/* Copyright 1997 Acorn Computers Ltd
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
/***************************************************/
/* File : URLutils.c */
/* */
/* Purpose: URL manipulation for the browser. */
/* */
/* Author : Merlyn Kline for Customer browser */
/* This source adapted by A.D.Hodgkinson */
/* from various original functions */
/* */
/* History: 06-Feb-97: Created. */
/***************************************************/
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include "swis.h"
#include "URI.h" /* URI handler API, in URILib:h */
#include "wimp.h"
#include "event.h"
#include "svcprint.h"
#include "Global.h"
#include "Utils.h"
#include "Fetch.h" /* (Which itself includes URLstat.h) */
#include "Filetypes.h"
#include "URLutils.h"
/* Local definitions */
#define ExtensionMatches(url, len, ext) (len) > strlen(ext) && !strncmp((url) + (len) - strlen(ext), (ext), strlen(ext))
/* Local variables */
/* Pointer to first item in queue of URLs dispatched through */
/* the URI handler (structure defined in URLutils.h) */
static uri_queue * uri_queue_base = NULL;
/*************************************************/
/* urlutils_pathname_to_url() */
/* */
/* Takes a pathname, and turns it into a File */
/* URL, if it isn't one already. The pathname */
/* that you give is altered directly, so if you */
/* want to remember the path as well as the URL, */
/* ensure there is a second copy of it in */
/* another buffer somewhere. */
/* */
/* Parameters: Pointer to the pathname; */
/* */
/* Size of the buffer the pathname */
/* is stored in. */
/*************************************************/
void urlutils_pathname_to_url(char * path, int buffersize)
{
int len;
len = strlen(FileMethod ProtocolSepShort); /* (urlutils.h) */
/* If the first part of the string doesn't match the FileMethod */
/* specifier (see URLutils.h) then insert this text and convert */
/* the rest of the path to a URL. */
if (strncmp(path, FileMethod ProtocolSepShort, len))
{
memmove(path + len, path, buffersize - len);
strncpy(path, FileMethod ProtocolSepShort, len);
/* Ensure the string is terminated */
path[buffersize - 1] = 0;
/* Now translate the pathname part of the URL to a Unix-style */
/* path scheme. */
urlutils_translate_pathname(path + len);
}
return;
}
/*************************************************/
/* urlutils_url_to_pathname() */
/* */
/* Takes a file:// URL, and turns it into a RISC */
/* OS pathname, if it isn't one already. The URL */
/* that you give is altered directly, so if you */
/* want to remember the URL as well as the new */
/* path, ensure there is a second copy of it in */
/* another buffer somewhere. */
/* */
/* Parameters: Pointer to the URL; */
/* */
/* Size of the buffer the URL is */
/* stored in. */
/*************************************************/
void urlutils_url_to_pathname(char * url, int buffersize)
{
int len;
len = strlen(FileMethod); /* (urlutils.h) */
/* If the first part of the string doesn't match the FileMethod */
/* specifier (see URLutils.h) then we can't do anything. */
if (!url || strncmp(url, FileMethod, len)) return;
/* Copy over the file: specifier and any '/'s */
while (url[len] == '/') len++;
memmove(url, url + len, buffersize - len);
/* Convert the path component of the URL back to RISC OS style */
urlutils_translate_pathname(url);
return;
}
/*************************************************/
/* urlutils_translate_pathname() */
/* */
/* Takes a RISC OS-style pathname and turns it */
/* into a Unix-style pathname, or vice versa. */
/* */
/* The pathname you give is altered directly, so */
/* if you want to remember the path before */
/* translation, ensure there is a second copy of */
/* in another buffer somewhere. */
/* */
/* Parameters: Pointer to the pathname. */
/*************************************************/
void urlutils_translate_pathname(char * path)
{
char * p;
p = path;
/* Skip past any filing system separators (e.g. as in the */
/* colons in 'ADFS::<disc>.<path>'). */
while (*p && *p != ':') p++;
/* Swap '/' for '.' */
while (*p)
{
if (*p == '/') *p = '.';
else if (*p == '.') *p = '/';
p++;
}
return;
}
/*************************************************/
/* urlutils_leafname_from_url() */
/* */
/* Returns a pointer to a string containing a */
/* possible leafname, based upon the URL passed */
/* into the function. */
/* */
/* Parameters: Pointer to a URL string; */
/* */
/* Pointer to a buffer into which to */
