sheetjs/js-cfb
1
Fork 0
Code Issues 3 Pull Requests Projects Releases Wiki Activity

version bump 1.2.1: DEFLATE

Browse Source
This commit is contained in:
SheetJS 2021-09-06 15:53:23 -04:00
parent efb96a2c2c
commit 1e70aae159
20 changed files with 1298 additions and 303 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
CHANGELOG.md
View File

@ -4,6 +4,13 @@ This log is intended to keep track of backwards-incompatible changes, including
but not limited to API changes and file location changes. Minor behavioral
changes may not be included if they are not expected to break existing code.
## 1.2.1 (2021-09-06)
* CFB write optimizations (h/t @rossj Ross Johnson)
* `read` in NodeJS will treat `Buffer` input as type `"buffer"` by default
* `deflate` / ZIP support fixed Huffman compression
* `inflate` more aggressive reallocs
## 1.2.0 (2020-07-09)
* Support for MAD file format (MIME aggregate document)

2
README.md
View File

@ -74,6 +74,7 @@ name, if there are no slashes) and returns an entry object or null if not found.
|------------|:----------------------------------------------------------------|
| `"base64"` | string: Base64 encoding of the file |
| `"binary"` | string: binary string (byte `n` is `data.charCodeAt(n)`) |
| `"buffer"` | nodejs Buffer |
| `"file"` | string: path of file that will be read (nodejs only) |
| (default) | buffer or array of 8-bit unsigned int (byte `n` is `data[n]`) |
@ -88,6 +89,7 @@ name, if there are no slashes) and returns an entry object or null if not found.
|------------|:----------------------------------------------------------------|
| `"base64"` | string: Base64 encoding of the file |
| `"binary"` | string: binary string (byte `n` is `data.charCodeAt(n)`) |
| `"buffer"` | nodejs Buffer |
| `"file"` | string: path of file that will be created (nodejs only) |
| (default) | buffer if available, array of 8-bit unsigned int otherwise |

2
bits/31_version.js
View File

@ -1 +1 @@
exports.version = '1.2.0';
exports.version = '1.2.1';

6
bits/49_readutils.js
View File

@ -4,7 +4,11 @@ function read_file(filename/*:string*/, options/*:CFBReadOpts*/) {
}
function read(blob/*:RawBytes|string*/, options/*:CFBReadOpts*/) {
switch(options && options.type || "base64") {
var type = options && options.type;
if(!type) {
if(has_buf && Buffer.isBuffer(blob)) type = "buffer";
}
switch(type || "base64") {
case "file": /*:: if(typeof blob !== 'string') throw "Must pass a filename when type='file'"; */return read_file(blob, options);
case "base64": /*:: if(typeof blob !== 'string') throw "Must pass a base64-encoded binary string when type='file'"; */return parse(s2a(Base64.decode(blob)), options);
case "binary": /*:: if(typeof blob !== 'string') throw "Must pass a binary string when type='file'"; */return parse(s2a(blob), options);

