/* 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;
}
PangoAttrList *
ms_container_read_markup (MSContainer const *c,
guint8 const *data, size_t txo_len,
char const *str)
{
TXORun txo_run;
size_t str_len;
XL_CHECK_CONDITION_VAL (txo_len >= 16,
pango_attr_list_new ()); /* min two records */
str_len = g_utf8_strlen (str, -1);
txo_run.last = G_MAXINT;
txo_run.accum = NULL;
for (txo_len -= 16 ; (gssize)txo_len >= 0 ; txo_len -= 8) {
guint16 o = GSF_LE_GET_GUINT16 (data + txo_len);
guint16 idx = GSF_LE_GET_GUINT16 (data + txo_len + 2);
XL_CHECK_CONDITION_VAL (o <= str_len, txo_run.accum);
txo_run.first = g_utf8_offset_to_pointer (str, o) - str;
XL_CHECK_CONDITION_VAL (txo_run.first < txo_run.last, txo_run.accum);
if (idx != 0) {
if (!txo_run.accum)
txo_run.accum = pango_attr_list_new ();
pango_attr_list_filter
(ms_container_get_markup (c, idx),
(PangoAttrFilterFunc) append_txorun,
&txo_run);
}
txo_run.last = txo_run.first;
}
return txo_run.accum;
}
static guint8 const *
qpro_get_record (QProReadState *state, guint16 *id, guint16 *len)
{
guint8 const *data;
data = gsf_input_read (state->input, 4, NULL);
Q_CHECK_CONDITION (data != NULL);
*id = GSF_LE_GET_GUINT16 (data + 0);
*len = GSF_LE_GET_GUINT16 (data + 2);
#if 0
g_printerr ("%hd with %hd\n", *id, *len);
#endif
if (*len == 0)
return "";
data = gsf_input_read (state->input, *len, NULL);
switch (*id) {
case QPRO_UNDOCUMENTED_837:
case QPRO_UNDOCUMENTED_907:
break; /* Nothing. */
default:
Q_CHECK_CONDITION (*len < 0x2000);
}
Q_CHECK_CONDITION (data != NULL);
return data;
error:
return NULL;
}
static GsfZipDirent *
zip_dirent_new_in (GsfInfileZip *zip, gsf_off_t *offset)
{
static guint8 const dirent_signature[] =
{ 'P', 'K', 0x01, 0x02 };
GsfZipDirent *dirent;
guint8 const *data;
guint16 name_len, extras_len, comment_len, compr_method, flags;
guint32 crc32, csize, usize, off;
gchar *name;
/* Read data and check the header */
if (gsf_input_seek (zip->source, *offset, G_SEEK_SET) ||
NULL == (data = gsf_input_read (zip->source, ZIP_DIRENT_SIZE, NULL)) ||
0 != memcmp (data, dirent_signature, sizeof (dirent_signature))) {
return NULL;
}
name_len = GSF_LE_GET_GUINT16 (data + ZIP_DIRENT_NAME_SIZE);
extras_len = GSF_LE_GET_GUINT16 (data + ZIP_DIRENT_EXTRAS_SIZE);
comment_len = GSF_LE_GET_GUINT16 (data + ZIP_DIRENT_COMMENT_SIZE);
flags = GSF_LE_GET_GUINT32 (data + ZIP_DIRENT_FLAGS);
compr_method = GSF_LE_GET_GUINT16 (data + ZIP_DIRENT_COMPR_METHOD);
crc32 = GSF_LE_GET_GUINT32 (data + ZIP_DIRENT_CRC32);
csize = GSF_LE_GET_GUINT32 (data + ZIP_DIRENT_CSIZE);
usize = GSF_LE_GET_GUINT32 (data + ZIP_DIRENT_USIZE);
off = GSF_LE_GET_GUINT32 (data + ZIP_DIRENT_OFFSET);
if ((data = gsf_input_read (zip->source, name_len, NULL)) == NULL)
return NULL;
name = g_new (gchar, (gulong) (name_len + 1));
memcpy (name, data, name_len);
name[name_len] = '\0';
dirent = gsf_zip_dirent_new ();
dirent->name = name;
dirent->flags = flags;
dirent->compr_method = compr_method;
dirent->crc32 = crc32;
dirent->csize = csize;
dirent->usize = usize;
dirent->offset = off;
#if 0
g_print ("%s = 0x%x @ %" GSF_OFF_T_FORMAT "\n", name, off, *offset);
#endif
*offset += ZIP_DIRENT_SIZE + name_len + extras_len + comment_len;
return dirent;
}
/** Given a data pointer, returns a color string (like cccccc for a medium gray).
