void *copyx(void *arg)
{
struct info *inf = (struct info *)arg;
FILE *fp;
char c;
int i;
int byte = blocksz;
int myblock = inf->whichblock;
int myfile = inf->fileno;
int shift = (myblock-1) * blocksz;
if(myblock == block_cnt[myfile] && size[myfile]%blocksz != 0)
byte = size[myfile]%blocksz;
fp = fopen(name[myfile],"r");
fseek(fp, shift, SEEK_SET);
sem_wait(&sem[myfile]);
fseek(fc[myfile], shift, SEEK_SET);
for(i=1; i<=byte; i++)
{
c = fgetc(fp);
fprintf(fc[myfile],"%c",c);
}
sem_post(&sem[myfile]);
fclose(fp);
if(blockdone[whichfile] < block_cnt[whichfile])
{
struct info temp;
sem_wait(&sem2);
blockdone[whichfile]++;
temp.whichblock = blockdone[whichfile];
temp.fileno = whichfile;
sem_post(&sem2);
copyx((void *)&temp);
}
else if(blockdone[whichfile] == block_cnt[whichfile] && whichfile < ARGC-3)
{
struct info temp;
sem_wait(&sem2);
whichfile++;
blockdone[whichfile]++;
temp.whichblock = blockdone[whichfile];
temp.fileno = whichfile;
sem_post(&sem2);
copyx((void *)&temp);
}
}
int main(int argc, char **argv) {
struct chunk c;
off_t a, b = 0;
int fin, fout;
if (argc != 3) {
fprintf(stderr, "usage: mksparse <infile> <outfile>");
return -1;
}
if ((fin = open(argv[1], O_RDONLY)) < 0) {
fprintf(stderr, "error: cannot open '%s' for reading\n", argv[1]);
return -1;
}
if (!strcmp(argv[2], "-")) {
fout = 1;
} else if ((fout = open(argv[2], O_WRONLY | O_CREAT, 0600)) < 0) {
fprintf(stderr, "error: cannot open '%s' for writing\n", argv[2]);
return -1;
}
while ((a = lseek(fin, b, SEEK_DATA)) >= 0) {
b = lseek(fin, a, SEEK_HOLE);
c.start = a;
c.length = b - a;
if (lseek(fin, a, SEEK_SET) != a)
goto fail;
if (writex(fout, &c, sizeof(c)))
goto fail;
if (copyx(fin, fout, c.length, buffer, sizeof(buffer)))
goto fail;
/* fprintf(stderr, "%lu bytes at %lu\n", c.length, c.start); */
}
c.start = c.length = 0;
if (writex(fout, &c, sizeof(c)))
goto fail;
if (close(fout))
goto fail;
return 0;
fail:
fprintf(stderr, "error: %s\n", strerror(errno));
return -1;
}
bool Cxml::get_node(char* xml_string)
{
int k = m_cursor;
int j = 0; //second level cursor;
bool bCDATA = false;
bool bIsPI = false; //Set to true if the cursor is curently inside a processing instruction
char cDelim = 0;
char c = xml_string[k];
const char COPEN = '<';
const char CCLOSE = '>';
const char CSLASH = '/';
const char CSPACE = ' ';
const char CQUOTE = '\'';
const char CDQUOTE = '\"';
const char CEQUAL = '=';
const char CNEW = '\n';
const char CTAB = '\t';
const char CEXCLAMATION = '!';
const char CMINUS = '-';
const char CSQRPO = '[';
const char CSQRPC = ']';
const char SZCDATA[9] = "![CDATA[";
const char CQM = '?';
const char CRET = 13; // carriage return
char *szNodeNameBuff = (char *)calloc(256,sizeof(char));
char *szNodeValBuff = (char *)calloc(256,sizeof(char));
char *szAttrNameBuff = (char *)calloc(256,sizeof(char));
char *szAttrValBuff = (char *)calloc(256,sizeof(char));
if(k >= m_length)
return false;
m_root_node->set_name("XML_DOC");
m_root_node->set_value(""); // just in case we want to access it later...