96
bits/79_flate.js
View File

@ -42,6 +42,38 @@ function read_bits_n(buf, bl, n) {
return v & f;
}
/* helpers for unaligned bit writes */
function write_bits_3(buf, bl, v) { var w = bl & 7, h = bl >>> 3;
if(w <= 5) buf[h] |= (v & 7) << w;
else {
buf[h] |= (v << w) & 0xFF;
buf[h+1] = (v&7) >> (8-w);
}
return bl + 3;
}
function write_bits_1(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v = (v&1) << w;
buf[h] |= v;
return bl + 1;
}
function write_bits_8(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v <<= w;
buf[h] |= v & 0xFF; v >>>= 8;
buf[h+1] = v;
return bl + 8;
}
function write_bits_16(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v <<= w;
buf[h] |= v & 0xFF; v >>>= 8;
buf[h+1] = v & 0xFF;
buf[h+2] = v >>> 8;
return bl + 16;
}
/* until ArrayBuffer#realloc is a thing, fake a realloc */
function realloc(b, sz/*:number*/) {
var L = b.length, M = 2*L > sz ? 2*L : sz + 5, i = 0;
@ -55,7 +87,7 @@ function realloc(b, sz/*:number*/) {
} else if(use_typed_arrays) {
var a = new Uint8Array(M);
if(a.set) a.set(b);
else for(; i < b.length; ++i) a[i] = b[i];
else for(; i < L; ++i) a[i] = b[i];
return a;
}
b.length = M;
@ -67,4 +99,64 @@ function zero_fill_array(n) {
var o = new Array(n);
for(var i = 0; i < n; ++i) o[i] = 0;
return o;
}
}
/* build tree (used for literals and lengths) */
function build_tree(clens, cmap, MAX/*:number*/)/*:number*/ {
var maxlen = 1, w = 0, i = 0, j = 0, ccode = 0, L = clens.length;
var bl_count = use_typed_arrays ? new Uint16Array(32) : zero_fill_array(32);
for(i = 0; i < 32; ++i) bl_count[i] = 0;
for(i = L; i < MAX; ++i) clens[i] = 0;
L = clens.length;
var ctree = use_typed_arrays ? new Uint16Array(L) : zero_fill_array(L); // []
/* build code tree */
for(i = 0; i < L; ++i) {
bl_count[(w = clens[i])]++;
if(maxlen < w) maxlen = w;
ctree[i] = 0;
}
bl_count[0] = 0;
for(i = 1; i <= maxlen; ++i) bl_count[i+16] = (ccode = (ccode + bl_count[i-1])<<1);
for(i = 0; i < L; ++i) {
ccode = clens[i];
if(ccode != 0) ctree[i] = bl_count[ccode+16]++;
}
/* cmap[maxlen + 4 bits] = (off&15) + (lit<<4) reverse mapping */
var cleni = 0;
for(i = 0; i < L; ++i) {
cleni = clens[i];
if(cleni != 0) {
ccode = bit_swap_n(ctree[i], maxlen)>>(maxlen-cleni);
for(j = (1<<(maxlen + 4 - cleni)) - 1; j>=0; --j)
cmap[ccode|(j<<cleni)] = (cleni&15) | (i<<4);
}
}
return maxlen;
}
/* Fixed Huffman */
var fix_lmap = use_typed_arrays ? new Uint16Array(512) : zero_fill_array(512);
var fix_dmap = use_typed_arrays ? new Uint16Array(32) : zero_fill_array(32);
if(!use_typed_arrays) {
for(var i = 0; i < 512; ++i) fix_lmap[i] = 0;
for(i = 0; i < 32; ++i) fix_dmap[i] = 0;
}
(function() {
var dlens/*:Array<number>*/ = [];
var i = 0;
for(;i<32; i++) dlens.push(5);
build_tree(dlens, fix_dmap, 32);
var clens/*:Array<number>*/ = [];
i = 0;
for(; i<=143; i++) clens.push(8);
for(; i<=255; i++) clens.push(9);
for(; i<=279; i++) clens.push(7);
for(; i<=287; i++) clens.push(8);
build_tree(clens, fix_lmap, 288);
})();