* throws BOGUS_DOCUMENT on error. */
UT_String makeColor(UT_uint8* aData, UT_uint32 aDataLen) {
// from tools/source/generic/color.cxx line 183ff
#define COL_NAME_USER (0x8000)
#define COL_RED_1B (0x0001)
#define COL_RED_2B (0x0002)
#define COL_GREEN_1B (0x0010)
#define COL_GREEN_2B (0x0020)
#define COL_BLUE_1B (0x0100)
#define COL_BLUE_2B (0x0200)
UT_String rv;
if (aDataLen < 2)
throw UT_IE_BOGUSDOCUMENT;
UT_uint16 colorName = GSF_LE_GET_GUINT16(aData);
if (colorName & COL_NAME_USER) {
// XXX TODO. Awaiting reply in mailinglist about what CompressMode is.
return "000000";
}
else {
if (colorName < COLOR_SIZE) {
UT_String_sprintf(rv, "%02x%02x%02x", gColors[colorName].red,
gColors[colorName].green, gColors[colorName].blue);
return rv;
}
else {
UT_DEBUGMSG(("SDW: COLOR OUT OF RANGE! has num %u\n", colorName));
return "000000";
}
}
}
static gboolean
qpro_check_signature (GsfInput *input)
{
guint8 const *header;
guint16 version;
if (gsf_input_seek (input, 0, G_SEEK_SET) ||
NULL == (header = gsf_input_read (input, 2+2+2, NULL)) ||
GSF_LE_GET_GUINT16 (header + 0) != 0 ||
GSF_LE_GET_GUINT16 (header + 2) != 2)
return FALSE;
version = GSF_LE_GET_GUINT16 (header + 4);
return (version == 0x1001 || /* 'WB1' format, documented */
version == 0x1002 || /* 'WB2' format, documented */
version == 0x1006 || /* qpro 6.0 ?? */
version == 0x1007); /* qpro 7.0 ?? */
}
/**
* gsf_vba_inflate:
* @input: stream to read from
* @offset: offset into it for start byte of compressed stream
* @size: size of the returned array
* @add_null_terminator: whenever add or not null at the end of array
*
* Decompresses VBA stream.
*
* Return value: A pointer to guint8 array
**/
guint8 *
gsf_vba_inflate (GsfInput *input, gsf_off_t offset, int *size, gboolean add_null_terminator)
{
guint8 sig;
GByteArray *res;
gsf_off_t length;
res = g_byte_array_new();
gsf_input_read (input, 1, &sig);
if (1 != sig) /* should start with 0x01 */
return NULL;
offset++;
length = gsf_input_size (input);
while (offset < length) {
GsfInput *chunk;
guint16 chunk_hdr;
guint8 const *tmp;
tmp = gsf_input_read (input, 2, NULL);
if (!tmp)
break;
chunk_hdr = GSF_LE_GET_GUINT16 (tmp);
offset += 2;
if (0xB000 == (chunk_hdr&0xF000) && (chunk_hdr&0xFFF) > 0 && (length - offset < 4094)){
if (length < offset + (chunk_hdr&0xFFF))
break;
chunk = gsf_input_proxy_new_section (input, offset, (gsf_off_t) (chunk_hdr&0xFFF) + 1);
offset += (chunk_hdr&0xFFF) + 1;
} else {
if (length < offset + 4094){
chunk = gsf_input_proxy_new_section (input, offset, length-offset);
offset = length;
} else {
chunk = gsf_input_proxy_new_section (input, offset, 4094);
offset += 4094;
}
}
if (chunk) {
GByteArray *tmpres = gsf_msole_inflate (chunk, 0);
gsf_input_seek (input, offset, G_SEEK_CUR);
g_byte_array_append (res, tmpres->data, tmpres->len);
g_byte_array_free (tmpres, TRUE);
g_object_unref (chunk);
}
}
if (res == NULL)
return NULL;
if (add_null_terminator)
g_byte_array_append (res, "", 1);
*size = res->len;
return g_byte_array_free (res, 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);
}
}
void streamRead(GsfInput* aStream, TextAttr& aAttr, gsf_off_t aEoa)
throw(UT_Error)
{
UT_uint8 flags;
gsf_off_t newPos;
readFlagRec(aStream, flags, &newPos);
streamRead(aStream, aAttr.which);
streamRead(aStream, aAttr.ver);
if (flags & 0x10) {
aAttr.startSet = true;
streamRead(aStream, aAttr.start);
}
else
aAttr.startSet = false;
if (flags & 0x20) {
aAttr.endSet = true;
streamRead(aStream, aAttr.end);
}
else
aAttr.endSet = false;
if (gsf_input_seek(aStream, newPos, G_SEEK_SET))
throw UT_IE_BOGUSDOCUMENT;
