int ohcount_is_binary_filename(const char *filename) {
char *p = (char *)filename + strlen(filename);
while (p > filename && *(p - 1) != '.') p--;
if (p > filename) {
struct ExtensionMap *re;
int length = strlen(p);
re = ohcount_hash_language_from_ext(p, length);
if (re) return ISBINARY(re->value);
// Try the lower-case version of this extension.
char lowerext[length];
strncpy(lowerext, p, length);
lowerext[length] = '\0';
for (p = lowerext; p < lowerext + length; p++) *p = tolower(*p);
re = ohcount_hash_language_from_ext(lowerext, length);
if (re) return ISBINARY(re->value);
}
return 0;
}
const char *ohcount_detect_language(SourceFile *sourcefile) {
const char *language = NULL;
char *p, *pe;
int length;
// Attempt to detect based on file extension.
length = strlen(sourcefile->ext);
struct ExtensionMap *re = ohcount_hash_language_from_ext(sourcefile->ext,
length);
if (re) language = re->value;
if (language == NULL) {
// Try the lower-case version of this extension.
char lowerext[length + 1];
strncpy(lowerext, sourcefile->ext, length);
lowerext[length] = '\0';
for (p = lowerext; p < lowerext + length; p++) *p = tolower(*p);
struct ExtensionMap *re = ohcount_hash_language_from_ext(lowerext, length);
if (re) return re->value;
}
if (language) {
if (ISAMBIGUOUS(language)) {
// Call the appropriate function for disambiguation.
length = strlen(DISAMBIGUATEWHAT(language));
struct DisambiguateFuncsMap *rd =
ohcount_hash_disambiguate_func_from_id(DISAMBIGUATEWHAT(language),
length);
if (rd) return rd->value(sourcefile);
} else return ISBINARY(language) ? NULL : language;
}
// Attempt to detect based on filename.
length = strlen(sourcefile->filename);
struct FilenameMap *rf =
ohcount_hash_language_from_filename(sourcefile->filename, length);
if (rf) return rf->value;
char line[81] = { '\0' }, buf[81];
// Attempt to detect using Emacs mode line (/^-\*-\s*mode[\s:]*\w/i).
p = ohcount_sourcefile_get_contents(sourcefile);
pe = p;
char *eof = p + ohcount_sourcefile_get_contents_size(sourcefile);
while (pe < eof) {
// Get the contents of the first line.
while (pe < eof && *pe != '\r' && *pe != '\n') pe++;
length = (pe - p <= sizeof(line)) ? pe - p : sizeof(line);
strncpy(line, p, length);
line[length] = '\0';
if (*line == '#' && *(line + 1) == '!') {
// First line was sh-bang; loop to get contents of second line.
while (*pe == '\r' || *pe == '\n') pe++;
p = pe;
} else break;
}
char *eol = line + strlen(line);
for (p = line; p < eol; p++) *p = tolower(*p);
p = strstr(line, "-*-");
if (p) {
p += 3;
while (*p == ' ' || *p == '\t') p++;
if (strncmp(p, "mode", 4) == 0) {
p += 4;
while (*p == ' ' || *p == '\t' || *p == ':') p++;
}
pe = p;
while (isalnum(*pe)) pe++;
length = pe - p;
strncpy(buf, p, length);
buf[length] = '\0';
struct LanguageMap *rl = ohcount_hash_language_from_name(buf, length);
if (rl) return rl->name;
}
// Attempt to detect based on Unix 'file' command.