105
bits/80_deflate.js
View File

@ -1,23 +1,116 @@
var _deflate = (function() {
var _deflateRaw = (function() {
return function deflateRaw(data, out) {
var DST_LN_RE = use_typed_arrays ? new Uint8Array(0x8000) : [];
for(var j = 0, k = 0; j < DST_LN.length; ++j) {
for(; k < DST_LN[j+1]; ++k) DST_LN_RE[k] = j;
}
for(;k < 32768; ++k) DST_LN_RE[k] = 29;
var LEN_LN_RE = use_typed_arrays ? new Uint8Array(0x102) : [];
for(j = 0, k = 0; j < LEN_LN.length; ++j) {
for(; k < LEN_LN[j+1]; ++k) LEN_LN_RE[k] = j;
}
function write_stored(data, out) {
var boff = 0;
while(boff < data.length) {
var L = Math.min(0xFFFF, data.length - boff);
var h = boff + L == data.length;
/* TODO: this is only type 0 stored */
out.write_shift(1, +h);
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
}
return out.l;
}
/* Fixed Huffman */
function write_huff_fixed(data, out) {
var bl = 0;
var boff = 0;
var addrs = use_typed_arrays ? new Uint16Array(0x8000) : [];
while(boff < data.length) {
var L = /* data.length - boff; */ Math.min(0xFFFF, data.length - boff);
/* write a stored block for short data */
if(L < 10) {
bl = write_bits_3(out, bl, +!!(boff + L == data.length)); // jshint ignore:line
if(bl & 7) bl += 8 - (bl & 7);
out.l = (bl / 8) | 0;
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
bl = out.l * 8;
continue;
}
bl = write_bits_3(out, bl, +!!(boff + L == data.length) + 2); // jshint ignore:line
var hash = 0;
while(L-- > 0) {
var d = data[boff];
hash = ((hash << 5) ^ d) & 0x7FFF;
var match = -1, mlen = 0;
if((match = addrs[hash])) {
match |= boff & ~0x7FFF;
if(match > boff) match -= 0x8000;
if(match < boff) while(data[match + mlen] == data[boff + mlen] && mlen < 250) ++mlen;
}
if(mlen > 2) {
/* Copy Token */
d = LEN_LN_RE[mlen];
if(d <= 22) bl = write_bits_8(out, bl, bitswap8[d+1]>>1) - 1;
else {
write_bits_8(out, bl, 3);
bl += 5;
write_bits_8(out, bl, bitswap8[d-23]>>5);
bl += 3;
}
var len_eb = (d < 8) ? 0 : ((d - 4)>>2);
if(len_eb > 0) {
write_bits_16(out, bl, mlen - LEN_LN[d]);
bl += len_eb;
}
d = DST_LN_RE[boff - match];
bl = write_bits_8(out, bl, bitswap8[d]>>3);
bl -= 3;
var dst_eb = d < 4 ? 0 : (d-2)>>1;
if(dst_eb > 0) {
write_bits_16(out, bl, boff - match - DST_LN[d]);
bl += dst_eb;
}
for(var q = 0; q < mlen; ++q) {
addrs[hash] = boff & 0x7FFF;
hash = ((hash << 5) ^ data[boff]) & 0x7FFF;
++boff;
}
L-= mlen - 1;
} else {
/* Literal Token */
if(d <= 143) d = d + 48;
else bl = write_bits_1(out, bl, 1);
bl = write_bits_8(out, bl, bitswap8[d]);
addrs[hash] = boff & 0x7FFF;
++boff;
}
}
bl = write_bits_8(out, bl, 0) - 1;
}
out.l = ((bl + 7)/8)|0;
return out.l;
}
return function _deflateRaw(data, out) {
if(data.length < 8) return write_stored(data, out);
return write_huff_fixed(data, out);
};
})();
return function(data) {
function _deflate(data) {
var buf = new_buf(50+Math.floor(data.length*1.1));
var off = _deflateRaw(data, buf);
return buf.slice(0, off);
};
})();
}