gsf_off_t curPos = gsf_input_tell(aStream);
if (curPos != aEoa) {
// there is data
aAttr.dataLen = aEoa - curPos;
aAttr.data = new UT_uint8[aAttr.dataLen];
streamRead(aStream, aAttr.data, aAttr.dataLen);
}
// LIST OF THE VALUES: http://ooo.ximian.com/lxr/source/sw/sw/inc/hintids.hxx#086
// together with http://ooo.ximian.com/lxr/source/sw/sw/source/core/sw3io/sw3fmts.cxx#172
switch (aAttr.which) {
case 0x1004: // strikethrough
aAttr.attrName = "text-decoration";
if (!aAttr.data || aAttr.data[0])
aAttr.attrVal = "line-through";
else
aAttr.isOff = true;
break;
case 0x1005: { // sub-/superscript
if (aAttr.dataLen < 3)
break;
// first byte is size of text % of normal size
UT_sint16 height = GSF_LE_GET_GINT16(aAttr.data + 1);
aAttr.attrName = "text-position";
if (height > 0)
aAttr.attrVal = "superscript";
else if (height < 0)
aAttr.attrVal = "subscript";
else
aAttr.isOff = true;
break; }
case 0x1006: { // font family
if (!aAttr.data || aAttr.dataLen < 7)
// 7 = 3 byte family etc., 2 byte name length, 2 byte style length
break;
aAttr.attrName = "font-family";
// XXX TODO This code here assumes that the font names are in latin1
UT_uint16 fontLen = GSF_LE_GET_GUINT16(aAttr.data + 3);
UT_String_sprintf(aAttr.attrVal, "%.*s", fontLen, (aAttr.data + 5));
break; }
case 0x1007: // font height
// structure: | height (2 byte, twips) | prop (?) (2 byte) (if version >= 2, if ver=1 1 byte) | unit (if version>=2) |
// XXX we ignore "prop" and unit for now, they seem not used much
aAttr.attrName = "font-size";
if (aAttr.data)
aAttr.attrVal = twipsToSizeString(GSF_LE_GET_GUINT16(aAttr.data));
break;
case 0x100a: // Italic
aAttr.attrName = "font-style";
if (!aAttr.data || aAttr.data[0]) // if there is data, first byte must be != 0
// abiword doesn't support oblique, so always set italic
aAttr.attrVal = "italic";
else
aAttr.isOff = true;
break;
case 0x100d: // Underline
aAttr.attrName = "text-decoration";
if (!aAttr.data || aAttr.data[0])
aAttr.attrVal = "underline";
else
aAttr.isOff = true;
break;
case 0x100e: // Bold
aAttr.attrName = "font-weight";
if (!aAttr.data || aAttr.data[0] >= 8) // 8=Bold.
aAttr.attrVal = "bold";
else
aAttr.isOff = true;
break;
case 0x4000: // line spacing
aAttr.attrName = "line-height";
aAttr.isPara = true;
// prop space (s8) | inter space (s16) | height (u16) | rule (s8) | interrule (s8)
if (aAttr.data && aAttr.dataLen >= 7) {
// Abiword wants it as float value, StarOffice saves as percentage (e.g.
// 150 for 1.5)
float proportionalLineSpace = float(aAttr.data[0])/100;
// But maybe we need to use the height - stored as twips, need points
// (used for "exact" and "minimum" line spacing)
// XXX inter-line spacing not supported by abiword (would be rule=0x00
// interrule=0x02, value to use=inter space, unit twips)
UT_String lineHeight = twipsToSizeString(GSF_LE_GET_GINT16(aAttr.data + 3));
// We'll turn the bytes at 5 and 6 into a single integer, for easier
// evaluation
switch (GSF_LE_GET_GUINT16(aAttr.data + 5)) {
case 0x0100: // proportional
aAttr.attrVal = std_size_string(proportionalLineSpace);
break;
case 0x0001:
case 0x0002:
aAttr.attrVal = lineHeight;
if (aAttr.data[5] == 2) // "minimum" case
aAttr.attrVal += '+';
break;
default:
UT_DEBUGMSG(("Unsupported linespacing: %02x %02x\n", aAttr.data[5], aAttr.data[6]));
}
}
break;
case 0x4001: // Alignment
aAttr.attrName = "text-align";
aAttr.isPara = true;
if (aAttr.data) {
switch (aAttr.data[0]) {
case 0:
aAttr.attrVal = "left";
break;
case 1:
aAttr.attrVal = "right";
break;
case 2:
case 4: // BLOCKLINE!? what's BLOCKLINE? I'm guessing justify.