/* place the leafname (not the same */
/* as the URL string); */
/* */
/* Size of the buffer. */
/* */
/* Returns: Will fill the buffer in with some */
/* leafname, even if one could not */
/* be worked out from the URL. */
/* Returns the buffer pointer for */
/* convenience (even though the */
/* caller will almost certainly know */
/* this). */
/* */
/* Assumes: Neither pointer may be NULL. The */
/* buffer must be at least 2 bytes */
/* in size. If either condition is */
/* not met, NULL is returned and the */
/* buffer is left untouched. */
/*************************************************/
char * urlutils_leafname_from_url(char * url, char * leaf, int size)
{
int l = 0;
if (!url || !leaf || size < 2) return NULL;
memset(leaf, 0, size);
/* l holds the string length if b->urlfdata exists */
l = (int) strlen(url);
/* If the string exists and is not null, try to extract */
/* a leafname from it. */
if (l)
{
/* Set 'e' to point at the last character in the string, */
/* and we will use 's' to point to the first character */
/* after the host name. */
int e = l - 1;
int s = 0;
int found = 0;
/* We want to do the following, bearing in mind that whilst */
/* a protocol at the start of the URL is assumed, it may be */
/* separated by nothing more than a colon (not even ':/' or */
/* '://'). The use of square brackets implies one or more */
/* analogous optional terms. */
/* */
/* a.b.c/[dir/]name.html -> name */
/* a.b.c/dir[/] -> dir */
/* a.b.c[/] -> Generic */
/* a.b.c/name.txt -> name */
/* a.b.c/name.tar.gz -> name/tar (as a consequence) */
/* a.b.c/name_1.2.1.gz -> name/1/2/1 (as a consequence) */
/* a.b.c/name.html#anc -> anc */
/* */
/* The tricky part is determining what bit is host name and */
/* what bit is path; it gets easier after that. First, look */
/* for a ':'. This will either be a protocol separator, or */
/* a port number separator. Trouble is, you need to skip */
/* any '/'s after the ':' if it's the first case, so check */
/* if any '.'s are passed when looking for the ':'. If so, */
/* we're on a port number. The only possible failure case */
/* is a URL with no protocol specified and a single word */
/* host name with a port number; but then, just about every */
/* other system will fail as the host name will be taken to */
/* be the protocol (e.g. 'dylan:8080'). So this isn't a */
/* problem, really - we'd return a generic response. */
while (s < l && url[s] != ':')
{
if (url[s] == '.') found = 1;
s++;
}
/* Did we run out of string? If so, go back to the start; */
/* we may just have no protocol or port specified. */
if (s == l) s = 0;
/* If we've not found any dots, may need to skip a few slashes; */
/* but not more than two, as three slashes (for example) implies */
/* that no host name is present. */
if (!found)
{
s++; /* Skip past the ':' */
if (url[s] == '/') s++;
if (url[s] == '/') s++;
/* If we ran out of string, must use a generic response */
if (s == l) goto return_generic;
}
/* We're either now on the ':' separating the host name */
/* and port number, or on the first character after any */
/* '/'s or the ':' separating the protocol from the */
/* host name. In either case, we now search forward for */
/* a single slash - the separator between host name and */
/* path. */
while (s < l && url[s] != '/')
{
s++;
}
/* Skip past the '/' */
if (url[s] == '/') s++;
/* If we're run out string, go for a generic response - */
/* e.g. 'http://www/' or just 'http://www'. */
if (s >= l) goto return_generic;
/* Otherwise, we can now start searching backwards for a */
/* usable leafname, knowing that reaching 's' is the */
/* exit condition. First, look for anchor names. */
if (strrchr(url, '#') > strrchr(url, '/')) /* Want 'a/b#c' -> 'c', but not 'a#b/c' -> 'b/c' */
{
while (e > s && url[e] != '#')
{
e--;
}
/* Step forward past the '#' */
e++;
/* If it turns out that the '#' is all that there is after */
/* the host name, it may be possible to step back and get */
/* whatever comes before it; or if we hit 's' in trying, */
/* we'll have to return a generic response. */
if (e == l)
{
e --;
if (e <= s) goto return_generic;
else goto get_directory_name;
}
/* Right, copy the string from this point forwards */
strncpy(leaf, url + e, size - 1);
/* Finished, so jump to the final stripping routine */
goto strip_illegal_chars;
}
/* Right, if we have a trailing '/' or '#' (from the above code), */
/* following a directory name, then get that directory name. */
get_directory_name:
