/* returns TRUE on error */
static gboolean
zip_child_init (GsfInfileZip *child, GError **errmsg)
{
static guint8 const header_signature[] =
{ 'P', 'K', 0x03, 0x04 };
guint8 const *data = NULL;
guint16 name_len, extras_len;
char *err = NULL;
GsfZipDirent *dirent = child->vdir->dirent;
/* skip local header
* should test tons of other info, but trust that those are correct
**/
if (gsf_input_seek (child->source, (gsf_off_t) dirent->offset, G_SEEK_SET))
err = g_strdup_printf ("Error seeking to zip header @ %" GSF_OFF_T_FORMAT,
dirent->offset);
else if (NULL == (data = gsf_input_read (child->source, ZIP_FILE_HEADER_SIZE, NULL)))
err = g_strdup_printf ("Error reading %d bytes in zip header", ZIP_FILE_HEADER_SIZE);
else if (0 != memcmp (data, header_signature, sizeof (header_signature))) {
err = g_strdup_printf ("Error incorrect zip header @ %" GSF_OFF_T_FORMAT,
dirent->offset);
g_print ("Header is :\n");
gsf_mem_dump (data, sizeof (header_signature));
g_print ("Header should be :\n");
gsf_mem_dump (header_signature, sizeof (header_signature));
}
if (NULL != err) {
if (errmsg != NULL)
*errmsg = g_error_new_literal (gsf_input_error_id (), 0, err);
g_free (err);
return TRUE;
}
name_len = GSF_LE_GET_GUINT16 (data + ZIP_FILE_HEADER_NAME_SIZE);
extras_len = GSF_LE_GET_GUINT16 (data + ZIP_FILE_HEADER_EXTRAS_SIZE);
dirent->data_offset = dirent->offset + ZIP_FILE_HEADER_SIZE + name_len + extras_len;
child->restlen = dirent->usize;
child->crestlen = dirent->csize;
if (dirent->compr_method != GSF_ZIP_STORED) {
int err;
if (!child->stream)
child->stream = g_new0 (z_stream, 1);
err = inflateInit2 (child->stream, -MAX_WBITS);
if (err != Z_OK) {
if (errmsg != NULL)
*errmsg = g_error_new (gsf_input_error_id (), 0,
"problem uncompressing stream");
return TRUE;
}
}
return FALSE;
}
static void
dump_biff_stream (GsfInput *stream)
{
guint8 const *data;
guint16 len, opcode;
unsigned pos = gsf_input_tell (stream);
while (NULL != (data = gsf_input_read (stream, 4, NULL))) {
gboolean enable_dump = TRUE;
opcode = GSF_LE_GET_GUINT16 (data);
len = GSF_LE_GET_GUINT16 (data+2);
if (len > 15000) {
enable_dump = TRUE;
g_warning ("Suspicious import of biff record > 15,000 (0x%x) for opcode 0x%hx",
len, opcode);
} else if ((opcode & 0xff00) > 0x1000) {
enable_dump = TRUE;
g_warning ("Suspicious import of biff record with opcode 0x%hx",
opcode);
}
if (enable_dump)
printf ("Opcode 0x%3hx : %15s, length 0x%hx (=%hd)\n",
opcode, get_biff_opcode_name (opcode),
len, len);
if (len > 0) {
data = gsf_input_read (stream, len, NULL);
if (data == NULL)
break;
if (enable_dump)
gsf_mem_dump (data, len);
}
pos = gsf_input_tell (stream);
}
}
static void
qpro_parse_formula (QProReadState *state, int col, int row,
guint8 const *data, guint8 const *end)
