diff --git a/Makefile.am b/Makefile.am index b8cbb4a6..3941baa7 100644 --- a/Makefile.am +++ b/Makefile.am @@ -6,7 +6,7 @@ SUBDIRS = include . doc example xstc @PYTHON_SUBDIR@ DIST_SUBDIRS = include . doc example python xstc -INCLUDES = -I$(top_builddir)/include -I@srcdir@/include @THREAD_CFLAGS@ @Z_CFLAGS@ +INCLUDES = -I$(top_builddir)/include -I@srcdir@/include @THREAD_CFLAGS@ @Z_CFLAGS@ @LZMA_CFLAGS@ noinst_PROGRAMS=testSchemas testRelax testSAX testHTML testXPath testURI \ testThreads testC14N testAutomata testRegexp \ @@ -18,7 +18,7 @@ bin_PROGRAMS = xmllint xmlcatalog bin_SCRIPTS=xml2-config lib_LTLIBRARIES = libxml2.la -libxml2_la_LIBADD = @THREAD_LIBS@ @Z_LIBS@ $(ICONV_LIBS) @M_LIBS@ @WIN32_EXTRA_LIBADD@ +libxml2_la_LIBADD = @THREAD_LIBS@ @Z_LIBS@ @LZMA_LIBS@ $(ICONV_LIBS) @M_LIBS@ @WIN32_EXTRA_LIBADD@ if USE_VERSION_SCRIPT LIBXML2_VERSION_SCRIPT = $(VERSION_SCRIPT_FLAGS)$(srcdir)/libxml2.syms @@ -54,7 +54,7 @@ libxml2_la_SOURCES = SAX.c entities.c encoding.c error.c parserInternals.c \ endif DEPS = $(top_builddir)/libxml2.la -LDADDS = @STATIC_BINARIES@ $(top_builddir)/libxml2.la @THREAD_LIBS@ @Z_LIBS@ $(ICONV_LIBS) @M_LIBS@ @WIN32_EXTRA_LIBADD@ +LDADDS = @STATIC_BINARIES@ $(top_builddir)/libxml2.la @THREAD_LIBS@ @Z_LIBS@ @LZMA_LIBS@ $(ICONV_LIBS) @M_LIBS@ @WIN32_EXTRA_LIBADD@ man_MANS = xml2-config.1 libxml.3 diff --git a/configure.in b/configure.in index 5dca51ab..fa803750 100644 --- a/configure.in +++ b/configure.in @@ -197,6 +197,14 @@ AC_ARG_WITH(zlib, LDFLAGS="${LDFLAGS} -L$withval/lib" fi ]) +AC_ARG_WITH(lzma, +[ --with-lzma[[=DIR]] use liblzma in DIR],[ + if test "$withval" != "no" -a "$withval" != "yes"; then + LZMA_DIR=$withval + CPPFLAGS="${CPPFLAGS} -I$withval/include" + LDFLAGS="${LDFLAGS} -L$withval/lib" + fi +]) AC_ARG_WITH(coverage, [ --with-coverage build for code coverage with GCC (off)]) @@ -399,6 +407,30 @@ AC_SUBST(Z_CFLAGS) AC_SUBST(Z_LIBS) AC_SUBST(WITH_ZLIB) +echo Checking lzma + +dnl Checks for lzma library. + +WITH_LZMA=0 +if test "$with_lzma" = "no"; then + echo "Disabling compression support" +else + AC_CHECK_HEADERS(lzma.h, + AC_CHECK_LIB(lzma, lzma_code,[ + AC_DEFINE([HAVE_LIBLZMA], [1], [Have compression library]) + WITH_LZMA=1 + if test "x${LZMA_DIR}" != "x"; then + LZMA_CFLAGS="-I${LZMA_DIR}/include" + LZMA_LIBS="-L${LZMA_DIR}/lib -llzma" + else + LZMA_LIBS="-llzma" + fi])) +fi + +AC_SUBST(LZMA_CFLAGS) +AC_SUBST(LZMA_LIBS) +AC_SUBST(WITH_LZMA) + CPPFLAGS=${_cppflags} LDFLAGS=${_ldflags} diff --git a/include/libxml/parser.h b/include/libxml/parser.h index aabb96cf..54f16608 100644 --- a/include/libxml/parser.h +++ b/include/libxml/parser.h @@ -1224,6 +1224,7 @@ typedef enum { XML_WITH_DEBUG_RUN = 30, XML_WITH_ZLIB = 31, XML_WITH_ICU = 32, + XML_WITH_LZMA = 33, XML_WITH_NONE = 99999 /* just to be sure of allocation size */ } xmlFeature; diff --git a/include/libxml/xmlversion.h.in b/include/libxml/xmlversion.h.in index 95856541..ddd46339 100644 --- a/include/libxml/xmlversion.h.in +++ b/include/libxml/xmlversion.h.in @@ -391,6 +391,15 @@ XMLPUBFUN void XMLCALL xmlCheckVersion(int version); #define LIBXML_ZLIB_ENABLED #endif +/** + * LIBXML_LZMA_ENABLED: + * + * Whether the Lzma support is compiled in + */ +#if @WITH_LZMA@ +#define LIBXML_LZMA_ENABLED +#endif + #ifdef __GNUC__ #ifdef HAVE_ANSIDECL_H #include diff --git a/parser.c b/parser.c index c55e41df..72a467ca 100644 --- a/parser.c +++ b/parser.c @@ -79,6 +79,9 @@ #ifdef HAVE_ZLIB_H #include #endif +#ifdef HAVE_LZMA_H +#include +#endif static void xmlFatalErr(xmlParserCtxtPtr ctxt, xmlParserErrors error, const char *info); @@ -954,6 +957,12 @@ xmlHasFeature(xmlFeature feature) return(1); #else return(0); +#endif + case XML_WITH_LZMA: +#ifdef LIBXML_LZMA_ENABLED + return(1); +#else + return(0); #endif case XML_WITH_ICU: #ifdef LIBXML_ICU_ENABLED diff --git a/xmlIO.c b/xmlIO.c index 8fc00e3c..7a6fb0b8 100644 --- a/xmlIO.c +++ b/xmlIO.c @@ -35,6 +35,9 @@ #ifdef HAVE_ZLIB_H #include #endif +#ifdef HAVE_LZMA_H +#include +#endif #if defined(WIN32) || defined(_WIN32) #include @@ -1306,6 +1309,807 @@ xmlGzfileClose (void * context) { } #endif /* HAVE_ZLIB_H */ +#ifdef HAVE_LZMA_H +/************************************************************************ + * * + * I/O for compressed file accesses * + * * + ************************************************************************/ +typedef void* xzFile; /* opaque lzma file descriptor */ +static xzFile xzopen(const char *path, const char *mode); +static xzFile xzdopen(int fd, const char *mode); +static int xzread(xzFile file, void *buf, unsigned len); +static int xzclose(xzFile file); +/* values for xz_state how */ +#define LOOK 0 /* look for a gzip/lzma header */ +#define COPY 1 /* copy input directly */ +#define GZIP 2 /* decompress a gzip stream */ +#define LZMA 3 /* decompress a lzma stream */ +/* internal lzma file state data structure */ +typedef struct { + int mode; /* see lzma modes above */ + int fd; /* file descriptor */ + char *path; /* path or fd for error messages */ + uint64_t pos; /* current position in uncompressed data */ + unsigned size; /* buffer size, zero if not allocated yet */ + unsigned want; /* requested buffer size, default is BUFSIZ */ + unsigned char *in; /* input buffer */ + unsigned char *out; /* output buffer (double-sized when reading) */ + unsigned char *next; /* next output data to deliver or write */ + unsigned have; /* amount of output data unused at next */ + int eof; /* true if end of input file reached */ + uint64_t start; /* where the lzma data started, for rewinding */ + uint64_t raw; /* where the raw data started, for seeking */ + int how; /* 0: get header, 1: copy, 2: decompress */ + int direct; /* true if last read direct, false if lzma */ + /* seek request */ + uint64_t skip; /* amount to skip (already rewound if backwards) */ + int seek; /* true if seek request pending */ + /* error information */ + int err; /* error code */ + char *msg; /* error message */ + /* lzma stream */ + lzma_stream strm; /* stream structure in-place (not a pointer) */ +#ifdef HAVE_ZLIB_H + /* zlib inflate or deflate stream */ + z_stream zstrm; /* stream structure in-place (not a pointer) */ +#endif +} xz_state, *xz_statep; + +static void xz_error(xz_statep state, int err, const char *msg) +{ + /* free previously allocated message and clear */ + if (state->msg != NULL) { + if (state->err != LZMA_MEM_ERROR) + free(state->msg); + state->msg = NULL; + } + + /* set