int tmpfile = 0;
char *path = sourcefile->filepath;
if (sourcefile->diskpath)
path = sourcefile->diskpath;
if (access(path, F_OK) != 0) { // create temporary file
path = malloc(21);
strncpy(path, "/tmp/ohcount_XXXXXXX", 20);
*(path + 21) = '\0';
int fd = mkstemp(path);
char *contents = ohcount_sourcefile_get_contents(sourcefile);
log_it("contents:");
log_it(contents);
length = contents ? strlen(contents) : 0;
write(fd, contents, length);
close(fd);
tmpfile = 1;
}
char command[strlen(path) + 11];
sprintf(command, "file -b '%s'", path);
FILE *f = popen(command, "r");
if (f) {
fgets(line, sizeof(line), f);
char *eol = line + strlen(line);
for (p = line; p < eol; p++) *p = tolower(*p);
p = strstr(line, "script text");
if (p && p == line) { // /^script text(?: executable)? for \w/
p = strstr(line, "for ");
if (p) {
p += 4;
pe = p;
while (isalnum(*pe)) pe++;
length = pe - p;
strncpy(buf, p, length);
buf[length] = '\0';
struct LanguageMap *rl = ohcount_hash_language_from_name(buf, length);
if (rl) language = rl->name;
}
} else if (p) { // /(\w+)(?: -\w+)* script text/
do {
p--;
pe = p;
while (*p == ' ') p--;
while (p != line && isalnum(*(p - 1))) p--;
if (p != line && *(p - 1) == '-') p--;
} while (*p == '-'); // Skip over any switches.
length = pe - p;
strncpy(buf, p, length);
buf[length] = '\0';
struct LanguageMap *rl = ohcount_hash_language_from_name(buf, length);
if (rl) language = rl->name;
} else if (strstr(line, "xml")) language = LANG_XML;
pclose(f);
if (tmpfile) {
remove(path);
free(path);
}
if (language) return language;
}
return NULL;
}
const char *disambiguate_h(SourceFile *sourcefile) {
char *p, *pe;
int length;
// If the directory contains a matching *.m file, likely Objective C.
length = strlen(sourcefile->filename);
if (strcmp(sourcefile->ext, "h") == 0) {
char path[length];
strncpy(path, sourcefile->filename, length);
path[length] = '\0';
*(path + length - 1) = 'm';
char **filenames = ohcount_sourcefile_get_filenames(sourcefile);
if (filenames) {
int i;
for (i = 0; filenames[i] != NULL; i++)
if (strcmp(path, filenames[i]) == 0)
return LANG_OBJECTIVE_C;
}
}
// Attempt to detect based on file contents.
char line[81], buf[81];
p = ohcount_sourcefile_get_contents(sourcefile);
pe = p;
char *eof = p + ohcount_sourcefile_get_contents_size(sourcefile);
while (pe < eof) {
// Get a line at a time.
while (pe < eof && *pe != '\r' && *pe != '\n') pe++;
length = (pe - p <= sizeof(line)) ? pe - p : sizeof(line);
strncpy(line, p, length);
line[length] = '\0';
char *eol = line + strlen(line);
char *line_end = pe;
// Look for C++ headers.
if (*line == '#') {
p = line + 1;
while (*p == ' ' || *p == '\t') p++;
if (strncmp(p, "include", 7) == 0 &&
(*(p + 7) == ' ' || *(p + 7) == '\t')) {
// /^#\s*include\s+[<"][^>"]+[>"]/
p += 8;
while (*p == ' ' || *p == '\t') p++;
if (*p == '<' || *p == '"') {
// Is the header file a C++ header file?
p++;
pe = p;
while (pe < eol && *pe != '>' && *pe != '"') pe++;
length = pe - p;
strncpy(buf, p, length);
buf[length] = '\0';
if (ohcount_hash_is_cppheader(buf, length))
return LANG_CPP;
// Is the extension for the header file a C++ file?
p = pe;
while (p > line && *(p - 1) != '.') p--;
length = pe - p;
strncpy(buf, p, length);
buf[length] = '\0';
struct ExtensionMap *re = ohcount_hash_language_from_ext(buf, length);
if (re && strcmp(re->value, LANG_CPP) == 0)
return LANG_CPP;
}
}
}
// Look for C++ keywords.
p = line;
while (p < eol) {
if (islower(*p) && p != line && !isalnum(*(p - 1)) && *(p - 1) != '_') {
pe = p;
while (islower(*pe)) pe++;
if (!isalnum(*pe) && *pe != '_') {
length = pe - p;
strncpy(buf, p, length);
buf[length] = '\0';
if (strcmp(buf, "class") == 0 ||
strcmp(buf, "namespace") == 0 ||
strcmp(buf, "template") == 0 ||
strcmp(buf, "typename") == 0)
return LANG_CPP;
}
p = pe + 1;
} else p++;
}
// Next line.