{
guint16 magic, ref_offset;
#if 0
int flags = GSF_LE_GET_GUINT16 (data + 8);
int length = GSF_LE_GET_GUINT16 (data + 10);
#endif
GnmValue *val;
GSList *stack = NULL;
GnmExprTop const *texpr = NULL;
guint8 const *refs, *fmla;
#ifdef DEBUG_MISSING
dump_missing_functions ();
#endif
Q_CHECK_CONDITION (end - data >= 14);
magic = GSF_LE_GET_GUINT16 (data + 6) & 0x7ff8;
ref_offset = GSF_LE_GET_GUINT16 (data + 12);
fmla = data + 14;
refs = fmla + ref_offset;
Q_CHECK_CONDITION (refs <= end);
#if 0
puts (cell_coord_name (col, row));
gsf_mem_dump (data, 14);
gsf_mem_dump (fmla, refs-fmla);
gsf_mem_dump (refs, end-refs);
#endif
while (fmla < refs && *fmla != QPRO_OP_EOF) {
QProOperators op = *fmla++;
GnmExpr const *expr = NULL;
#if 0
g_print ("Operator %d.\n", op);
#endif
switch (op) {
case QPRO_OP_CONST_FLOAT:
Q_CHECK_CONDITION (refs - fmla >= 8);
expr = gnm_expr_new_constant (value_new_float (
gsf_le_get_double (fmla)));
fmla += 8;
break;
case QPRO_OP_CELLREF: {
GnmCellRef ref;
guint16 tmp;
Q_CHECK_CONDITION (end - refs >= 6);
tmp = GSF_LE_GET_GUINT16 (refs + 4);
ref.sheet = NULL;
ref.col = *((gint8 *)(refs + 2));
ref.col_relative = (tmp & 0x4000) ? TRUE : FALSE;
ref.row_relative = (tmp & 0x2000) ? TRUE : FALSE;
if (ref.row_relative)
ref.row = (int)(((gint16)((tmp & 0x1fff) << 3)) >> 3);
else
ref.row = tmp & 0x1fff;
expr = gnm_expr_new_cellref (&ref);
refs += 6;
break;
}
case QPRO_OP_RANGEREF: {
GnmCellRef a, b;
guint16 tmp;
Q_CHECK_CONDITION (end - refs >= 10);
tmp = GSF_LE_GET_GUINT16 (refs + 4);
a.sheet = NULL;
a.col = *((gint8 *)(refs + 2));
a.col_relative = (tmp & 0x4000) ? TRUE : FALSE;
a.row_relative = (tmp & 0x2000) ? TRUE : FALSE;
if (a.row_relative)
a.row = (int)(((gint16)((tmp & 0x1fff) << 3)) >> 3);
else
a.row = tmp & 0x1fff;
tmp = GSF_LE_GET_GUINT16 (refs + 8);
b.sheet = NULL;
b.col = *((gint8 *)(refs + 6));
b.col_relative = (tmp & 0x4000) ? TRUE : FALSE;
b.row_relative = (tmp & 0x2000) ? TRUE : FALSE;
if (b.row_relative)
b.row = (int)(((gint16)((tmp & 0x1fff) << 3)) >> 3);
else
b.row = tmp & 0x1fff;
expr = gnm_expr_new_constant (
value_new_cellrange_unsafe (&a, &b));
refs += 10;
break;
}
static int
test (unsigned argc, char *argv[])
{
static char const * const stream_names[] = {
"Workbook", "WORKBOOK", "workbook",
"Book", "BOOK", "book"
};
GsfInput *input, *stream, *pcache_dir;
GsfInfile *infile;
GError *err = NULL;
unsigned i, j;
for (i = 1 ; i < argc ; i++) {
fprintf( stderr, "%s\n",argv[i]);
input = gsf_input_mmap_new (argv[i], NULL);
if (input == NULL) /* Only report error if stdio fails too */
input = gsf_input_stdio_new (argv[i], &err);
if (input == NULL) {
g_return_val_if_fail (err != NULL, 1);
g_warning ("'%s' error: %s", argv[i], err->message);
