InputSource SourceFromStream(const char8 *description, FILE *file)
{
Entity e;
e = NewExternalEntity("",0,description,0,0);
if(!strchr8(description, '/'))
EntitySetBaseURL(e, default_base_url());
return NewInputSource(e, MakeFILE16FromFILE(file, "r"));
}
InputSource SourceFromFILE16(const char8 *description, FILE16 *file16)
{
Entity e;
e = NewExternalEntity(0, 0, description, 0, 0);
if(!strchr8(description, '/'))
{
char8 *base = default_base_url();
EntitySetBaseURL(e, base);
Free(base);
}
return NewInputSource(e, file16);
}
char *url_merge(const char *url, const char *base,
char **_scheme, char **_host, int *_port, char **_path)
{
char *merged_scheme, *merged_host, *merged_path, *merged_url;
char *scheme=0, *host=0, *path=0;
char *base_scheme=0, *base_host=0, *base_path=0;
char *default_base=0;
int port, base_port, merged_port, i, j;
char *p;
/* First see if we have an absolute URL */
parse_url(url, &scheme, &host, &port, &path);
if(scheme && (host || *path == '/'))
{
merged_scheme = scheme;
merged_host = host;
merged_port = port;
merged_path = path;
merged_url = strdup8(url);
goto ok;
}
/* Relative URL, so we need the base URL */
if(!base)
base = default_base = default_base_url();
parse_url(base, &base_scheme, &base_host, &base_port, &base_path);
if(base_scheme && (base_host || *base_path == '/'))
;
else
{
LT_ERROR1(LEFILE, "Error: bad base URL <%s>\n", base);
goto bad;
}
/* Determine merged path */
if(path[0] == '/')
{
/* not relative, use as-is */
merged_path = path;
path = 0;
}
else
{
/* relative, append to base path */
merged_path = Malloc(strlen(base_path) + strlen(path) + 1);
strcpy(merged_path, base_path);
/* strip last component of base */
for(i=strlen(merged_path)-1; i>=0 && merged_path[i] != '/'; i--)
merged_path[i] = '\0';
/* append relative path */
strcat(merged_path, path);
/* Remove . and .. components from path */
p = merged_path;
for(i=0; p[i]; )
{
assert(p[i] == '/');
/* find next segment */
for(j=i+1; p[j] && p[j] != '/'; j++)
;
/* Do we have "." ? */
if(j - i == 2 && p[i+1] == '.')
{
strcpy(&p[i+1], p[j] ? &p[j+1] : &p[j]);
continue;
}
/* Do we have "<segment>/.." with <segment> != ".." ? */
/* (We know we're not looking at "./" so we don't have to
* worry about "./..")
*/
if(p[j] == '/' && p[j+1] == '.' && p[j+2] == '.' &&
(p[j+3] == '/' || p[j+3] == '\0') &&
(j - i != 3 || p[i+1] != '.' || p[i+2] != '.'))
{
strcpy(&p[i+1], p[j+3] ? &p[j+4] : &p[j+3]);
i = 0; /* start again from beginning */
continue;
}
/* move to next segment */
i = j;
}
}
/* Check for deviant relative URLs like file:foo */
if(scheme && !host && *path != '/')
{
if(strcmp(scheme, base_scheme) == 0)
{
WARN1(LEFILE,
"Warning: relative URL <%s> contains scheme, contrary to RFC 1808\n",
url);
}
else
{
LT_ERROR2(LEFILE,
"Error: relative URL <%s> has scheme different from base <%s>\n",
url, base);
goto bad;
}
}
/* Return the parts and the whole thing */
merged_scheme = base_scheme; if(scheme) Free(scheme);
if(host)
{
merged_host = host; Free(base_host);
merged_port = port;
}
else
{
merged_host = base_host;
merged_port = base_port;
}
Free(path); Free(base_path);
merged_url = Malloc(strlen(merged_scheme) + 1 +
(merged_host ? 2 + strlen(merged_host) + 10 : 0) +
strlen(merged_path) + 1);
if(merged_host)
{
if(merged_port == -1)
sprintf(merged_url, "%s://%s%s",
merged_scheme, merged_host, merged_path);
else
sprintf(merged_url, "%s://%s:%d%s",
merged_scheme, merged_host, merged_port, merged_path);
}
else
sprintf(merged_url, "%s:%s", merged_scheme, merged_path);
ok:
Free(default_base);
if(_scheme) *_scheme = merged_scheme; else Free(merged_scheme);
if(_host) *_host = merged_host; else Free(merged_host);
if(_port) *_port = merged_port;
if(_path) *_path = merged_path; else Free(merged_path);
return merged_url;
bad:
Free(default_base);
Free(scheme);
Free(host);
Free(path);
Free(base_scheme);
Free(base_host);
Free(base_path);
return NULL;
}