if (url[e] == '/' || url[e] == '#')
{
int chars;
found = e--;
/* Find the start point of the name */
while (e > s && url[e] != '/') e--;
/* If e is greater than s, we're sitting on the '/' */
/* found above, so advance past it. */
if (e > s) e++;
/* Copy the name in */
chars = found - e;
if (chars > size - 1) chars = size - 1;
strncpy(leaf, url + e, chars);
/* Finished - just need to take care of any illegal characters */
goto strip_illegal_chars;
}
/* Otherwise, continue, looking for a '/'. If a '.' is found on the */
/* way, remember the offset for the first time it is encountered. */
found = 0;
while (e > s && url[e] != '/')
{
if (url[e] == '.' && !found) found = e;
e--;
}
if (url[e] == '/')
{
e++;
if (e >= l) goto return_generic;
}
/* We now have 'e' pointing to the start of a leafname */
/* and possibly 'found' pointing to the start of a */
/* filename extension. If the latter is true, check */
/* that there is something between the two... */
if (found && found - 1 == e) goto return_generic;
/* OK, we have a string */
if (!found)
{
strncpy(leaf, url + e, size - 1);
}
else
{
int chars = found - e;
if (chars > size - 1) chars = size - 1;
strncpy(leaf, url + e, chars);
}
/* Right, that's the worst of it over...! */
goto strip_illegal_chars;
return_generic:
lookup_token("NoURLleaf:Index",0,0);
strncpy(leaf, tokens, size - 1);
}
/* There was apparently no URL in the buffer, so offer a */
/* neutral filename of HTMLfile. */
else
{
lookup_token("NoURLdata:HTMLfile",0,0); /* Will put the string in the 'tokens' global buffer */
strncpy(leaf, tokens, size - 1);
}
strip_illegal_chars:
/* Scan the leaf for illegal characters */
l = 0;
while (leaf[l])
{
/* A few we can replace with meaningful alternatives */
if (leaf[l] == '.') leaf[l] = '/';
if (leaf[l] == '\\') leaf[l] = '/';
if (leaf[l] == '&') leaf[l] = '+';
if (leaf[l] == '"') leaf[l] = '\'';
/* And the rest, replace with underscores */
if (
leaf[l] == '$' ||
leaf[l] == '%' ||
leaf[l] == '@' ||
leaf[l] == '^' ||
leaf[l] == ':' ||
leaf[l] == '#' ||
leaf[l] == '*' ||
leaf[l] == '"' ||
leaf[l] == '|'
)
leaf[l] = '_';
l++;
}
/* Finished. */
return leaf;
}
/*************************************************/
/* urlutils_host_name_from_url() */
/* */
/* Extracts the host name from a given URL. */
/* */
/* Parameters: Pointer to the URL string; */
/* */
/* Pointer to a buffer to write the */
/* host name into; */
/* */
/* Size of the buffer. */
/*************************************************/
void urlutils_host_name_from_url(char * url, char * host, int size)
{
char * p;
host[0] = 0;
/* First look for '//', as in 'http://' */
p = strstr(url, ProtocolSeparator);
if (p)
{
/* If found, copy everything after that into 'host' */
p += 2;
strncpy(host, p, size - 1);
host[size - 1] = 0;
/* Now search for a '/', as in 'http://www.acorn/', */
/* and if found force a terminator there. */
p = strchr(host, '/');
if (p) *p = 0;
}
}
/*************************************************/
/* urlutils_filetype_from_url() */
/* */
/* Examines a URL and returns a RISC OS filetype */
/* based on the filename extension in the URL. */
/* */
/* Parameters: Pointer to a null-terminated URL */
/* string. */
/* */
/* Returns: A RISC OS filetype. */
/*************************************************/
int urlutils_filetype_from_url(const char * url)
{
// For now, hard code it. In future, use a mime mapper module.
int len;
if (!url || !*url) return NULL;
else len = strlen(url);
/* Document types */
if (ExtensionMatches(url, len, ".html")) return FileType_HTML;
if (ExtensionMatches(url, len, ".htm")) return FileType_HTML;
if (ExtensionMatches(url, len, ".txt")) return FileType_TEXT;
if (ExtensionMatches(url, len, ".shtml")) return FileType_HTML;
if (ExtensionMatches(url, len, ".shtm")) return FileType_HTML;
if (ExtensionMatches(url, len, ".pdf")) return FileType_PDF;
if (ExtensionMatches(url, len, ".doc")) return FileType_WORD;
if (ExtensionMatches(url, len, ".ps")) return FileType_PS;
if (ExtensionMatches(url, len, ".eps")) return FileType_PS;
if (ExtensionMatches(url, len, ".wri")) return FileType_DOS;
if (ExtensionMatches(url, len, ".xls")) return FileType_XLS;
/* Images */
if (ExtensionMatches(url, len, ".gif")) return FileType_GIF;
if (ExtensionMatches(url, len, ".jpg")) return FileType_JPEG;
if (ExtensionMatches(url, len, ".jpeg")) return FileType_JPEG;
if (ExtensionMatches(url, len, ".tiff")) return FileType_TIFF;
if (ExtensionMatches(url, len, ".tif")) return FileType_TIFF;