pe = line_end;
while (*pe == '\r' || *pe == '\n') pe++;
p = pe;
}
// Nothing to suggest C++.
return LANG_C;
}
/* Use all available means to detect file language
*/
const char *ohcount_detect_language(SourceFile *sourcefile) {
const char *language = NULL;
char *p, *pe;
int length;
// Attempt to detect using Emacs mode line (/^-\*-\s*mode[\s:]*\w/i).
char line[81] = { '\0' }, buf[81];
p = ohcount_sourcefile_get_contents(sourcefile);
pe = p;
char *eof = p + ohcount_sourcefile_get_contents_size(sourcefile);
while (pe < eof) {
// Get the contents of the first line.
while (pe < eof && *pe != '\r' && *pe != '\n') pe++;
length = (pe - p <= sizeof(line)) ? pe - p : sizeof(line);
strncpy(line, p, length);
line[length] = '\0';
if (*line == '#' && *(line + 1) == '!') {
// First line was sh-bang; loop to get contents of second line.
while (*pe == '\r' || *pe == '\n') pe++;
p = pe;
} else break;
}
p = strstr(line, "-*-");
if (p) {
p += 3;
while (*p == ' ' || *p == '\t') p++;
// detect "mode" (any capitalization)
if (strncasecmp(p, "mode", 4) == 0) {
p += 4;
while (*p == ' ' || *p == '\t' || *p == ':') p++;
}
pe = p;
while (!isspace(*pe) && *pe != ';' && pe != strstr(pe, "-*-")) pe++;
length = (pe - p <= sizeof(buf)) ? pe - p : sizeof(buf);
strncpy(buf, p, length);
buf[length] = '\0';
// Special case for "c" or "C" emacs mode header: always means C, not C++
if (strcasecmp(buf, "c") == 0) {
return LANG_C;
}
// First try it with the language name.
struct LanguageMap *rl = ohcount_hash_language_from_name(buf, length);
if (rl) language = rl->name;
if(!language) {
// Then try it with the extension table.
struct ExtensionMap *re = ohcount_hash_language_from_ext(buf, length);
if (re) language = re->value;
}
if (!language) {
// Try the lower-case version of this modeline.
for (pe = buf; pe < buf+length; pe++) *pe = tolower(*pe);
// First try it with the language name.
rl = ohcount_hash_language_from_name(buf, length);
if (rl) language = rl->name;
}
if (!language) {
// Then try it with the extension table.
struct ExtensionMap *re = ohcount_hash_language_from_ext(buf, length);
if (re) language = re->value;
}
}
// Attempt to detect based on file extension.
if(!language) {
length = strlen(sourcefile->ext);
struct ExtensionMap *re = ohcount_hash_language_from_ext(sourcefile->ext,
length);
if (re) language = re->value;
if (!language) {
// Try the lower-case version of this extension.
char lowerext[length + 1];
strncpy(lowerext, sourcefile->ext, length);
lowerext[length] = '\0';
for (p = lowerext; p < lowerext + length; p++) *p = tolower(*p);
struct ExtensionMap *re = ohcount_hash_language_from_ext(lowerext, length);
if (re) language = re->value;
}
}
// Attempt to detect based on filename.
if(!language) {
length = strlen(sourcefile->filename);
struct FilenameMap *rf =
ohcount_hash_language_from_filename(sourcefile->filename, length);
if (rf) language = rf->value;
}
// Attempt to detect based on Unix 'file' command.
if(!language) {
language = detect_language_magic(sourcefile);
}
if (language) {
if (ISAMBIGUOUS(language)) {
// Call the appropriate function for disambiguation.
length = strlen(DISAMBIGUATEWHAT(language));
struct DisambiguateFuncsMap *rd =
ohcount_hash_disambiguate_func_from_id(DISAMBIGUATEWHAT(language),
length);
if (rd) language = rd->value(sourcefile);
} else language = ISBINARY(language) ? NULL : language;
}
return language;
}