element* Current = m_root_node->add_child_element();
// We are going to march through the file, one character at a time
// (k counts the file position, and sort the characters into XML
// elements as they appear.
while(k<m_length)
{
c = xml_string[k];
if(c == CNEW || c == CTAB || c == CRET)
{
k++; // This is white space. Eat it.
continue;
}
if(c == COPEN) // Found an "open" character (<)
{
if(xml_string[k+1] == CEXCLAMATION && xml_string[k+2] == CMINUS && xml_string[k+3] == CMINUS) // this is a comment
{ //the comment section
clean_str(szAttrValBuff);
k+=4;
c = xml_string[k];
while(!(xml_string[k] == CMINUS && xml_string[k+1] == CMINUS && xml_string[k+2] == CCLOSE)) // Find the end of the comment.
{
szAttrNameBuff = concat(szAttrNameBuff, c);
c = xml_string[++k];
}
k+=3;
// MODIFIED BY TS
// We don't really want to process comments at all, so just
// discard them.
//if(Current->get_name() != NULL) //we have set this node, navigate to a child of it
// Current = Current->add_child_element();
//Current->set_comment(szAttrNameBuff); //set it as a comment node
//Current = Current->get_parent(); //return to the previous level
continue;
}
while(k+10 < m_length && j < 9)
{
if(xml_string[k+1+j] != SZCDATA[j])
break;
j++;
if(j==8)
{
// This is definitely a CDATA section
k = k + j;
int start = k;
while((k + 3) < m_length && xml_string[k+1] != CSQRPC && xml_string[k+2] != CSQRPC && xml_string[k+3] != CCLOSE)
{
k++;
}
int stop = k;
k+=5;
char* buffer = (char*)calloc(stop - start + 2, sizeof(char));
copyx(buffer, xml_string, start, stop);
Current->set_value(buffer);
free(buffer);
j = 0;
bCDATA = true;
break;
}
}
if(bCDATA)
{
bCDATA = false;
continue;
}
clean_str(szNodeNameBuff);
if(xml_string[k+1] == CSLASH)
{// This is a close tag, hopefully for the last opened node.
Current = Current->get_parent();
k++;
while(xml_string[k] != CCLOSE)
{
k++;
}
k++;
continue;
}
if(xml_string[k+1] == CQM)
{
c = xml_string[++k];
bIsPI = true;
}
// If we are here, this is an open tag. So create a new node.
c = xml_string[++k];
while(c != CSLASH && c != CSPACE && c != CCLOSE)
{//Loop until the node name has been entirely read.
if(c != CNEW && c != CTAB && c != CRET)
szNodeNameBuff = concat(szNodeNameBuff,c);
c = xml_string[++k];
}
if(Current != NULL) // this node is set, navigate to a child of it
if(Current->get_name() != NULL)
Current = Current->add_child_element();
Current->set_name(szNodeNameBuff);
// We are inside the element tag, though at the end of the
// element name. Therefore, if there's a space here,
// there must be an attribute coming.
while(c == CSPACE)
{
c = xml_string[++k];
if(c == CSLASH)
{
break;
}
if(c == CQM && bIsPI)
{
break;
}
clean_str(szAttrNameBuff);
clean_str(szAttrValBuff);
// Get ready for a new attribute.
attribute* pA = new attribute();
// Accumulate characters until there is an equal sign, when
// we'll know the attribute name has been read.
while(c != CEQUAL)
{
if (c != CNEW && c != CTAB && c != CRET)
szAttrNameBuff = concat(szAttrNameBuff, c);
c = xml_string[++k];
}
// Is the attribute value in single or double quotes?
c = xml_string[++k];
if (c == CQUOTE || c == CDQUOTE)
{
cDelim = c;
c = xml_string[++k];
}
// Accumulate characters until the next delimiter, when
// we'll know the attribute value has been read.
while(c != cDelim && cDelim != 0)
{
if(c != CNEW && c != CTAB && c != CRET)
szAttrValBuff = concat(szAttrValBuff, c);
c = xml_string[++k];
}
// Reset the delimiter indicator, and advance one character.
cDelim = 0;
c = xml_string[++k];
// Set the name and value of our new attribute.
pA->set_name(szAttrNameBuff);
pA->set_value(szAttrValBuff);
Current->add_attribute(pA);
} // Repeat if the next character is a space.
if(c == CSLASH) // A slash here indicates a singleton element.