if (ExtensionMatches(url, len, ".png")) return FileType_PNG;
/* Archives */
if (ExtensionMatches(url, len, ".zip")) return FileType_ARC;
if (ExtensionMatches(url, len, ".arc")) return FileType_ARC;
if (ExtensionMatches(url, len, ".spk")) return FileType_ARC;
if (ExtensionMatches(url, len, ".arj")) return FileType_ARC;
if (ExtensionMatches(url, len, ".gz")) return FileType_GZ;
if (ExtensionMatches(url, len, ".tar")) return FileType_ARC;
if (ExtensionMatches(url, len, ".zoo")) return FileType_ARC;
/* Sounds */
if (ExtensionMatches(url, len, ".wav")) return FileType_WAVE;
if (ExtensionMatches(url, len, ".arm")) return FileType_ARMA;
if (ExtensionMatches(url, len, ".mod")) return FileType_MOD;
/* Movies */
if (ExtensionMatches(url, len, ".mov")) return FileType_AVI;
if (ExtensionMatches(url, len, ".avi")) return FileType_AVI;
if (ExtensionMatches(url, len, ".qt")) return FileType_AVI;
if (ExtensionMatches(url, len, ".qtvr")) return FileType_AVI;
if (ExtensionMatches(url, len, ".rpl")) return FileType_ARMO;
if (ExtensionMatches(url, len, ".rep")) return FileType_ARMO;
/* Miscellaneous */
if (ExtensionMatches(url, len, ".bin")) return FileType_DATA;
if (ExtensionMatches(url, len, ".dat")) return FileType_DATA;
if (ExtensionMatches(url, len, ".data")) return FileType_DATA;
if (ExtensionMatches(url, len, ".exe")) return FileType_DOS;
if (ExtensionMatches(url, len, ".com")) return FileType_DOS;
/* Otherwise, return text */
return FileType_TEXT;
}
/*************************************************/
/* urlutils_create_hotlist_url() */
/* */
/* Creates a URL though which a hotlist file may */
/* be fetched. This is done by looking at a */
/* system variable 'Browse$HotlistURL'. If that */
/* isn't set it looks at 'Browse$HotlistURIFile' */
/* which can hold the path of a URI file to */
/* load. Lastly, it looks at the Choices file */
/* token 'HotlistPath', where a RISC OS pathname */
/* pointing to the file should be placed. This */
/* will be turned into a URL for fetching, so */
/* care must be taken over the path used. */
/* */
/* Parameters: Pointer to a buffer to place the */
/* URL in (not in a flex block!); */
/* */
/* Size of the buffer. */
/*************************************************/
void urlutils_create_hotlist_url(char * buffer, int size)
{
_kernel_oserror * e;
/* See if the variable exists. */
memset(buffer, 0, size);
/* Equivalent to getenv, but the RISC OS implementation evaluates */
/* the system variable as an expression which we don't want; hence */
/* the direct use of the SWI. */
e = _swix(OS_ReadVarVal,
_INR(0,4),
"Browse$HotlistURL", /* Variable name */
buffer, /* Buffer */
size, /* Buffer size, -1 to check it exists */
0, /* Name pointer (0 for 1st call) */
4); /* Variable type (4 = literal string) */
/* First lookup failed, so try the URI file. */
if (e)
{
e = _swix(OS_ReadVarVal,
_INR(0,4),
"Browse$HotlistURIFile",
buffer,
size,
0,
4);
if (e)
{
/* If the above gives an error, the variable doesn't exist; get */
/* the HotlistPath string from the Messages file instead. */
strncpy(buffer, lookup_choice("HotlistPath",1,0), size - 1);
urlutils_pathname_to_url(buffer, size);
}
else
{
char path[Limits_OS_Pathname];
StrNCpy0(path, buffer);
urlutils_load_uri_file(buffer, size, path);
}
}
}
/*************************************************/
/* urlutils_create_home_url() */
/* */
/* Creates a URL though which a home page may be */
/* fetched. This is done by looking at a system */
/* variable 'Browse$HomeURL'. If that isn't set, */
/* it looks at 'Browse$HomeURIFile', which can� */
/* hold the path of a URI file to load. Lastly, */
/* it looks at the Choices file token */
/* 'HomePage'. */
/* */
/* Parameters: Pointer to a buffer to place the */
/* URL in (not in a flex block!); */
/* */
/* Size of the buffer. */
/*************************************************/
void urlutils_create_home_url(char * buffer, int size)
{
_kernel_oserror * e;
/* See if the variable exists */
memset(buffer, 0, size);
/* Equivalent to getenv, but the RISC OS implementation evaluates */
/* the system variable as an expression which we don't want; hence */
/* the direct use of the SWI. */
e = _swix(OS_ReadVarVal,
_INR(0,4),
"Browse$HomeURL", /* Variable name */
buffer, /* Buffer */
size, /* Buffer size, -1 to check it exists */
0, /* Name pointer (0 for 1st call) */
4); /* Variable type (4 = literal string) */
/* First lookup failed, so try the URI file. */
if (e)
{
