase(fidx, olen);
insert(fidx, newtext);
fidx = find(oldtext, fidx + newtext.size() - olen + 1);
}
return(rcnt);
}
bool unc_text::equals(const unc_text& ref) const
{
int len = size();
if (ref.size() != len)
{
return false;
}
for (int idx = 0; idx < len; idx++)
{
if (m_chars[idx] != ref.m_chars[idx])
{
return false;
}
}
return true;
}
bool unc_text::equals(const unc_text &ref) const
{
const size_t len = size();
if (ref.size() != len)
{
return(false);
}
for (size_t idx = 0; idx < len; idx++)
{
if (m_chars[idx] != ref.m_chars[idx])
{
return(false);
}
}
return(true);
}
void unc_text::set(const unc_text& ref, int idx, int len)
{
int size = ref.size();
fix_len_idx(size, idx, len);
m_logok = false;
if ((idx == 0) && (len == size))
{
m_chars = ref.m_chars;
}
else
{
m_chars.resize(len);
int di = 0;
while (len-- > 0)
{
m_chars[di++] = ref.m_chars[idx++];
}
}
}
int unc_text::replace(const char *search_text, const unc_text &replace_text)
{
const size_t s_len = strlen(search_text);
const size_t r_len = replace_text.size();
int rcnt = 0;
int fidx = find(search_text);
while (fidx >= 0)
{
rcnt++;
erase(static_cast<size_t>(fidx), s_len);
(static_cast<size_t>(fidx) >= m_chars.size())
? append(replace_text)
: insert(static_cast<size_t>(fidx), replace_text);
fidx = find(search_text, static_cast<size_t>(fidx) + r_len);
}
return(rcnt);
}
void unc_text::set(const unc_text &ref, size_t idx, size_t len)
{
const auto ref_size = ref.size();
if (len == ref_size)
{
m_chars = ref.m_chars;
update_logtext();
return;
}
m_chars.resize(len);
len = fix_len_idx(ref_size, idx, len);
for (size_t di = 0;
len > 0;
di++, idx++, len--)
{
m_chars[di] = ref.m_chars[idx];
}
update_logtext();
}
void unc_text::insert(size_t idx, const unc_text &ref)
{
if (ref.size() == 0)
{
return;
}
if (idx >= m_chars.size())
{
throw out_of_range(string(__func__) + ":" + to_string(__LINE__)
+ " - idx >= m_chars.size()");
}
const auto utf8_idx = getLogTextUtf8Len(m_chars, idx);
// (A+B) remove \0 from both containers, add back a single at the end
m_logtext.pop_back(); // A
m_logtext.insert(std::next(std::begin(m_logtext), utf8_idx),
std::begin(ref.m_logtext),
std::prev(std::end(ref.m_logtext))); // B
m_logtext.push_back('\0');
m_chars.insert(std::next(std::begin(m_chars), idx),
std::begin(ref.m_chars), std::end(ref.m_chars));
}
/**
* Scans a multiline comment to determine the following:
* - the extra indent of the non-first line (0 or 1)
* - the continuation text ('' or '* ')
*
* The decision is based on:
* - cmt_indent_multi
* - cmt_star_cont
* - the first line length
* - the second line leader length
* - the last line length
*
* If the first and last line are the same length and don't contain any alnum
* chars and (the first line len > 2 or the second leader is the same as the
* first line length), then the indent is 0.
*
* If the leader on the second line is 1 wide or missing, then the indent is 1.
*
* Otherwise, the indent is 0.
*
* @param str The comment string
* @param len Length of the comment
* @param start_col Starting column
* @return 0 or 1
*/
static void calculate_comment_body_indent(cmt_reflow &cmt, const unc_text& str)
{
int idx = 0;
int first_len = 0;
int last_len = 0;
int width = 0;
int len = str.size();
cmt.xtra_indent = 0;
if (!cpd.settings[UO_cmt_indent_multi].b)
{
return;
}
if (cpd.settings[UO_cmt_multi_check_last].b)
{
/* find the last line length */
for (idx = len - 1; idx > 0; idx--)
{
if ((str[idx] == '\n') || (str[idx] == '\r'))
{
idx++;
while ((idx < len) && ((str[idx] == ' ') || (str[idx] == '\t')))
{
idx++;
}
last_len = len - idx;
break;
}
}
}
/* find the first line length */
for (idx = 0; idx < len; idx++)
{
if ((str[idx] == '\n') || (str[idx] == '\r'))
{
first_len = idx;
while ((str[first_len - 1] == ' ') || (str[first_len - 1] == '\t'))
{
first_len--;
}
/* handle DOS endings */
if ((str[idx] == '\r') && (str[idx + 1] == '\n'))
{
idx++;
}
idx++;
break;
}
}
/* Scan the second line */
width = 0;
for (/* nada */; idx < len - 1; idx++)
{
if ((str[idx] == ' ') || (str[idx] == '\t'))
{
if (width > 0)
{
break;
}
continue;
}
if ((str[idx] == '\n') || (str[idx] == '\r'))
{
/* Done with second line */
break;
}
/* Count the leading chars */
if ((str[idx] == '*') ||
(str[idx] == '|') ||
(str[idx] == '\\') ||
(str[idx] == '#') ||
(str[idx] == '+'))
{
width++;
}
else
{
if ((width != 1) || (str[idx - 1] != '*'))
{
width = 0;
}
break;
}
}
//LOG_FMT(LSYS, "%s: first=%d last=%d width=%d\n", __func__, first_len, last_len, width);
/*TODO: make the first_len minimum (4) configurable? */
if ((first_len == last_len) && ((first_len > 4) || (first_len == width)))
{
return;
}
cmt.xtra_indent = ((width == 2) ? 0 : 1);
}
/**
* text starts with '$('. see if this matches a keyword and add text based
* on that keyword.