aAttr.attrVal = "justify";
break;
case 3:
aAttr.attrVal = "center";
break;
}
}
break;
case 0x4005: {// Tabstops
aAttr.attrName = "tabstops";
aAttr.isPara = true;
// Data structure:
// Count(8) | Position (in twips) (32) | Adjustment (8) | Decimal Separator (?) (8) | Fill character (8)
// (total size per tab = 7)
// UT_sint8 count = aAttr.data[0];
for (UT_uint32 i = 1; (i + 6) < aAttr.dataLen; i += 7) {
// Abiword wants: 12.3cm/L0, where 0 indicates what to fill with
UT_uint16 posInTwips = GSF_LE_GET_GUINT32(aAttr.data + i);
UT_String pos = twipsToSizeString(posInTwips);
aAttr.attrVal += pos;
aAttr.attrVal += '/';
if (aAttr.data[i + 4] < sizeof(sTabAlignment)/sizeof(*sTabAlignment))
aAttr.attrVal += sTabAlignment[aAttr.data[i + 4]];
else
aAttr.attrVal += 'L'; // fallback
char fillIndex = '0';
// Fill character
switch (aAttr.data[i + 6]) {
case '.':
fillIndex = '1';
break;
case '-':
fillIndex = '2';
break;
case '_':
fillIndex = '3';
break;
case ' ':
fillIndex = '0';
break;
default:
UT_DEBUGMSG(("Filling with '%c' is not supported\n", aAttr.data[i + 6]));
}
aAttr.attrVal += fillIndex;
aAttr.attrVal += ',';
}
}
break;
default:
UT_DEBUGMSG(("SDW: unknown attribute 0x%x, compressed %d\n", aAttr.which, lcl_sw3io__CompressWhich(aAttr.which)));
}
}
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 gboolean
check_header (GsfInputGZip *input)
{
if (input->raw) {
input->header_size = 0;
input->trailer_size = 0;
} else {
static guint8 const signature[2] = {0x1f, 0x8b};
guint8 const *data;
unsigned flags, len;
/* Check signature */
if (NULL == (data = gsf_input_read (input->source, 2 + 1 + 1 + 6, NULL)) ||
0 != memcmp (data, signature, sizeof (signature)))
return TRUE;
/* verify flags and compression type */
flags = data[3];
if (data[2] != Z_DEFLATED || (flags & ~GZIP_HEADER_FLAGS) != 0)
return TRUE;
/* If we have the size, don't bother seeking to the end. */
if (input->uncompressed_size < 0) {
/* Get the uncompressed size */
if (gsf_input_seek (input->source, (gsf_off_t) -4, G_SEEK_END) ||
NULL == (data = gsf_input_read (input->source, 4, NULL)))
return TRUE;
/* FIXME, but how? The size read here is modulo 2^32. */
input->uncompressed_size = GSF_LE_GET_GUINT32 (data);
if (input->uncompressed_size / 1000 > gsf_input_size (input->source)) {
g_warning ("Suspiciously well compressed file with better than 1000:1 ratio.\n"
"It is probably truncated or corrupt");
}
}
if (gsf_input_seek (input->source, 2 + 1 + 1 + 6, G_SEEK_SET))
return TRUE;
if (flags & GZIP_EXTRA_FIELD) {
if (NULL == (data = gsf_input_read (input->source, 2, NULL)))
return TRUE;
len = GSF_LE_GET_GUINT16 (data);
if (NULL == gsf_input_read (input->source, len, NULL))
return TRUE;
}
if (flags & GZIP_ORIGINAL_NAME) {
/* Skip over the filename (which is in ISO 8859-1 encoding). */
do {
if (NULL == (data = gsf_input_read (input->source, 1, NULL)))
return TRUE;
} while (*data != 0);
}
if (flags & GZIP_HAS_COMMENT) {
/* Skip over the comment (which is in ISO 8859-1 encoding). */
do {
if (NULL == (data = gsf_input_read (input->source, 1, NULL)))
return TRUE;
} while (*data != 0);
}
if (flags & GZIP_HEADER_CRC &&
NULL == (data = gsf_input_read (input->source, 2, NULL)))
return TRUE;
input->header_size = input->source->cur_offset;