e = _swix(OS_ReadVarVal,
_INR(0,4),
"Browse$HomeURIFile",
buffer,
size,
0,
4);
if (e)
{
/* If the above gives an error, the variable doesn't exist; get */
/* the HotlistPath string from the global choices structure */
/* instead. */
strncpy(buffer, choices.home_page, size - 1);
}
else
{
char path[Limits_OS_Pathname];
StrNCpy0(path, buffer);
urlutils_load_uri_file(buffer, size, path);
}
}
}
/*************************************************/
/* urlutils_fix_url() */
/* */
/* Takes a URL and 'fixes' it, e.g. appends a */
/* '/' character to a URL which is missing one. */
/* The contents of the buffer you give with the */
/* URL inside are altered directly, so if you */
/* want to remember the old URL, ensure there is */
/* a second copy of it in another buffer */
/* somewhere. */
/* */
/* Parameters: Pointer to the URL; */
/* */
/* Size of the buffer the URL is */
/* stored in. */
/* */
/* Returns: Pointer to the fixed URL (which */
/* at the moment is the buffer that */
/* you passed in). */
/*************************************************/
char * urlutils_fix_url(char * buffer, int buffersize)
{
int len, shl;
int flen, fshl;
int glen, gshl;
int blen, plen;
plen = strlen(ProtocolSeparator);
shl = strlen(HTTPmethod);
len = shl + plen;
fshl = strlen(FTPmethod);
flen = fshl + plen;
gshl = strlen(GopherMethod);
glen = gshl + plen;
blen = strlen(buffer);
/* If there's no ':' in the string, insert a protocol type */
if (
!strchr(buffer, ':') &&
blen + len < buffersize &&
blen + flen < buffersize
)
{
/* If the site appears to be an FTP site, insert the FTP protocol */
/* at the start. Similarly for Gopher; else insert HTTP. */
if (!strncmp(buffer, FTPmethod, fshl - 1)) /* -1 as we don't want to compare the ':' */
{
memmove(buffer + flen, buffer, buffersize - flen);
strncpy(buffer, FTPmethod ProtocolSeparator, flen);
}
else if (!strncmp(buffer, GopherMethod, gshl - 1))
{
memmove(buffer + glen, buffer, buffersize - glen);
strncpy(buffer, GopherMethod ProtocolSeparator, glen);
}
else
{
memmove(buffer + len, buffer, buffersize - len);
strncpy(buffer, HTTPmethod ProtocolSeparator, len);
}
buffer[buffersize - 1] = 0;
}
/* If there are at least 2 unused bytes in the buffer, and the */
/* front of the string matches the HTTPMethod specifier (again */
/* this is defined at the top of this file) then search for a */
/* '/' character which isn't part of a '//' sequence. If none */
/* is found, append a '/'. This is why 2 bytes free are needed */
/* - one for the '/', one for the string terminator. */
if (strlen(buffer) < buffersize - 2 && !strncmp(buffer, HTTPmethod, shl))
{
int i, s = 0;
for (i = 0;
!s && buffer[i] && (i < (buffersize - 1));
i ++)
{
/* If we have a '/' but not a '//' sequence, mark this with s = 1 */
if (buffer[i] == '/' && buffer[i + 1] != '/') s = 1;
/* If at start of a '//' sequence, skip past it */
else if (buffer[i] == '/') i++;
}
if (!s) strcat(buffer,"/");
}
return buffer;
}
/*************************************************/
/* urlutils_load_uri_file() */
/* */
/* Loads a URI file. Will take ANT URL files */
/* too. The given buffer will be filled with a */
/* null-terminated URI from the file. */
/* */
/* The load terminates when the buffer is full */
/* except for the last byte (to allow for a */
/* forced terminator), or a control code is met */
/* in the URI file. Note that the buffer is */
/* initialised to hold null bytes before the URI */
/* file is opened. */
/* */
/* If there is an error opening the file or the */
/* file is empty, the contents of the buffer are */
/* undefined. */
/* */
/* Parameters: Pointer to the buffer; */
/* */
/* Size of the buffer; */
/* */
/* Pointer to the pathname of the */
/* URI file. */
/* */
/* Assumes: The buffer and path must NOT be */
/* the same area in memory. */
/*************************************************/
void urlutils_load_uri_file(char * buffer, size_t size, char * path)
{
FILE * fp;
int byte, counter = 0;
int type, found;
/* Warning - heavy use of 'goto's coming up shortly...! */
if (!size) return;
memset(buffer, 0, size);
/* Find the filetype */
if (
_swix(OS_File,
_INR(0,1) | _OUT(0) | _OUT(6),
23, /* Read catalogue info for named, stamped object */
path,
&found,
&type)
)
return;
/* 'found' should be 1 (a file, rather than not found, a directory, etc.). */
if (found != 1) return;
/* Don't load if not the correct type */
if (type != FileType_URI && type != FileType_URL) return;
/* Open the file */
fp = fopen(path, "rb");