* @return the number of characters eaten from the text
*/
static int add_comment_kw(const unc_text& text, int idx, cmt_reflow& cmt)
{
if (text.startswith("$(filename)", idx))
{
add_text(path_basename(cpd.filename));
return(11);
}
if (text.startswith("$(class)", idx))
{
chunk_t *tmp = get_next_class(cmt.pc);
if (tmp != NULL)
{
add_text(tmp->str);
return(8);
}
}
/* If we can't find the function, we are done */
chunk_t *fcn = get_next_function(cmt.pc);
if (fcn == NULL)
{
return(0);
}
if (text.startswith("$(message)", idx))
{
add_text(fcn->str);
chunk_t *tmp = chunk_get_next_ncnl(fcn);
chunk_t *word = NULL;
while (tmp)
{
if ((tmp->type == CT_BRACE_OPEN) || (tmp->type == CT_SEMICOLON))
{
break;
}
if (tmp->type == CT_OC_COLON)
{
if (word != NULL)
{
add_text(word->str);
word = NULL;
}
add_text(":");
}
if (tmp->type == CT_WORD)
{
word = tmp;
}
tmp = chunk_get_next_ncnl(tmp);
}
return(10);
}
if (text.startswith("$(function)", idx))
{
if (fcn->parent_type == CT_OPERATOR)
{
add_text("operator ");
}
add_text(fcn->str);
return(11);
}
if (text.startswith("$(javaparam)", idx))
{
add_comment_javaparam(fcn, cmt);
return(12);
}
if (text.startswith("$(fclass)", idx))
{
chunk_t *tmp = chunk_get_prev_ncnl(fcn);
if ((tmp != NULL) && (tmp->type == CT_OPERATOR))
{
tmp = chunk_get_prev_ncnl(tmp);
}
if ((tmp != NULL) && ((tmp->type == CT_DC_MEMBER) ||
(tmp->type == CT_MEMBER)))
{
tmp = chunk_get_prev_ncnl(tmp);
add_text(tmp->str);
return(9);
}
}
return(0);
}
/**
* Outputs a comment. The initial opening '//' may be included in the text.
* Subsequent openings (if combining comments), should not be included.
* The closing (for C/D comments) should not be included.
*
* TODO:
* If reflowing text, the comment should be added one word (or line) at a time.
* A newline should only be sent if a blank line is encountered or if the next
* line is indented beyond the current line (optional?).
* If the last char on a line is a ':' or '.', then the next line won't be
* combined.
*/
static void add_comment_text(const unc_text& text,
cmt_reflow& cmt, bool esc_close)
{
bool was_star = false;
bool was_slash = false;
bool was_dollar = false;
bool in_word = false;
int tmp;
int len = text.size();
for (int idx = 0; idx < len; idx++)
{
if (!was_dollar && cmt.kw_subst &&
(text[idx] == '$') && (len > (idx + 3)) && (text[idx + 1] == '('))
{
idx += add_comment_kw(text, idx, cmt);
if (idx >= len)
{
break;
}
}
/* Split the comment */
if (text[idx] == '\n')
{
in_word = false;
add_char('\n');
cmt_output_indent(cmt.brace_col, cmt.base_col, cmt.column);
if (cmt.xtra_indent)
{
add_char(' ');
}
/* hack to get escaped newlines to align and not dup the leading '//' */
tmp = next_up(text, idx + 1, cmt.cont_text);
if (tmp < 0)
{
add_text(cmt.cont_text);
}
else
{
idx += tmp;
}
}
else if (cmt.reflow &&
(text[idx] == ' ') &&
(cpd.settings[UO_cmt_width].n > 0) &&
((cpd.column > cpd.settings[UO_cmt_width].n) ||
next_word_exceeds_limit(text, idx)))
{
in_word = false;
add_char('\n');
cmt_output_indent(cmt.brace_col, cmt.base_col, cmt.column);
if (cmt.xtra_indent)
{
add_char(' ');
}
add_text(cmt.cont_text);
}
else
{
/* Escape a C closure in a CPP comment */
if (esc_close &&
((was_star && (text[idx] == '/')) ||
(was_slash && (text[idx] == '*'))))
{
add_char(' ');
}
if (!in_word && !unc_isspace(text[idx]))
{
cmt.word_count++;
}
in_word = !unc_isspace(text[idx]);
add_char(text[idx]);
was_star = (text[idx] == '*');
was_slash = (text[idx] == '/');
was_dollar = (text[idx] == '$');
}
}
}
static void cmt_trim_whitespace(unc_text& line, bool in_preproc)
{
/* Remove trailing whitespace on the line */
while ((line.size() > 0) &&
((line.back() == ' ') ||
(line.back() == '\t')))
{
line.pop_back();
}
/* If in a preproc, shift any bs-nl back to the comment text */
if (in_preproc && (line.size() > 1) && (line.back() == '\\'))
{
bool do_space = false;
/* If there was any space before the backslash, change it to 1 space */
line.pop_back();
while ((line.size() > 0) &&
((line.back() == ' ') ||
(line.back() == '\t')))
{
do_space = true;
line.pop_back();
}
if (do_space)
{
line.append(' ');
}
line.append('\\');
}
}
