/***********************************************************************
			Global include files
***********************************************************************/

#include <stdio.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <time.h>
#include <string.h>
#include <signal.h>
#include <unistd.h>
#include <dirent.h>


/***********************************************************************
		        Global routine declarations
***********************************************************************/

char *getvalue();
char *lstrstr();
char *decode();
void badconnection();
void badsegviol();
void badtimeout();
char *strdup();
char *GetCoden();
char *GetDirectory();


/***********************************************************************
		      Defined files for use by program.
***********************************************************************/

#define	CODENFILE	"/disk2/suitespot/epaps_files/coden_lookup"
#define	TITLEFILE	"/disk2/suitespot/epaps_files/title_lookup"
#define	HEAD		"/disk2/suitespot/epaps_files/head"
#define	EXTFILE		"/disk2/suitespot/epaps_files/ext_includes"
#define	TAIL		"/disk2/suitespot/epaps_files/tail"
#define	ROOTDIR		"/ftp/public/epaps"

/*
#define	CODENFILE	"/netscape/src/epaps/coden_lookup"
#define	TITLEFILE	"/netscape/src/epaps/title_lookup"
#define	HEAD		"/netscape/src/epaps/head"
#define	EXTFILE		"/netscape/src/epaps/ext_includes"
#define	TAIL		"/netscape/src/epaps/tail"
#define	ROOTDIR		"/ftp/edir"
*/


/***********************************************************************
			General define statements.
***********************************************************************/

#define MAXSTRLEN       1024

#define TIMEOUT 	45

#define	HEXLEN		5

/***********************************************************************
			Structure declarations
***********************************************************************/


/***********************************************************************
		      Global variables (yuck) are here.
***********************************************************************/

int skip[5120];
static char query_string[5120];

int main( int argc, char **argv)
{
int i, len;
char *method, *query;
char PapId[ MAXSTRLEN], *Coden, *SubDir, Directory[ 1024];
char IdType[ MAXSTRLEN], DoiCoden[MAXSTRLEN], FileName[MAXSTRLEN], *Ptr;

	/* Close the stderr. */
	fclose(stderr);

	/* Send certain signals to routines so we can catch errors from the
	 * httpd server.
	 */
	signal(SIGSEGV, badsegviol);
	signal(SIGPIPE, badconnection);
	signal(SIGALRM, badtimeout);

	/* First, get all the normal environment CGI things... */
	method = (char *) getenv("REQUEST_METHOD");

	/* If it's a bad method..... */
	if (method == NULL) progerr("Unknown method.");

	/* The POST method */
	if (!strncmp(method, "POST", 4)) {
		len = atoi(getenv("CONTENT_LENGTH"));

		i = 0;
		while (len-- && i < MAXSTRLEN)
			query_string[i++] = fgetc(stdin);

		query_string[i] = '\0';
		}

	/* The GET method. */
	else if (!strncmp(method, "GET", 3)) {
		query = (char *) getenv("QUERY_STRING");

		if (query == NULL) query_string[0] = '\0';
		else strcpy(query_string, query);
		}
	else
		progerr("Unknown method.");


	strncpy( PapId, getvalue( "ID=", ""), MAXSTRLEN);
	strncpy( IdType, getvalue( "IDTYPE=", ""), MAXSTRLEN);
	strncpy( DoiCoden, getvalue( "CODEN=", ""), MAXSTRLEN);
	strncpy( FileName, getvalue( "FILE=", ""), MAXSTRLEN);

	if( !*PapId) {
		progerr2();
		}

	if( *IdType && !strcasecmp( IdType, "doi")) {
		if( !*DoiCoden) progerr2();

		Coden = strdup( DoiCoden);
		Ptr = PapId+8;

		strcpy( PapId, Ptr);
		}
	else Coden = strdup( GetCoden( PapId));

	SubDir = strdup( GetDirectory( Coden));

	sprintf( Directory, "%s/%s/%s", ROOTDIR, SubDir, PapId);

	if( access( Directory, R_OK) != 0) progerr2();

	if( *FileName) {
		printf( "Location: ftp://ftp.aip.org/%s/%s\n\n", Directory+12,
		    FileName);

		exit( 0);
		}

	PrintHead();
	PrintFile( HEAD);

	PrintTitle( Coden);

	ScanFiles( Directory);

	PrintFile( TAIL);

	return( 0);
}


char *GetCoden( char *Id)
{
char *Temp, *p;

	if( strncasecmp( Id, "E-", 2)) progerr2();

	Temp = strdup( Id);

	Temp += 2;

	if( !(p = strchr( Temp, '-'))) progerr2();