if (fp)
{
byte = getc(fp);
/* For URL files, go straight to URL reading - they're just */
/* a terminated, or unterminated URL in a file. */
if (type == FileType_URL) goto urlutils_load_uri_file_read_uri;
/* If it is a URI file, it must start with 'URI'. */
if (byte != 'U') goto urlutils_load_uri_file_not_a_uri_file;
byte = getc(fp);
if (byte != 'R') goto urlutils_load_uri_file_not_a_uri_file;
byte = getc(fp);
if (byte != 'I') goto urlutils_load_uri_file_not_a_uri_file;
/* Get to first character after the 'I' */
byte = getc(fp);
/* Now find the version number - skip white space */
urlutils_load_uri_file_skip_white_space_1:
while (byte != EOF && byte < ' ') byte = getc(fp);
/* If we've hit a hash, this is a comment line - terminated by white */
/* space or end of file. */
if (byte == '#')
{
/* Now want to *find* the white space to skip the comment */
while (byte != EOF && byte >= ' ') byte = getc(fp);
/* If we're not at the end of the file, loop back and continue */
/* skipping white space. */
if (byte != EOF) goto urlutils_load_uri_file_skip_white_space_1;
}
/* By now, we're either at EOF or the start of the first real entry */
/* after 'URL', which will be the file version. URI file versions */
/* are guaranteed backwards compatible, so we can skip this bit. */
while (byte != EOF && byte >= ' ') byte = getc(fp);
/* Now we want to get to the URI, so again, skip white space */
/* and comment lines. After that, we'll be on the first byte */
/* of the URI (or end of file, if it's broken). */
urlutils_load_uri_file_skip_white_space_2:
while (byte != EOF && byte < ' ') byte = getc(fp);
if (byte == '#')
{
while (byte != EOF && byte >= ' ') byte = getc(fp);
if (byte != EOF) goto urlutils_load_uri_file_skip_white_space_2;
}
/* Load the URI component - assume everything from here up */
/* to any control chracter or EOF is part of the URI. Note */
/* that 'byte' already contains the first character. */
urlutils_load_uri_file_read_uri:
while (byte != EOF && byte >= ' ' && counter < size - 1)
{
buffer[counter++] = byte;
byte = getc(fp);
}
fclose(fp);
}
return;
/* Error condition exits */
urlutils_load_uri_file_not_a_uri_file:
/* Not a URI / URL file */
if (fp) fclose(fp);
erb.errnum = Utils_Error_Custom_Message;
StrNCpy0(erb.errmess,
lookup_token("NotAURI:This is not a valid URI file.",
0,
0));
show_error_ret(&erb);
return;
}
/*************************************************/
/* urlutils_extract_uri() */
/* */
/* Looks at a URI file loaded into a buffer, and */
/* overwrites it with the URL contents extracted */
/* from the body. */
/* */
/* Parameters: Pointer to the buffer holding the */
/* URI file; */
/* */
/* Size of the file (the buffer is */
/* assumed to be at least this size */
/* but not assumed to be larger). */
/* */
/* Returns: Contents of the buffer are */
/* updated to hold a null terminated */
/* URI. */
/*************************************************/
void urlutils_extract_uri(char * buffer, size_t file_size)
{
int copy, counter;
if (!buffer || !file_size) return;
/* Ensure it starts with 'URI' */
if (strncmp(buffer, "URI", 3)) goto urlutils_extract_uri_not_a_uri_file;
/* Point to the first byte after the 'URI' indentifier */
counter = 3;
/* Skip white space to find the version number */
urlutils_extract_uri_skip_white_space_1:
while (
counter < file_size &&
buffer[counter] &&
buffer[counter] < ' '
)
counter++;
/* If we find a comment, skip the comment body */
if (buffer[counter] == '#')
{
while (
counter < file_size &&
buffer[counter] &&
buffer[counter] >= ' '
)
counter++;
/* Now go back to skip the white space after the comment, and */
/* thus any other comments that follow. */
if (buffer[counter]) goto urlutils_extract_uri_skip_white_space_1;
}
/* Skip the file version */
while (
counter < file_size &&
buffer[counter] &&
buffer[counter] >= ' '
)
counter++;
/* Again, skip white space and comments */
urlutils_extract_uri_skip_white_space_2:
while (
counter < file_size &&
buffer[counter] &&
buffer[counter] < ' '
)
counter++;
if (buffer[counter] == '#')
{
while (
counter < file_size &&
buffer[counter] &&
buffer[counter] >= ' '
)
counter++;
if (buffer[counter]) goto urlutils_extract_uri_skip_white_space_2;
}
/* Now we're at the URI. Copy it to the start of the buffer. */
copy = 0;
while (
counter < file_size &&
buffer[counter] &&
buffer[counter] >= ' '
)
buffer[copy++] = buffer[counter++];
/* Need to make sure that the string is terminated - if we're */