error code, and if no message, then done */ + state->err = err; + if (msg == NULL) + return; + + /* for an out of memory error, save as static string */ + if (err == LZMA_MEM_ERROR) { + state->msg = (char *)msg; + return; + } + + /* construct error message with path */ + if ((state->msg = malloc(strlen(state->path) + strlen(msg) + 3)) == NULL) { + state->err = LZMA_MEM_ERROR; + state->msg = (char *)"out of memory"; + return; + } + strcpy(state->msg, state->path); + strcat(state->msg, ": "); + strcat(state->msg, msg); + return; +} + +static void xz_reset(xz_statep state) +{ + state->have = 0; /* no output data available */ + state->eof = 0; /* not at end of file */ + state->how = LOOK; /* look for gzip header */ + state->direct = 1; /* default for empty file */ + state->seek = 0; /* no seek request pending */ + xz_error(state, LZMA_OK, NULL); /* clear error */ + state->pos = 0; /* no uncompressed data yet */ + state->strm.avail_in = 0; /* no input data yet */ +#ifdef HAVE_ZLIB_H + state->zstrm.avail_in = 0; /* no input data yet */ +#endif +} + +static xzFile xz_open(const char *path, int fd, const char *mode ATTRIBUTE_UNUSED) +{ + xz_statep state; + + /* allocate xzFile structure to return */ + state = malloc(sizeof(xz_state)); + if (state == NULL) + return NULL; + state->size = 0; /* no buffers allocated yet */ + state->want = BUFSIZ; /* requested buffer size */ + state->msg = NULL; /* no error message yet */ + + /* save the path name for error messages */ + state->path = malloc(strlen(path) + 1); + if (state->path == NULL) { + free(state); + return NULL; + } + strcpy(state->path, path); + + /* open the file with the appropriate mode (or just use fd) */ + state->fd = fd != -1 ? fd : + open(path, +#ifdef O_LARGEFILE + O_LARGEFILE | +#endif +#ifdef O_BINARY + O_BINARY | +#endif + O_RDONLY, + 0666); + if (state->fd == -1) { + free(state->path); + free(state); + return NULL; + } + + /* save the current position for rewinding (only if reading) */ + state->start = lseek(state->fd, 0, SEEK_CUR); + if (state->start == (uint64_t) -1) state->start = 0; + + /* initialize stream */ + xz_reset(state); + + /* return stream */ + return (xzFile)state; +} + +xzFile xzopen(const char *path, const char *mode) +{ + return xz_open(path, -1, mode); +} + +xzFile xzdopen(int fd, const char *mode) +{ + char *path; /* identifier for error messages */ + xzFile xz; + + if (fd == -1 || (path = malloc(7 + 3 * sizeof(int))) == NULL) + return NULL; + sprintf(path, "", fd); /* for debugging */ + xz = xz_open(path, fd, mode); + free(path); + return xz; +} + +static int xz_load(xz_statep state, unsigned char *buf, unsigned len, unsigned *have) +{ + int ret; + + *have = 0; + do { + ret = read(state->fd, buf + *have, len - *have); + if (ret <= 0) + break; + *have += ret; + } while (*have < len); + if (ret < 0) { + xz_error(state, -1, strerror(errno)); + return -1; + } + if (ret == 0) + state->eof = 1; + return 0; +} + +static int xz_avail(xz_statep state) +{ + lzma_stream *strm = &(state->strm); + + if (state->err != LZMA_OK) + return -1; + if (state->eof == 0) { + if (xz_load(state, state->in, state->size, + (unsigned *)&(strm->avail_in)) == -1) + return -1; + strm->next_in = state->in; + } + return 