{
Current = Current->get_parent();
c=xml_string[++k];
while(c != CCLOSE)
{
c = xml_string[++k];
}
}
if(c == CQM && bIsPI)
{
Current->set_as_pi();
Current = Current->get_parent();
c=xml_string[++k];
bIsPI = false;
while(c != CCLOSE)
{
c = xml_string[++k];
}
}
if(c == CCLOSE)
{
;
}
}
// If there isn't an open character (<) here, this is the
// contents of an element. Record the characters. Originally
// (pre-TS), this would not allow leading spaces or leading
// slashes (/) in an element. Don't know why, but these have
// been tentatively commented out by TS 6/11/16
if(c != COPEN && c != CCLOSE /*&& c != CSLASH*//* && c != CSPACE*/)
{
clean_str(szNodeValBuff);
while(c != COPEN)
{
if(c != CNEW && c != CTAB && c != CRET/* && c != CSPACE*/)
szNodeValBuff =concat(szNodeValBuff,c);
c = xml_string[++k];
}
Current->set_value(szNodeValBuff);
continue;
}
k++;
}
free(szNodeNameBuff);
free(szNodeValBuff);
free(szAttrNameBuff);
free(szAttrValBuff);
return true;
}
bool Cxml::get_node(char* xml_string)
{
int k = m_cursor;
int j = 0; //second level cursor;
bool bCDATA = false;
bool bIsPI = false; //Set to true if the cursor is curently inside a processing instruction
char cDelim = 0;
char c = xml_string[k];
const char COPEN = '<';
const char CCLOSE = '>';
const char CSLASH = '/';
const char CSPACE = ' ';
const char CQUOTE = '\'';
const char CDQUOTE = '\"';
const char CEQUAL = '=';
const char CNEW = '\n';
const char CTAB = '\t';
const char CEXCLAMATION = '!';
const char CMINUS = '-';
const char CSQRPO = '[';
const char CSQRPC = ']';
const char SZCDATA[9] = "![CDATA[";
const char CQM = '?';
const char CRET = 13; // carriage return
char *szNodeNameBuff = (char *)calloc(256,sizeof(char));
char *szNodeValBuff = (char *)calloc(256,sizeof(char));
char *szAttrNameBuff = (char *)calloc(256,sizeof(char));
char *szAttrValBuff = (char *)calloc(256,sizeof(char));
if(k >= m_length)
return false;
m_root_node->set_name("XML_DOC");
m_root_node->set_value(""); // just in case we want to access it later...