	*p = '\0';

	return( Temp);
}


char *GetDirectory( char *Coden)
{
FILE *Inp = NULL;
char Temp[ 1024], *Directory = NULL;

	if( !(Inp = fopen( CODENFILE, "r")))
		progerr2();

	while( fgets( Temp, 1024, Inp)) {
		if( Temp[0] == '#') continue;

		Temp[ strlen( Temp) - 1] = '\0';

		if( !strncmp( Temp, Coden, strlen( Coden)))
			Directory = strdup( Temp+strlen( Coden) + 1);
		}

	fclose( Inp);

	if( !Directory)
		progerr2();

	return( Directory);
} 


PrintTitle( char *Coden)
{
FILE *Inp = NULL;
char Temp[ 1024];
char *p = NULL;

	if( !(Inp = fopen( TITLEFILE, "r"))) return( 1);

	while( fgets( Temp, 1024, Inp)) {
		if( Temp[0] == '#') continue;

		Temp[ strlen( Temp) - 1] = '\0';

		if( !(p = strchr( Temp, ' '))) continue;
		*p++ = '\0';

		if( !strcmp( Temp, Coden))
/*
			printf( "%s\n", Temp+strlen( Coden) + 1);
*/
			printf( "%s\n", p);
		}

	printf( "<hr><p>\n");

	fclose( Inp);

}


PrintFile( char *Filename)
{
FILE *Inp = NULL;
char Temp[ 1024];

	if( !(Inp = fopen( Filename, "r")))
		return( 1);

	while( fgets( Temp, 1024, Inp))
		printf( "%s", Temp);

	fclose( Inp);

	return( 0);
}


OutputURL( int Http, char *Directory, char *Name)
{
struct stat buf;
char Temp[1024];
float Size = 0.0;

	printf( "<a href=\"%s://ftp.aip.org/%s/%s\">%s</a>",
	    Http ? "http" : "ftp", Directory+12, Name, Name);

	sprintf( Temp, "%s/%s", Directory, Name);

	if( !Http && stat( Temp, &buf) == 0) {
		Size = buf.st_size / 1024.0;

		printf( " (%.1f kB)", Size);
		}
}


OutputText( char *Name)
{
FILE *Inp = NULL;
char Temp[ 5120];
int Found = 0;
char *p, *s;

	if( !(p = strchr( Name, '.'))) return( 0);

	if( !(Inp = fopen( EXTFILE, "r"))) return( 0);

	while( fgets( Temp, 5120, Inp)) {
		if( Temp[0] == '#') continue;

		s = strchr( Temp, ' ');
		*s++ = '\0';

		if( !strcmp( Temp, p)) {
			Found++;
			printf( " %s", s);
			}
		}

	fclose( Inp);

	if( Found) printf( "<p>\n");

	return( Found);
}


GetReadmeLine()
{
FILE *Inp = NULL;
char Temp[ 5120];

	if( !(Inp = fopen( EXTFILE, "r"))) return( 0);

	while( fgets( Temp, 5120, Inp)) {
		if( Temp[0] == '#') continue;

		if( !strncasecmp( Temp, "README", 6))
			printf( " %s\n<p>", Temp + 7);
		}

	fclose( Inp);

	return( 1);
}


ScanFiles( char *Directory)
{
DIR *dirp;
struct dirent *direntp;
char Store[500][1024];
int Count = 0, i = 0;

	if( !(dirp = opendir( Directory)))
		progerr( "Cannot open directory...");

	while(( direntp = readdir( dirp)) != NULL) {
		if( *direntp->d_name == '.') continue;

		if( !strncasecmp( direntp->d_name, "index.html", 10)) {
			printf( "<img src=\"/images/epaps_image.gif\"><font size=+1>The author has provided an ");
			printf( "<a href=\"ftp://ftp.aip.org/%s/%s\">HTML index page</a>",
			    Directory+12, direntp->d_name);
			printf( " to facilitate the access of these files.</font><p>");
			}
		else if( !strncasecmp( direntp->d_name, "index.htm", 9)) {
			printf( "<img src=\"/images/epaps_image.gif\"><font size=+1>The author has provided an ");
			printf( "<a href=\"ftp://ftp.aip.org/%s/%s\">HTML index page</a>",
			    Directory+12, direntp->d_name);
			printf( " to facilitate the access of these files.</font><p>");
			}
		}

	rewinddir( dirp);

	while(( direntp = readdir( dirp)) != NULL) {
		if( *direntp->d_name == '.') continue;

