10a48773fSEric van Gyzen /*
20a48773fSEric van Gyzen __ __ _
30a48773fSEric van Gyzen ___\ \/ /_ __ __ _| |_
40a48773fSEric van Gyzen / _ \\ /| '_ \ / _` | __|
50a48773fSEric van Gyzen | __// \| |_) | (_| | |_
60a48773fSEric van Gyzen \___/_/\_\ .__/ \__,_|\__|
70a48773fSEric van Gyzen |_| XML parser
80a48773fSEric van Gyzen
90a48773fSEric van Gyzen Copyright (c) 1997-2000 Thai Open Source Software Center Ltd
10*cc68614dSXin LI Copyright (c) 2000 Clark Cooper <coopercc@users.sourceforge.net>
11*cc68614dSXin LI Copyright (c) 2001-2002 Fred L. Drake, Jr. <fdrake@users.sourceforge.net>
12*cc68614dSXin LI Copyright (c) 2006 Karl Waclawek <karl@waclawek.net>
13*cc68614dSXin LI Copyright (c) 2016-2017 Sebastian Pipping <sebastian@pipping.org>
14*cc68614dSXin LI Copyright (c) 2017 Rhodri James <rhodri@wildebeest.org.uk>
150a48773fSEric van Gyzen Licensed under the MIT license:
160a48773fSEric van Gyzen
170a48773fSEric van Gyzen Permission is hereby granted, free of charge, to any person obtaining
180a48773fSEric van Gyzen a copy of this software and associated documentation files (the
190a48773fSEric van Gyzen "Software"), to deal in the Software without restriction, including
200a48773fSEric van Gyzen without limitation the rights to use, copy, modify, merge, publish,
210a48773fSEric van Gyzen distribute, sublicense, and/or sell copies of the Software, and to permit
220a48773fSEric van Gyzen persons to whom the Software is furnished to do so, subject to the
230a48773fSEric van Gyzen following conditions:
240a48773fSEric van Gyzen
250a48773fSEric van Gyzen The above copyright notice and this permission notice shall be included
260a48773fSEric van Gyzen in all copies or substantial portions of the Software.
270a48773fSEric van Gyzen
280a48773fSEric van Gyzen THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
290a48773fSEric van Gyzen EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
300a48773fSEric van Gyzen MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN
310a48773fSEric van Gyzen NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,
320a48773fSEric van Gyzen DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
330a48773fSEric van Gyzen OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
340a48773fSEric van Gyzen USE OR OTHER DEALINGS IN THE SOFTWARE.
355bb6a25fSPoul-Henning Kamp */
365bb6a25fSPoul-Henning Kamp
375bb6a25fSPoul-Henning Kamp #include <sys/types.h>
385bb6a25fSPoul-Henning Kamp #include <sys/mman.h>
395bb6a25fSPoul-Henning Kamp #include <sys/stat.h>
405bb6a25fSPoul-Henning Kamp #include <fcntl.h>
415bb6a25fSPoul-Henning Kamp #include <errno.h>
425bb6a25fSPoul-Henning Kamp #include <string.h>
435bb6a25fSPoul-Henning Kamp #include <stdio.h>
445bb6a25fSPoul-Henning Kamp #include <unistd.h>
455bb6a25fSPoul-Henning Kamp
465bb6a25fSPoul-Henning Kamp #ifndef MAP_FILE
475bb6a25fSPoul-Henning Kamp # define MAP_FILE 0
485bb6a25fSPoul-Henning Kamp #endif
495bb6a25fSPoul-Henning Kamp
500a48773fSEric van Gyzen #include "xmltchar.h"
515bb6a25fSPoul-Henning Kamp #include "filemap.h"
525bb6a25fSPoul-Henning Kamp
530a48773fSEric van Gyzen #ifdef XML_UNICODE_WCHAR_T
540a48773fSEric van Gyzen # define XML_FMT_STR "ls"
550a48773fSEric van Gyzen #else
560a48773fSEric van Gyzen # define XML_FMT_STR "s"
570a48773fSEric van Gyzen #endif
580a48773fSEric van Gyzen
595bb6a25fSPoul-Henning Kamp int
filemap(const tchar * name,void (* processor)(const void *,size_t,const tchar *,void * arg),void * arg)600a48773fSEric van Gyzen filemap(const tchar *name,
610a48773fSEric van Gyzen void (*processor)(const void *, size_t, const tchar *, void *arg),
626b2c1e49SXin LI void *arg) {
635bb6a25fSPoul-Henning Kamp int fd;
645bb6a25fSPoul-Henning Kamp size_t nbytes;
655bb6a25fSPoul-Henning Kamp struct stat sb;
665bb6a25fSPoul-Henning Kamp void *p;
675bb6a25fSPoul-Henning Kamp
680a48773fSEric van Gyzen fd = topen(name, O_RDONLY);
695bb6a25fSPoul-Henning Kamp if (fd < 0) {
700a48773fSEric van Gyzen tperror(name);
715bb6a25fSPoul-Henning Kamp return 0;
725bb6a25fSPoul-Henning Kamp }
735bb6a25fSPoul-Henning Kamp if (fstat(fd, &sb) < 0) {
740a48773fSEric van Gyzen tperror(name);
755bb6a25fSPoul-Henning Kamp close(fd);
765bb6a25fSPoul-Henning Kamp return 0;
775bb6a25fSPoul-Henning Kamp }
785bb6a25fSPoul-Henning Kamp if (! S_ISREG(sb.st_mode)) {
795bb6a25fSPoul-Henning Kamp close(fd);
800a48773fSEric van Gyzen fprintf(stderr, "%" XML_FMT_STR ": not a regular file\n", name);
815bb6a25fSPoul-Henning Kamp return 0;
825bb6a25fSPoul-Henning Kamp }
830a48773fSEric van Gyzen if (sb.st_size > XML_MAX_CHUNK_LEN) {
840a48773fSEric van Gyzen close(fd);
850a48773fSEric van Gyzen return 2; /* Cannot be passed to XML_Parse in one go */
860a48773fSEric van Gyzen }
875bb6a25fSPoul-Henning Kamp
885bb6a25fSPoul-Henning Kamp nbytes = sb.st_size;
89220ed979SColeman Kane /* mmap fails for zero length files */
90220ed979SColeman Kane if (nbytes == 0) {
91220ed979SColeman Kane static const char c = '\0';
92220ed979SColeman Kane processor(&c, 0, name, arg);
93220ed979SColeman Kane close(fd);
94220ed979SColeman Kane return 1;
95220ed979SColeman Kane }
966b2c1e49SXin LI p = (void *)mmap((void *)0, (size_t)nbytes, PROT_READ, MAP_FILE | MAP_PRIVATE,
976b2c1e49SXin LI fd, (off_t)0);
985bb6a25fSPoul-Henning Kamp if (p == (void *)-1) {
990a48773fSEric van Gyzen tperror(name);
1005bb6a25fSPoul-Henning Kamp close(fd);
1015bb6a25fSPoul-Henning Kamp return 0;
1025bb6a25fSPoul-Henning Kamp }
1035bb6a25fSPoul-Henning Kamp processor(p, nbytes, name, arg);
104be8aff81SXin LI munmap((void *)p, nbytes);
1055bb6a25fSPoul-Henning Kamp close(fd);
1065bb6a25fSPoul-Henning Kamp return 1;
1075bb6a25fSPoul-Henning Kamp }
108