/* likely to overflow the buffer, we must overwrite the last */
/* char with a terminator. You never know, it could be a non- */
/* essential last character (e.g. trailing '/'). */
if (copy && copy < file_size) buffer[copy] = 0;
else buffer[copy - 1] = 0;
return;
urlutils_extract_uri_not_a_uri_file:
/* Not a URI / URL file */
erb.errnum = Utils_Error_Custom_Message;
StrNCpy0(erb.errmess,
lookup_token("NotAURI:This is not a valid URI file.",
0,
0));
show_error_ret(&erb);
return;
}
/*************************************************/
/* urlutils_internal_extra() */
/* */
/* Returns an offset into a given string at */
/* which extra data in an internal URL may be */
/* found. */
/* */
/* Parameters: Pointer to the URL string. */
/* */
/* Returns: Offset for the extra data, or 0 */
/* if none is found. */
/*************************************************/
int urlutils_internal_extra(char * iurl)
{
char * extra;
if (strncmp(iurl, Internal_URL, Int_URL_Len)) return 0;
extra = strchr(iurl, ':');
if (!extra) return 0;
else extra ++;
return (int) (extra - iurl);
}
/*************************************************/
/* urlutils_internal_tail() */
/* */
/* Returns an offset into a given string at */
/* which tail data (typically a URL leafname) */
/* may be found. */
/* */
/* Parameters: Pointer to the URL string. */
/* */
/* Returns: Offset for the tail data, or 0 if */
/* none is found. */
/*************************************************/
int urlutils_internal_tail(char * iurl)
{
char * tail, * extra;
int exoff, found = 0;
exoff = urlutils_internal_extra(iurl);
if (!exoff) return 0;
extra = iurl + exoff;
tail = iurl + strlen(iurl); /* No '-1' here as tail is decremented early in the while loop below */
while (tail > extra && !found)
{
tail--;
if (*tail == '/') found = 1;
}
if (!found) return 0;
else tail ++;
return (int) (tail - iurl);
}
/*************************************************/
/* urlutils_set_displayed() */
/* */
/* On the basis of a given internal URL, sets */
/* the 'displayed' field of a given browser_data */
/* structure. */
/* */
/* Parameters: Pointer to a browser_data struct */
/* that is to be altered; */
/* */
/* Pointer to the internal URL. */
/*************************************************/
void urlutils_set_displayed(browser_data * b, char * iurl)
{
if (!strncmp(iurl, Internal_URL, Int_URL_Len))
{
if (!strncmp(iurl + Int_URL_Len, ForExternalHImage, strlen(ForExternalHImage))) b->displayed = Display_External_Image;
else if (!strncmp(iurl + Int_URL_Len, ForExternalNImage, strlen(ForExternalNImage))) b->displayed = Display_External_Image;
else if (!strncmp(iurl + Int_URL_Len, ForScrapFile, strlen(ForScrapFile))) b->displayed = Display_Scrap_File;
else if (!strncmp(iurl + Int_URL_Len, ForGoBack, strlen(ForGoBack))) b->displayed = Display_Previous_Page;
else if (!strncmp(iurl + Int_URL_Len, ForGoRecover, strlen(ForGoRecover))) b->displayed = Display_Recovered_Page;
else if (!strncmp(iurl + Int_URL_Len, ForGoHome, strlen(ForGoHome))) b->displayed = Display_Home_Page;
else b->displayed = Display_Fetched_Page;
}
else b->displayed = Display_Fetched_Page;
}
/*************************************************/
/* urlutils_check_protocols() */
/* */
/* Checks a given URL to see if the fetch */
/* protocol it specifies can be handled. */
/* */
/* Parameters: Pointer to the URL string. */
/* */
/* Returns: 1 if the URL can be handled (i.e. */
/* the protocol at the start of the */
/* URL matches one that the Messages */
/* file says a module which is */
/* currently running copes with), */
/* else 0. */
/*************************************************/
int urlutils_check_protocols(char * url)
{
int protocols = 0;
int i;
char p[24];
if (!url || (url && !*url)) return 0;
/* Find the number of possible protocols */
protocols = atoi(lookup_control("ProtocolMax", 1, NULL));
/* Exit if not found / not a sensible number */
if (protocols <= 0) return 0;
/* Loop round all protocols */
for (i = 1; i <= protocols; i++)
{
/* Look up the module name by building a MessageTrans */
/* token of the appropriate format, and call OS_Module */
/* 18 (lookup module) for it; if the SWI doesn't */
/* raise an error, the module was found. */
sprintf(p, "ProtocolM%d", i);
if (!_swix(OS_Module,
_INR(0,1),
18,
lookup_control(p, 1, NULL)))
{
/* Module is present, so check the protocol */
sprintf(p, "ProtocolU%d", i);
lookup_control(p, 1, NULL);
/* If the protocol identifier can be found... */
if (tokens[0] != '!')