g_error_free (err);
err = NULL;
continue;
}
input = gsf_input_uncompress (input);
infile = gsf_infile_msole_new (input, &err);
if (infile == NULL) {
g_return_val_if_fail (err != NULL, 1);
g_warning ("'%s' Not an OLE file: %s", argv[i], err->message);
g_error_free (err);
err = NULL;
#ifdef DUMP_CONTENT
dump_biff_stream (input);
#endif
g_object_unref (G_OBJECT (input));
continue;
}
#if 0
stream = gsf_infile_child_by_name (infile, "\01CompObj");
if (stream != NULL) {
gsf_off_t len = gsf_input_size (stream);
guint8 const *data = gsf_input_read (stream, len, NULL);
if (data != NULL)
gsf_mem_dump (data, len);
g_object_unref (G_OBJECT (stream));
}
return 0;
#endif
stream = gsf_infile_child_by_name (infile, "\05SummaryInformation");
if (stream != NULL) {
GsfDocMetaData *meta_data = gsf_doc_meta_data_new ();
puts ( "SummaryInfo");
err = gsf_doc_meta_data_read_from_msole (meta_data, stream);
if (err != NULL) {
g_warning ("'%s' error: %s", argv[i], err->message);
g_error_free (err);
err = NULL;
} else
gsf_doc_meta_dump (meta_data);
g_object_unref (meta_data);
g_object_unref (G_OBJECT (stream));
}
stream = gsf_infile_child_by_name (infile, "\05DocumentSummaryInformation");
if (stream != NULL) {
GsfDocMetaData *meta_data = gsf_doc_meta_data_new ();
puts ( "DocSummaryInfo");
err = gsf_doc_meta_data_read_from_msole (meta_data, stream);
if (err != NULL) {
g_warning ("'%s' error: %s", argv[i], err->message);
g_error_free (err);
err = NULL;
} else
gsf_doc_meta_dump (meta_data);
g_object_unref (meta_data);
g_object_unref (G_OBJECT (stream));
}
for (j = 0 ; j < G_N_ELEMENTS (stream_names) ; j++) {
stream = gsf_infile_child_by_name (infile, stream_names[j]);
if (stream != NULL) {
puts (j < 3 ? "Excel97" : "Excel95");
#ifdef DUMP_CONTENT
dump_biff_stream (stream);
#endif
g_object_unref (G_OBJECT (stream));
break;
}
}
#ifdef DUMP_CONTENT
pcache_dir = gsf_infile_child_by_name (infile, "_SX_DB_CUR"); /* Excel 97 */
if (NULL == pcache_dir)
pcache_dir = gsf_infile_child_by_name (infile, "_SX_DB"); /* Excel 95 */
if (NULL != pcache_dir) {
int i, n = gsf_infile_num_children (infile);
for (i = 0 ; i < n ; i++) {
stream = gsf_infile_child_by_index (GSF_INFILE (pcache_dir), i);
if (stream != NULL) {
printf ("=================================================\nPivot cache '%04hX'\n\n", i);
dump_biff_stream (stream);
g_object_unref (G_OBJECT (stream));
}
}
g_object_unref (G_OBJECT (pcache_dir));
}
#endif
g_object_unref (G_OBJECT (infile));
g_object_unref (G_OBJECT (input));
}
return 0;
}
/**
* vba_dir_read:
* @vba: #GsfInfileMSVBA
* @err: (allow-none): place to store a #GError if anything goes wrong
*
* Read an VBA dirctory and its project file.
* along the way.
*
* Returns: %FALSE on error setting @err if it is supplied.