element* Current = m_root_node->add_child_element();
while(k<m_length)
{
c = xml_string[k];
if(c == CNEW || c == CTAB || c == CRET)
{
k++;
continue;
}
if(c == COPEN)
{
if(xml_string[k+1] == CEXCLAMATION && xml_string[k+2] == CMINUS && xml_string[k+3] == CMINUS) // this is a comment
{ //the comment section
clean_str(szAttrValBuff);
k+=4;
c = xml_string[k];
while(!(xml_string[k] == CMINUS && xml_string[k+1] == CMINUS && xml_string[k+2] == CCLOSE))
{
szAttrNameBuff = concat(szAttrNameBuff, c);
c = xml_string[++k];
}
k+=3;
if(Current->get_name() != NULL) //we have set this node, navigate to a child of it
Current = Current->add_child_element();
Current->set_comment(szAttrNameBuff); //set it as a comment node
Current = Current->get_parent(); //return to the previous level
continue;
}
while(k+10 < m_length && j < 9)
{
if(xml_string[k+1+j] != SZCDATA[j])
break;
j++;
if(j==8)
{
// definetly a CDATA section
k = k + j;
int start = k;
while((k + 3) < m_length && xml_string[k+1] != CSQRPC && xml_string[k+2] != CSQRPC && xml_string[k+3] != CCLOSE)
{
k++;
}
int stop = k;
k+=5;
char* buffer = (char*)calloc(stop - start + 2, sizeof(char));
copyx(buffer, xml_string, start, stop);
Current->set_value(buffer);
free(buffer);
j = 0;
bCDATA = true;
break;
}
}
if(bCDATA)
{
bCDATA = false;
continue;
}
clean_str(szNodeNameBuff);
if(xml_string[k+1] == CSLASH)
{ // closing tag for the last opened node
Current = Current->get_parent();
k++;
while(xml_string[k] != CCLOSE)
{
k++;
}
k++;
continue;
}
if(xml_string[k+1] == CQM)
{
c = xml_string[++k];
bIsPI = true;
}
// open tag. It means we have a node so we create it
c = xml_string[++k];
while(c != CSLASH && c != CSPACE && c != CCLOSE)
{//loops until the node name has been entirely read
if(c != CNEW && c != CTAB && c != CRET)
szNodeNameBuff = concat(szNodeNameBuff,c);
c = xml_string[++k];
}
if(Current != NULL) // this node is set, navigate to a child of it
if(Current->get_name() != NULL) // this node is set, navigate to a child of it
Current = Current->add_child_element();
Current->set_name(szNodeNameBuff);
// If there's a space here, there must be an attribute coming.
while(c == CSPACE)
{
c = xml_string[++k];
if(c == CSLASH)
{
break;
}
if(c == CQM && bIsPI)
{
break;
}
clean_str(szAttrNameBuff);
clean_str(szAttrValBuff);
// Get ready for a new attribute.
attribute* pA = new attribute();
// Accumulate characters until there is an equal sign, when
// we'll know the attribute name has been read.
while(c != CEQUAL)
{
if (c != CNEW && c != CTAB && c != CRET)
szAttrNameBuff = concat(szAttrNameBuff, c);
c = xml_string[++k];
}
// Is the attribute value in single or double quotes?
c = xml_string[++k];
if (c == CQUOTE || c == CDQUOTE)
{
cDelim = c;
c = xml_string[++k];
}
// Accumulate characters until the next delimiter, when
// we'll know the attribute value has been read.
while(c != cDelim && cDelim != 0)
{
if(c != CNEW && c != CTAB && c != CRET)
szAttrValBuff = concat(szAttrValBuff, c);
c = xml_string[++k];
}
// Reset the delimiter indicator, and advance one character.
cDelim = 0;
c = xml_string[++k];
// Set the name and value of our new attribute.
pA->set_name(szAttrNameBuff);
pA->set_value(szAttrValBuff);
Current->add_attribute(pA);
} // Repeat if the next character is a space.
if(c == CSLASH)
{
Current = Current->get_parent();
c=xml_string[++k];
while(c != CCLOSE)
{
c = xml_string[++k];
}
}
if(c == CQM && bIsPI)
{
Current->set_as_pi();
Current = Current->get_parent();
c=xml_string[++k];
bIsPI = false;
while(c != CCLOSE)
{
c = xml_string[++k];
}
}
if(c == CCLOSE)
{
;
}
}
if(c != COPEN && c != CCLOSE && c != CSLASH/* && c != CSPACE*/)
{
clean_str(szNodeValBuff);
while(c != COPEN)
{
if(c != CNEW && c != CTAB && c != CRET/* && c != CSPACE*/)
szNodeValBuff =concat(szNodeValBuff,c);
c = xml_string[++k];
}
Current->set_value(szNodeValBuff);
continue;
}
k++;
}
free(szNodeNameBuff);
free(szNodeValBuff);
free(szAttrNameBuff);
free(szAttrValBuff);
return true;
}