{
/* Compare it to the same number of characters in */
/* the given URL. If it matches, we can deal with */
/* the URL. */
if (!strncmp(tokens, url, strlen(tokens))) return 1;
/* Otherwise, loop on to the next protocol... */
}
}
}
/* If we reach here, no protocol was found. */
return 0;
}
/*************************************************/
/* urlutils_dispatch() */
/* */
/* Puts a given URI into the URI queue and sends */
/* it out to the URI handler. */
/* */
/* Parameters: Pointer to a browser_data struct */
/* for which the URI relates; */
/* */
/* Pointer to null-terminated URI */
/* string (not in a movable block, */
/* so not, e.g., in a flex block); */
/* */
/* URI queue flags (see URIutils.h). */
/* */
/* Assumes: That the caller has already made */
/* sure the URI handler is present. */
/*************************************************/
_kernel_oserror * urlutils_dispatch(browser_data * b, char * uri, unsigned int flags)
{
_kernel_oserror * e;
unsigned int return_code;
uri_queue * entry;
#ifdef TRACE
if (tl & (1u<<21)) Printf("urlutils_dispatch: Called for %p with '%s'\n",b,uri);
#endif
/* Claim memory for the new entry */
entry = malloc(sizeof(uri_queue));
/* Moan if the claim failed */
if (!entry)
{
#ifdef TRACE
if (tl & (1u<<21)) Printf("urlutils_dispatch: Memory claim for queue entry failed\n",b,uri);
#endif
return make_no_fetch_memory_error(15);
}
#ifdef TRACE
malloccount += sizeof(uri_queue);
if (tl & (1u<<13)) Printf("** malloccount (urlutils_dispatch): \0211%d\0217\n",malloccount);
if (tl & (1u<<21)) Printf("urlutils_dispatch: Claimed queue entry %p\n",entry);
#endif
/* Fill in part of the entry */
entry->flags = flags;
entry->b = b;
/* If there are no entries, set uri_queue_base to the */
/* address of this one. Otherwise, point this entry's */
/* 'next' to the current base item, and point that */
/* item's 'prev' back to this entry. Then replace the */
/* current base entry with this new one. */
entry->prev = NULL;
if (!uri_queue_base) entry->next = NULL;
else
{
entry->next = uri_queue_base;
uri_queue_base->prev = entry;
}
uri_queue_base = entry;
/* Now call the URI handler and get a handle to fill in */
/* the last uri_queue field. */
e = uri_dispatch(URI_Dispatch_Inform,
uri,
task_handle,
&return_code,
NULL,
&entry->uri_handle);
if (e)
{
#ifdef TRACE
if (tl & (1u<<21)) Printf("urlutils_dispatch: Exitting with error\n");
#endif
return e;
}
/* If the request was refused complain */
if (return_code != URI_Dispatch_RequestAccepted)
{
erb.errnum = Utils_Error_Custom_Message;
StrNCpy0(erb.errmess,
lookup_token("Refused:Cannot fetch this address as the fetch request was refused by the internal handler.",
0,0));
#ifdef TRACE
if (tl & (1u<<21)) Printf("urlutils_dispatch: Exitting with error\n");
#endif
return &erb;
}
/* Otherwise exit successfully */
#ifdef TRACE
if (tl & (1u<<21)) Printf("urlutils_dispatch: Successful\n");
#endif
return NULL;
}
/*************************************************/
/* urlutils_remove_from_queue() */
/* */
/* Removes a specified entry from the list of */
/* uri_queue structures, freeing the memory */
/* allocated for it. */
/* */
/* Parameters: The URI handle of the entry. */
/*************************************************/
_kernel_oserror * urlutils_remove_from_queue(URI_handle_t uri_handle)
{
uri_queue * entry = uri_queue_base;
#ifdef TRACE
if (tl & (1u<<21)) Printf("urlutils_remove_from_queue: Called with handle %p\n", uri_handle);
#endif
/* Try to find the entry */
while (entry && entry->uri_handle != uri_handle) entry = entry->next;
#ifdef TRACE
/* Complain if not found */
if (!entry)
{
erb.errnum = Utils_Error_Custom_Normal;
sprintf(erb.errmess, "Can't find URI handle %p in URI queue", uri_handle);
if (tl & (1u<<21)) Printf("urlutils_remove_from_queue: Exitting with error\n");
return &erb;
}
#else
/* Fail silently */
if (!entry) return NULL;
#endif
if (entry->prev) entry->prev->next = entry->next;
if (entry->next) entry->next->prev = entry->prev;
if (entry == uri_queue_base) uri_queue_base = entry->next;
#ifdef TRACE
malloccount -= sizeof(uri_queue);
if (tl & (1u<<13)) Printf("** malloccount (uriutils_remove_from_queue): \0212%d\0217\n",malloccount);
if (tl & (1u<<21)) Printf("urlutils_remove_from_queue: Freeing entry %p\n",entry);
#endif
free (entry);
#ifdef TRACE
if (tl & (1u<<21)) Printf("urlutils_remove_from_queue: Successful\n");
#endif
return NULL;
}
/*************************************************/
/* urlutils_find_queue_entry() */
/* */
/* Finds an entry in the list of uri_queue */
/* structures. */
/* */
/* Parameters: The URI handle of the entry. */
/* */
/* Returns: Pointer to the entry, or NULL if */
/* no entry with that handle could */
/* be found. */
/*************************************************/
uri_queue * urlutils_find_queue_entry(URI_handle_t uri_handle)
{
uri_queue * entry = uri_queue_base;
#ifdef TRACE
if (tl & (1u<<21)) Printf("urlutils_find_queue_entry: Called with handle %p\n", uri_handle);
#endif
while (entry && entry->uri_handle != uri_handle) entry = entry->next;
#ifdef TRACE
if (tl & (1u<<21)) Printf("urlutils_find_queue_entry: Returning with entry %p\n", entry);
#endif
return entry;
}