63
bits/81_inflate.js
View File

@ -1,64 +1,5 @@
/* modified inflate function also moves original read head */
/* build tree (used for literals and lengths) */
function build_tree(clens, cmap, MAX/*:number*/)/*:number*/ {
var maxlen = 1, w = 0, i = 0, j = 0, ccode = 0, L = clens.length;
var bl_count = use_typed_arrays ? new Uint16Array(32) : zero_fill_array(32);
for(i = 0; i < 32; ++i) bl_count[i] = 0;
for(i = L; i < MAX; ++i) clens[i] = 0;
L = clens.length;
var ctree = use_typed_arrays ? new Uint16Array(L) : zero_fill_array(L); // []
/* build code tree */
for(i = 0; i < L; ++i) {
bl_count[(w = clens[i])]++;
if(maxlen < w) maxlen = w;
ctree[i] = 0;
}
bl_count[0] = 0;
for(i = 1; i <= maxlen; ++i) bl_count[i+16] = (ccode = (ccode + bl_count[i-1])<<1);
for(i = 0; i < L; ++i) {
ccode = clens[i];
if(ccode != 0) ctree[i] = bl_count[ccode+16]++;
}
/* cmap[maxlen + 4 bits] = (off&15) + (lit<<4) reverse mapping */
var cleni = 0;
for(i = 0; i < L; ++i) {
cleni = clens[i];
if(cleni != 0) {
ccode = bit_swap_n(ctree[i], maxlen)>>(maxlen-cleni);
for(j = (1<<(maxlen + 4 - cleni)) - 1; j>=0; --j)
cmap[ccode|(j<<cleni)] = (cleni&15) | (i<<4);
}
}
return maxlen;
}
var fix_lmap = use_typed_arrays ? new Uint16Array(512) : zero_fill_array(512);
var fix_dmap = use_typed_arrays ? new Uint16Array(32) : zero_fill_array(32);
if(!use_typed_arrays) {
for(var i = 0; i < 512; ++i) fix_lmap[i] = 0;
for(i = 0; i < 32; ++i) fix_dmap[i] = 0;
}
(function() {
var dlens/*:Array<number>*/ = [];
var i = 0;
for(;i<32; i++) dlens.push(5);
build_tree(dlens, fix_dmap, 32);
var clens/*:Array<number>*/ = [];
i = 0;
for(; i<=143; i++) clens.push(8);
for(; i<=255; i++) clens.push(9);
for(; i<=279; i++) clens.push(7);
for(; i<=287; i++) clens.push(8);
build_tree(clens, fix_lmap, 288);
})();
var dyn_lmap = use_typed_arrays ? new Uint16Array(32768) : zero_fill_array(32768);
var dyn_dmap = use_typed_arrays ? new Uint16Array(32768) : zero_fill_array(32768);
var dyn_cmap = use_typed_arrays ? new Uint16Array(128) : zero_fill_array(128);
@ -177,8 +118,8 @@ function inflate(data, usz/*:number*/) {
boff = dyn(data, boff);
max_len_1 = dyn_len_1; max_len_2 = dyn_len_2;
}
if(!usz && (OL < woff + 32767)) { outbuf = realloc(outbuf, woff + 32767); OL = outbuf.length; }
for(;;) { // while(true) is apparently out of vogue in modern JS circles
if(!usz && (OL < woff + 32767)) { outbuf = realloc(outbuf, woff + 32767); OL = outbuf.length; }
/* ingest code and move read head */
var bits = read_bits_n(data, boff, max_len_1);
var code = (header>>>1) == 1 ? fix_lmap[bits] : dyn_lmap[bits];
@ -209,7 +150,7 @@ function inflate(data, usz/*:number*/) {
}
/* in the common case, manual byte copy is faster than TA set / Buffer copy */
if(!usz && OL < tgt) { outbuf = realloc(outbuf, tgt); OL = outbuf.length; }
if(!usz && OL < tgt) { outbuf = realloc(outbuf, tgt + 100); OL = outbuf.length; }
while(woff < tgt) { outbuf[woff] = outbuf[woff - dst]; ++woff; }
}
}

2
bits/82_zparse.js
View File

@ -82,7 +82,7 @@ function parse_local_file(blob/*:CFBlob*/, csz/*:number*/, usz/*:number*/, o/*:C
var data = blob.slice(blob.l, blob.l + _csz);
switch(meth) {
case 8: data = _inflateRawSync(blob, _usz); break;
case 0: break;
case 0: break; // TODO: scan for magic number
default: throw new Error("Unsupported ZIP Compression method " + meth);
}

3
bits/83_zwrite.js
View File

@ -46,7 +46,10 @@ function write_zip(cfb/*:CFBContainer*/, options/*:CFBWriteOpts*/)/*:RawBytes*/
start_cd += namebuf.length;
out.push(namebuf);
/* TODO: extra fields? */
/* TODO: encryption header ? */
start_cd += outbuf.length;
out.push(outbuf);