/* the last 8 bytes are the crc and size. */
input->trailer_size = 8;
}
gsf_input_set_size (GSF_INPUT (input), input->uncompressed_size);
if (gsf_input_remaining (input->source) < input->trailer_size)
return TRUE; /* No room for payload */
return FALSE;
}
/**
* vba_project_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_project_read (GsfInfileMSVBA *vba, GError **err)
{
/* NOTE : This seems constant, find some confirmation */
static guint8 const signature[] = { 0xcc, 0x61 };
static struct {
guint8 const signature[4];
char const * const name;
int const vba_version;
gboolean const is_mac;
} const versions [] = {
{ { 0x5e, 0x00, 0x00, 0x01 }, "Office 97", 5, FALSE},
{ { 0x5f, 0x00, 0x00, 0x01 }, "Office 97 SR1", 5, FALSE },
{ { 0x65, 0x00, 0x00, 0x01 }, "Office 2000 alpha?", 6, FALSE },
{ { 0x6b, 0x00, 0x00, 0x01 }, "Office 2000 beta?", 6, FALSE },
{ { 0x6d, 0x00, 0x00, 0x01 }, "Office 2000", 6, FALSE },
{ { 0x6f, 0x00, 0x00, 0x01 }, "Office 2000", 6, FALSE },
{ { 0x70, 0x00, 0x00, 0x01 }, "Office XP beta 1/2", 6, FALSE },
{ { 0x73, 0x00, 0x00, 0x01 }, "Office XP", 6, FALSE },
{ { 0x76, 0x00, 0x00, 0x01 }, "Office 2003", 6, FALSE },
{ { 0x79, 0x00, 0x00, 0x01 }, "Office 2003", 6, FALSE },
{ { 0x60, 0x00, 0x00, 0x0e }, "MacOffice 98", 5, TRUE },
{ { 0x62, 0x00, 0x00, 0x0e }, "MacOffice 2001", 5, TRUE },
{ { 0x63, 0x00, 0x00, 0x0e }, "MacOffice X", 6, TRUE },
{ { 0x64, 0x00, 0x00, 0x0e }, "MacOffice 2004", 6, TRUE },
};
guint8 const *data;
unsigned i, count, len;
gunichar2 *uni_name;
char *name;
GsfInput *dir;
dir = gsf_infile_child_by_name (vba->source, "dir");
if (dir == NULL) {
if (err != NULL)
*err = g_error_new (gsf_input_error_id (), 0,
_("Can't find the VBA directory stream"));
return FALSE;
}
if (gsf_input_seek (dir, 0, G_SEEK_SET) ||
NULL == (data = gsf_input_read (dir, VBA56_DIRENT_HEADER_SIZE, NULL)) ||
0 != memcmp (data, signature, sizeof (signature))) {
if (err != NULL)
*err = g_error_new (gsf_input_error_id (), 0,
_("No VBA signature"));
return FALSE;
}
for (i = 0 ; i < G_N_ELEMENTS (versions); i++)
if (!memcmp (data+2, versions[i].signature, 4))
break;
if (i >= G_N_ELEMENTS (versions)) {
if (err != NULL)
*err = g_error_new (gsf_input_error_id (), 0,
_("Unknown VBA version signature 0x%x%x%x%x"),
data[2], data[3], data[4], data[5]);
return FALSE;
}
puts (versions[i].name);
/* these depend strings seem to come in 2 blocks */
count = GSF_LE_GET_GUINT16 (data + VBA56_DIRENT_RECORD_COUNT);
for (; count > 0 ; count--) {
if (NULL == ((data = gsf_input_read (dir, 2, NULL))))
break;
len = GSF_LE_GET_GUINT16 (data);
if (NULL == ((data = gsf_input_read (dir, len, NULL)))) {
printf ("len == 0x%x ??\n", len);
break;
}
uni_name = g_new0 (gunichar2, len/2 + 1);
/* be wary about endianness */
for (i = 0 ; i < len ; i += 2)
uni_name [i/2] = GSF_LE_GET_GUINT16 (data + i);
name = g_utf16_to_utf8 (uni_name, -1, NULL, NULL, NULL);
g_free (uni_name);
printf ("%d %s\n", count, name);
/* ignore this blob ???? */
if (!strncmp ("*\\G", name, 3)) {
if (NULL == ((data = gsf_input_read (dir, 12, NULL)))) {
printf ("len == 0x%x ??\n", len);
break;
}
}
g_free (name);
}
g_return_val_if_fail (count == 0, FALSE);
return TRUE;
}
/**
* 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;
}