**/
static gboolean
vba_dir_read (GsfInfileMSVBA *vba, GError **err)
{
int inflated_size, element_count = -1;
char const *msg = NULL;
char *name, *elem_stream = NULL;
guint32 len;
guint16 tag;
guint8 *inflated_data, *end, *ptr;
GsfInput *dir;
gboolean failed = TRUE;
/* 0. get the stream */
dir = gsf_infile_child_by_name (vba->source, "dir");
if (dir == NULL) {
msg = _("Can't find the VBA directory stream");
goto fail_stream;
}
/* 1. decompress it */
ptr = inflated_data = gsf_vba_inflate (dir, 0, &inflated_size, TRUE);
if (inflated_data == NULL)
goto fail_compression;
end = inflated_data + inflated_size;
/* 2. GUESS : based on several xls with macros and XL8GARY this looks like a
* series of sized records. Be _extra_ careful */
do {
/* I have seen
* type len data
* 1 4 1 0 0 0
* 2 4 9 4 0 0
* 3 2 4 e4
* 4 <var> project name
* 5 0
* 6 0
* 7 4
* 8 4
* 0x3d 0
* 0x40 0
* 0x14 4 9 4 0 0
*
* 0x0f == number of elements
* 0x1c == (Size 0)
* 0x1e == (Size 4)
* 0x48 == (Size 0)
* 0x31 == stream offset of the compressed source !
*
* 0x16 == an ascii dependency name
* 0x3e == a unicode dependency name
* 0x33 == a classid for a dependency with no trialing data
*
* 0x2f == a dummy classid
* 0x30 == a classid
* 0x0d == the classid
* 0x2f, and 0x0d appear contain
* uint32 classid_size;
* <classid>
* 00 00 00 00 00 00
* and sometimes some trailing junk
**/
if ((ptr + 6) > end) {
msg = _("vba project header problem");
goto fail_content;
}
tag = GSF_LE_GET_GUINT16 (ptr);
len = GSF_LE_GET_GUINT32 (ptr + 2);
ptr += 6;
if ((ptr + len) > end) {
msg = _("vba project header problem");
goto fail_content;
}
switch (tag) {
case 4:
name = g_strndup (ptr, len);
#ifdef OLD_VBA_DUMP
puts ("<?xml version=\"1.0\" encoding=\"UTF-8\"?>");
printf ("<project name=\"%s\">", name);
#endif
g_free (name);
break;
case 9:
/* this seems to have an extra two bytes that are not
* part of the length ..?? */
len += 2;
break;
case 0xf :
if (len != 2) {
g_warning ("element count is not what we expected");
break;
}
if (element_count >= 0) {
g_warning ("More than one element count ??");
break;
}
element_count = GSF_LE_GET_GUINT16 (ptr);
break;
/* dependencies */
case 0x0d : break;
case 0x2f : break;
case 0x30 : break;
case 0x33 : break;
case 0x3e : break;
case 0x16:
#if 0
name = g_strndup (ptr, len);
g_print ("Depend Name : '%s'\n", name);
g_free (name);
#endif
break;
/* elements */
case 0x47 : break;
case 0x32 : break;
case 0x1a:
#if 0
name = g_strndup (ptr, len);
g_print ("Element Name : '%s'\n", name);
g_free (name);
#endif
break;
case 0x19:
g_free (elem_stream);
elem_stream = g_strndup (ptr, len);
break;
case 0x31:
if (len != 4) {
g_warning ("source offset property is not what we expected");
break;
}
vba_extract_module_source (vba, elem_stream,
GSF_LE_GET_GUINT32 (ptr));
g_free (elem_stream); elem_stream = NULL;
element_count--;
break;
default :
#if 0
g_print ("tag %hx : len %u\n", tag, len);
gsf_mem_dump (ptr, len);
#endif
break;
}
ptr += len;
} while (tag != 0x10);
if (element_count != 0)
g_warning ("Number of elements differs from expectations");
failed = FALSE;
fail_content :
g_free (inflated_data);
#ifdef OLD_VBA_DUMP
puts ("</project>");
#endif
fail_compression :
g_object_unref (dir);
fail_stream :
g_free (elem_stream);
if (failed) {
if (err != NULL)
*err = g_error_new_literal (gsf_input_error_id (), 0, msg);
return FALSE;
}
return TRUE;
}