0; +} + +static int +is_format_xz(xz_statep state) +{ + lzma_stream *strm = &(state->strm); + + return strm->avail_in >= 6 && memcmp(state->in, "\3757zXZ", 6) == 0; +} + +static int +is_format_lzma(xz_statep state) +{ + lzma_stream *strm = &(state->strm); + + lzma_filter filter; + lzma_options_lzma *opt; + uint32_t dict_size; + uint64_t uncompressed_size; + size_t i; + + if (strm->avail_in < 13) + return 0; + + filter.id = LZMA_FILTER_LZMA1; + if (lzma_properties_decode(&filter, NULL, state->in, 5) != LZMA_OK) + return 0; + + opt = filter.options; + dict_size = opt->dict_size; + free(opt); + + /* A hack to ditch tons of false positives: We allow only dictionary + * sizes that are 2^n or 2^n + 2^(n-1) or UINT32_MAX. LZMA_Alone + * created only files with 2^n, but accepts any dictionary size. + * If someone complains, this will be reconsidered. + */ + if (dict_size != UINT32_MAX) { + uint32_t d = dict_size - 1; + d |= d >> 2; + d |= d >> 3; + d |= d >> 4; + d |= d >> 8; + d |= d >> 16; + ++d; + if (d != dict_size || dict_size == 0) + return 0; + } + + /* Another hack to ditch false positives: Assume that if the + * uncompressed size is known, it must be less than 256 GiB. + * Again, if someone complains, this will be reconsidered. + */ + uncompressed_size = 0; + for (i = 0; i < 8; ++i) + uncompressed_size |= (uint64_t)(state->in[5 + i]) << (i * 8); + + if (uncompressed_size != UINT64_MAX + && uncompressed_size > (UINT64_C(1) << 38)) + return 0; + + return 1; +} + +#ifdef HAVE_ZLIB_H +/* Get next byte from input, or -1 if end or error. */ +#define NEXT() ((strm->avail_in == 0 && xz_avail(state) == -1) ? -1 : \ + (strm->avail_in == 0 ? -1 : \ + (strm->avail_in--, *(strm->next_in)++))) + +/* Get a four-byte little-endian integer and return 0 on success and the value + in *ret. Otherwise -1 is returned and *ret is not modified. */ +static int gz_next4(xz_statep state, unsigned long *ret) +{ + int ch; + unsigned long val; + z_streamp strm = &(state->zstrm); + + val = NEXT(); + val += (unsigned)NEXT() << 8; + val += (unsigned long)NEXT() << 16; + ch = NEXT(); + if (ch == -1) + return -1; + val += (unsigned long)ch << 24; + *ret = val; + return 0; +} +#endif + +static int xz_head(xz_statep state) +{ + lzma_stream *strm = &(state->strm); + lzma_stream init = LZMA_STREAM_INIT; + int flags; + unsigned len; + + /* allocate read buffers and inflate memory */ + if (state->size == 0) { + /* allocate buffers */ + state->in = malloc(state->want); + state->out = malloc(state->want << 1); + if (state->in == NULL || state->out == NULL) { + if (state->out != NULL) + free(state->out); + if (state->in != NULL) + free(state->in); + xz_error(state, LZMA_MEM_ERROR, "out of memory"); + return -1; + } + state->size = state->want; + + /* allocate decoder memory */ + state->strm = init; + state->strm.avail_in = 0; + state->strm.next_in = NULL; + if (lzma_auto_decoder(&state->strm, UINT64_MAX, 0) != LZMA_OK) { + free(state->out); + free(state->in); + state->size = 0; + xz_error(state, LZMA_MEM_ERROR, "out of memory"); + return -1; + } + +#ifdef HAVE_ZLIB_H + /* allocate inflate memory */ + state->zstrm.zalloc = Z_NULL; + state->zstrm.zfree = Z_NULL; + state->zstrm.opaque = Z_NULL; + state->zstrm.avail_in = 0; + state->zstrm.next_in = Z_NULL; + if (inflateInit2(&(state->zstrm), -15) != Z_OK) { /* raw inflate */ + free(state->out); + free(state->in); + state->size = 0; + xz_error(state, LZMA_MEM_ERROR, "out of memory"); + return -1; + } +#endif + } + + /* get some data in the input buffer */ + if (strm->avail_in == 0) { + if (xz_avail(state) == -1) + return -1; + if (strm->avail_in == 0) + return 0; + } + + /* look for the xz magic header bytes */ + if (is_format_xz(state) || is_format_lzma(state)) { + state->how = LZMA; + state->direct = 0; + return 0; + } + +#ifdef HAVE_ZLIB_H + /* look for the gzip magic header bytes 31 and 139 */ + if (strm->next_in[0] == 31) { + strm->avail_in--; + strm->next_in++; + if (strm->avail_in == 0 && xz_avail(state) == -1) + return -1; + if (strm->avail_in && strm->next_in[0] == 139) { + /* we have a gzip header, woo hoo! */ + strm->avail_in--; + strm->next_in++; + + /* skip rest of header */ + if (NEXT() != 8) { /* compression method */ + xz_error(state, LZMA_DATA_ERROR, "unknown compression method"); + return -1; + } + flags = NEXT(); + if (flags & 0xe0) { /* reserved flag bits */ + xz_error(state, LZMA_DATA_ERROR, "unknown header flags set"); + return -1; + } + NEXT(); /* modification time */ + NEXT(); + NEXT(); + NEXT(); + NEXT(); /* extra flags */ + NEXT(); /* operating system */ + if (flags & 4) { /* extra field */ + len = (unsigned)NEXT(); + len += (unsigned)NEXT() << 8; + while (len--) + if (NEXT() < 0) + break; + } + if (flags & 8) /* file name */ + while (NEXT() > 0) + ; + if (flags & 16) /* comment */ + while (NEXT() > 0) + ; + if (flags & 2) { /* header crc */ + NEXT(); + NEXT(); + } + /* an unexpected end of file is not checked for here -- it will be + noticed on the first request for uncompressed data */ + + /* set up for decompression */ + inflateReset(&state->zstrm); + state->zstrm.adler = crc32(0L, Z_NULL, 0); + state->how = GZIP; + state->direct = 0; + return 0; + } + else { + /* not a gzip file -- save first byte (31) and fall to raw i/o */ + state->out[0] = 31; + state->have = 1; + } + } +#endif + + /* doing raw i/o, save start of raw data for seeking, copy any leftover + input to output -- this assumes that the output buffer is larger than + the input buffer, which also assures space for gzungetc() */ + state->raw = state->pos; + state->next = state->out; + if (strm->avail_in) { + memcpy(state->next + state->have, strm->next_in, strm->avail_in); + state->have += strm->avail_in; + strm->avail_in = 0; + } + state->how = COPY; + state->direct = 1; + return 0; +} + +static int xz_decomp(xz_statep state) +{ + int ret; + unsigned had; + unsigned long crc, len; + lzma_stream *strm = &(state->strm); + + lzma_action action = LZMA_RUN; + + /* fill output buffer up to end of deflate stream */ + had = strm->avail_out; + do { + /* get more input for inflate() */ + if (strm->avail_in == 0 && xz_avail(state) == -1) + return -1; + if (strm->avail_in == 0) { + xz_error(state, LZMA_DATA_ERROR, "unexpected end of file"); + return -1; + } + if (state->eof) + action = LZMA_FINISH; + + /* decompress and handle errors */ +#ifdef HAVE_ZLIB_H + if (state->how == GZIP) { + state->zstrm.avail_in = (uInt) state->strm.avail_in; + state->zstrm.next_in = (Bytef*) state->strm.next_in; + state->zstrm.avail_out = (uInt) state->strm.avail_out; + state->zstrm.next_out = (Bytef*) state->strm.next_out; + ret = inflate(&state->zstrm, Z_NO_FLUSH); + if (ret == Z_STREAM_ERROR || ret == Z_NEED_DICT) { + xz_error(state, Z_STREAM_ERROR, + "internal error: inflate stream corrupt"); + return -1; + } + if (ret == Z_MEM_ERROR) ret = LZMA_MEM_ERROR; + if (ret == Z_DATA_ERROR) ret = LZMA_DATA_ERROR; + if (ret == Z_STREAM_END) ret = LZMA_STREAM_END; + state->strm.avail_in = state->zstrm.avail_in; + state->strm.next_in = state->zstrm.next_in; + state->strm.avail_out = state->zstrm.avail_out; + state->strm.next_out = state->zstrm.next_out; + } + else /* state->how == LZMA */ +#endif + ret = lzma_code(strm, action); + if (ret == LZMA_MEM_ERROR) { + xz_error(state, LZMA_MEM_ERROR, "out of memory"); + return -1; + } + if (ret == LZMA_DATA_ERROR) { + xz_error(state, LZMA_DATA_ERROR, "compressed data error"); + return -1; + } + } while (strm->avail_out && ret != LZMA_STREAM_END); + + /* update available output and crc check value */ + state->have = had - strm->avail_out; + state->next = strm->next_out - state->have; +#ifdef HAVE_ZLIB_H + state->zstrm.adler = crc32(state->zstrm.adler, state->next, state->have); +#endif + + if (ret == LZMA_STREAM_END) { +#ifdef HAVE_ZLIB_H + if (state->how == GZIP) { + if (gz_next4(state, &crc) == -1 || gz_next4(state, &len) == -1) { + xz_error(state, LZMA_DATA_ERROR, "unexpected end of file"); + return -1; + } + if (crc != state->zstrm.adler) { + xz_error(state, LZMA_DATA_ERROR, "incorrect data check"); + return -1; + } + if (len != (state->zstrm.total_out & 0xffffffffL)) { + xz_error(state, LZMA_DATA_ERROR, "incorrect length check"); + return -1; + } + state->strm.avail_in = 0; + state->strm.next_in = NULL; + state->strm.avail_out = 0; + state->strm.next_out = NULL; + } + else +#endif + if ( strm->avail_in != 0 || !state->eof) { + xz_error(state, LZMA_DATA_ERROR, "trailing garbage"); + return -1; + } + state->how = LOOK; /* ready for next stream, once have is 0 (leave + state->direct unchanged to remember how) */ + } + + /* good decompression */ + return 0; +} + +static int xz_make(xz_statep state) +{ + lzma_stream *strm = &(state->strm); + + if (state->how == LOOK) { /* look for lzma / gzip header */ + if (xz_head(state) == -1) + return -1; + if (state->have) /* got some data from xz_head() */ + return 0; + } + if (state->how == COPY) { /* straight copy */ + if (xz_load(state, state->out, state->size << 1, &(state->have)) == -1) + return -1; + state->next = state->out; + } + else if (state->how == LZMA || state->how == GZIP) { /* decompress */ + strm->avail_out = state->size << 1; + strm->next_out = state->out; + if (xz_decomp(state) == -1) + return -1; + } + return 0; +} + +static int xz_skip(xz_statep state, uint64_t len) +{ + unsigned n; + + /* skip over len bytes or reach end-of-file, whichever comes first */ + while (len) + /* skip over whatever is in output buffer */ + if (state->have) { + n = (uint64_t)state->have > len ? + (unsigned)len : state->have; + state->have -= n; + state->next += n; + state->pos += n; + len -= n; + } + + /* output buffer empty -- return