		if( !strncasecmp( direntp->d_name, "README", 6)) {
			OutputURL( 1, Directory, direntp->d_name);
			GetReadmeLine();
			}
		else if( !strncasecmp( direntp->d_name, "index.html", 10))
			continue;
		else if( !strncasecmp( direntp->d_name, "index.htm", 9))
			continue;
		else strcpy( Store[Count++], direntp->d_name);
		}

	closedir( dirp);

	for( i=0; i < Count; i++) {
		OutputURL( 0, Directory, Store[ i]);

		if( !OutputText( Store[ i]))
			printf( "<p>\n");
		}
}


/* The NEW getvalue routine. 
 * 2/27/96 - James T. Wonder
 *   The old getvalue routine had some problems with dealing with multiple
 * null variable pairs.  This stemed from it's use of the Boyer-Moore
 * algorithm from Sedgewick.  I removed this from the program and am using
 * the much faster strtok routine.  Most systems should now have strtok
 * included in the standard library.  This was the only reason for
 * Sedgewick :-).
 */
char *getvalue(var, def)
char *var;
char *def;
{
char String[MAXSTRLEN], *Token, *Ptr;
static char tmpstr[MAXSTRLEN];

	strcpy( String, query_string);

	for( Token = strtok( String, "&"); Token; Token = strtok( NULL, "&")) {

		/* Found it!  Yeah! */
		if( !strncasecmp( Token, var, strlen( var))) {
			Ptr = Token + strlen( var);
			if( !Ptr || !*Ptr) {
				if( strlen(def) <= 1)
					return "\0";
				strcpy(tmpstr, decode(def));
				return tmpstr;
				}
			strcpy(tmpstr, decode(Ptr));
			return tmpstr;
			}
		}

	if( strlen(def) <= 1)
		return "\0";

	strcpy(tmpstr, decode(def));
	return tmpstr;
}


PrintHead()
{
	/* Do the content type. */
	printf( "Content-type: text/html\n\n");

	printf( "<html>\n");
	printf( "<head>\n");
	printf( "<title>EPAPS</title>\n");
	printf( "</head>\n");
}


/* Whoops, caught a problem! */
progerr(errstring)
char *errstring;
{
	PrintHead();
	printf("<h1>Error</h1>\n");
	printf("This program encountered an error:<p>\n");
	printf("<p>%s\n<p>\n", errstring);

	exit(0);
}


progerr2()
{
	/*
        printf( "Location: http://scitation.aip.org/error/noepaps.html\n\n");
        */
	printf( "Location:  http://scitation.aip.org/error/noepaps.jsp\n\n");

	exit( 0);
}


/* Decodes URL-encoded strings.  */
char *decode(s)
char *s;
{
int i, j, ascint;
char firstnum, secondnum;
static char newstr[MAXSTRLEN];

        for (i = j = 0; s[i]; i++) {
                if (s[i] == '%') {
                        firstnum = s[++i];
                        secondnum = s[++i];
                        ascint = (getascii(firstnum) * 16)
                        + getascii(secondnum);
			if (ascint == 10 || ascint == 13)
				ascint = 32;
                        newstr[j++] = (char) ascint;
                        continue;
                }
                else if (s[i] == '+')
			newstr[j++] = ' ';
                else
                        newstr[j++] = s[i];
        }
        newstr[j] = '\0';

        return newstr;
}

/* Encodes strings so they can be put in URLs. */
char *encode(s)
char *s;
{
int i;
char hexstr[HEXLEN];
static char line[MAXSTRLEN];
 
	for (i = 0; *s; s++)
		if (*s == ' ')
			line[i++] = '+';
		else if (isdigit(*s) || isalpha(*s))
			line[i++] = *s;
		else {
			sprintf(hexstr, "%0X", *s);
			line[i++] = '%';
			line[i++] = hexstr[0];
			line[i] = hexstr[1];

			if (line[i] == '\0') {
				line[i] = line[i - 1];
				line[i - 1] = '0';
				}

			i++;
			}

		line[i] = '\0';

	return line;
}


/* Returns the right ASCII code for a hex value.  */
int getascii(hex)
char hex;
{
        if ((int) hex >= 48 && (int) hex <= 57)
                return (hex - 48);
        else
                return (hex - 65 + 10);
}


/* Error messages for broken pipes, SIGSEGV, and timing out.  */
void badconnection()
{
	progerr("Can't connect to this server (it is probably down).");
}

void badsegviol()
{
	progerr("Caught a memory violation (probably a parsing error).");
}

void badtimeout()
{
	char message[MAXSTRLEN];

	sprintf(message,
	"This service timed out (%d seconds) - it may be too busy.",
	TIMEOUT);
	progerr(message);
}