if we're at the end of the input */ + else if (state->eof && state->strm.avail_in == 0) + break; + + /* need more data to skip -- load up output buffer */ + else { + /* get more output, looking for header if required */ + if (xz_make(state) == -1) + return -1; + } + return 0; +} + +int xzread(xzFile file, void *buf, unsigned len) +{ + unsigned got, n; + xz_statep state; + lzma_stream *strm; + + /* get internal structure */ + if (file == NULL) + return -1; + state = (xz_statep)file; + strm = &(state->strm); + + /* check that we're reading and that there's no error */ + if (state->err != LZMA_OK) + return -1; + + /* since an int is returned, make sure len fits in one, otherwise return + with an error (this avoids the flaw in the interface) */ + if ((int)len < 0) { + xz_error(state, LZMA_BUF_ERROR, "requested length does not fit in int"); + return -1; + } + + /* if len is zero, avoid unnecessary operations */ + if (len == 0) + return 0; + + /* process a skip request */ + if (state->seek) { + state->seek = 0; + if (xz_skip(state, state->skip) == -1) + return -1; + } + + /* get len bytes to buf, or less than len if at the end */ + got = 0; + do { + /* first just try copying data from the output buffer */ + if (state->have) { + n = state->have > len ? len : state->have; + memcpy(buf, state->next, n); + state->next += n; + state->have -= n; + } + + /* output buffer empty -- return if we're at the end of the input */ + else if (state->eof && strm->avail_in == 0) + break; + + /* need output data -- for small len or new stream load up our output + buffer */ + else if (state->how == LOOK || len < (state->size << 1)) { + /* get more output, looking for header if required */ + if (xz_make(state) == -1) + return -1; + continue; /* no progress yet -- go back to memcpy() above */ + /* the copy above assures that we will leave with space in the + output buffer, allowing at least one gzungetc() to succeed */ + } + + /* large len -- read directly into user buffer */ + else if (state->how == COPY) { /* read directly */ + if (xz_load(state, buf, len, &n) == -1) + return -1; + } + + /* large len -- decompress directly into user buffer */ + else { /* state->how == LZMA */ + strm->avail_out = len; + strm->next_out = buf; + if (xz_decomp(state) == -1) + return -1; + n = state->have; + state->have = 0; + } + + /* update progress */ + len -= n; + buf = (char *)buf + n; + got += n; + state->pos += n; + } while (len); + + /* return number of bytes read into user buffer (will fit in int) */ + return (int)got; +} + +int xzclose(xzFile file) +{ + int ret; + xz_statep state; + + /* get internal structure */ + if (file == NULL) + return LZMA_DATA_ERROR; + state = (xz_statep)file; + + /* free memory and close file */ + if (state->size) { + lzma_end(&(state->strm)); +#ifdef HAVE_LIBZ_H + inflateEnd(&(state->zstrm)); +#endif + free(state->out); + free(state->in); + } + free(state->path); + ret = close(state->fd); + free(state); + return ret ? ret : LZMA_OK; +} + +/** + * xmlXzfileMatch: + * @filename: the URI for matching + * + * input from compressed file test + * + * Returns 1 if matches, 0 otherwise + */ +static int +xmlXzfileMatch (const char *filename ATTRIBUTE_UNUSED) { + return(1); +} + +/** + * xmlXzFileOpen_real: + * @filename: the URI for matching + * + * input from compressed file open + * if @filename is " " then the standard input is used + * + * Returns an I/O context or NULL in case of error + */ +static void * +xmlXzfileOpen_real (const char *filename) { + const char *path = NULL; + xzFile fd; + + if (!strcmp(filename, "-")) { + fd = xzdopen(dup(0), "rb"); + return((void *) fd); + } + + if (!xmlStrncasecmp(BAD_CAST filename, BAD_CAST "file://localhost/", 17)) { + path = &filename[16]; + } else if (!xmlStrncasecmp(BAD_CAST filename, BAD_CAST "file:///", 8)) { + path = &filename[7]; + } else if (!xmlStrncasecmp(BAD_CAST filename, BAD_CAST "file:/", 6)) { + /* lots of generators seems to lazy to read RFC 1738 */ + path = &filename[5]; + } else + path = filename; + + if (path == NULL) + return(NULL); + if (!xmlCheckFilename(path)) + return(NULL); + + fd = xzopen(path, "rb"); + return((void *) fd); +} + +/** + * xmlXzfileOpen: + * @filename: the URI for matching + * + * Wrapper around xmlXzfileOpen_real that try it with an unescaped + * version of @filename, if this fails fallback to @filename + * + * Returns a handler or NULL in case or failure + */ +static void * +xmlXzfileOpen (const char *filename) { + char *unescaped; + void *retval; + + retval = xmlXzfileOpen_real(filename); + if (retval == NULL) { + unescaped = xmlURIUnescapeString(filename, 0, NULL); + if (unescaped != NULL) { + retval = xmlXzfileOpen_real(unescaped); + } + xmlFree(unescaped); + } + + return retval; +} + +/** + * xmlXzfileRead: + * @context: the I/O context + * @buffer: where to drop data + * @len: number of bytes to write + * + * Read @len bytes to @buffer from the compressed I/O channel. + * + * Returns the number of bytes written + */ +static int +xmlXzfileRead (void * context, char * buffer, int len) { + int ret; + + ret = xzread((xzFile) context, &buffer[0], len); + if (ret < 0) xmlIOErr(0, "xzread()"); + return(ret); +} + +/** + * xmlXzfileClose: + * @context: the I/O context + * + * Close a compressed I/O channel + */ +static int +xmlXzfileClose (void * context) { + int ret; + + ret = (xzclose((xzFile) context) == LZMA_OK ) ? 0 : -1; + if (ret < 0) xmlIOErr(0, "xzclose()"); + return(ret); +} +#endif /* HAVE_LZMA_H */ + #ifdef LIBXML_HTTP_ENABLED /************************************************************************ * * @@ -2181,6 +2985,10 @@ xmlRegisterDefaultInputCallbacks(void) { xmlRegisterInputCallbacks(xmlGzfileMatch, xmlGzfileOpen, xmlGzfileRead, xmlGzfileClose); #endif /* HAVE_ZLIB_H */ +#ifdef HAVE_LZMA_H + xmlRegisterInputCallbacks(xmlXzfileMatch, xmlXzfileOpen, + xmlXzfileRead, xmlXzfileClose); +#endif /* HAVE_ZLIB_H */ #ifdef LIBXML_HTTP_ENABLED xmlRegisterInputCallbacks(xmlIOHTTPMatch, xmlIOHTTPOpen, diff --git a/xmllint.c b/xmllint.c index 14c64bec..2acc7c58 100644 --- a/xmllint.c +++ b/xmllint.c @@ -2958,6 +2958,7 @@ static void showVersion(const char *name) { if (xmlHasFeature(XML_WITH_DEBUG_MEM)) fprintf(stderr, "MemDebug "); if (xmlHasFeature(XML_WITH_DEBUG_RUN)) fprintf(stderr, "RunDebug "); if (xmlHasFeature(XML_WITH_ZLIB)) fprintf(stderr, "Zlib "); + if (xmlHasFeature(XML_WITH_LZMA)) fprintf(stderr, "Lzma "); fprintf(stderr, "\n"); }