2
bits/84_mht.js
View File

@ -203,4 +203,4 @@ function write_mad(cfb/*:CFBContainer*/, options/*:CFBWriteOpts*/)/*:string*/ {
}
out.push(boundary + '--\r\n');
return out.join("\r\n");
}
}

196
cfb.flow.js
View File

@ -295,7 +295,7 @@ CRC32.str = crc32_str;
/* [MS-CFB] v20171201 */
var CFB = (function _CFB(){
var exports/*:CFBModule*/ = /*::(*/{}/*:: :any)*/;
exports.version = '1.2.0';
exports.version = '1.2.1';
/* [MS-CFB] 2.6.4 */
function namecmp(l/*:string*/, r/*:string*/)/*:number*/ {
var L = l.split("/"), R = r.split("/");
@ -707,7 +707,11 @@ function read_file(filename/*:string*/, options/*:CFBReadOpts*/) {
}
function read(blob/*:RawBytes|string*/, options/*:CFBReadOpts*/) {
switch(options && options.type || "base64") {
var type = options && options.type;
if(!type) {
if(has_buf && Buffer.isBuffer(blob)) type = "buffer";
}
switch(type || "base64") {
case "file": /*:: if(typeof blob !== 'string') throw "Must pass a filename when type='file'"; */return read_file(blob, options);
case "base64": /*:: if(typeof blob !== 'string') throw "Must pass a base64-encoded binary string when type='file'"; */return parse(s2a(Base64.decode(blob)), options);
case "binary": /*:: if(typeof blob !== 'string') throw "Must pass a binary string when type='file'"; */return parse(s2a(blob), options);
@ -1105,6 +1109,38 @@ function read_bits_n(buf, bl, n) {
return v & f;
}
/* helpers for unaligned bit writes */
function write_bits_3(buf, bl, v) { var w = bl & 7, h = bl >>> 3;
if(w <= 5) buf[h] |= (v & 7) << w;
else {
buf[h] |= (v << w) & 0xFF;
buf[h+1] = (v&7) >> (8-w);
}
return bl + 3;
}
function write_bits_1(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v = (v&1) << w;
buf[h] |= v;
return bl + 1;
}
function write_bits_8(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v <<= w;
buf[h] |= v & 0xFF; v >>>= 8;
buf[h+1] = v;
return bl + 8;
}
function write_bits_16(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v <<= w;
buf[h] |= v & 0xFF; v >>>= 8;
buf[h+1] = v & 0xFF;
buf[h+2] = v >>> 8;
return bl + 16;
}
/* until ArrayBuffer#realloc is a thing, fake a realloc */
function realloc(b, sz/*:number*/) {
var L = b.length, M = 2*L > sz ? 2*L : sz + 5, i = 0;
@ -1118,7 +1154,7 @@ function realloc(b, sz/*:number*/) {
} else if(use_typed_arrays) {
var a = new Uint8Array(M);
if(a.set) a.set(b);
else for(; i < b.length; ++i) a[i] = b[i];
else for(; i < L; ++i) a[i] = b[i];
return a;
}
b.length = M;
@ -1130,30 +1166,7 @@ function zero_fill_array(n) {
var o = new Array(n);
for(var i = 0; i < n; ++i) o[i] = 0;
return o;
}var _deflate = (function() {
var _deflateRaw = (function() {
return function deflateRaw(data, out) {
var boff = 0;
while(boff < data.length) {
var L = Math.min(0xFFFF, data.length - boff);
var h = boff + L == data.length;
/* TODO: this is only type 0 stored */
out.write_shift(1, +h);
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
}
return out.l;
};
})();
return function(data) {
var buf = new_buf(50+Math.floor(data.length*1.1));
var off = _deflateRaw(data, buf);
return buf.slice(0, off);
};
})();
/* modified inflate function also moves original read head */
}
/* build tree (used for literals and lengths) */
function build_tree(clens, cmap, MAX/*:number*/)/*:number*/ {
@ -1193,6 +1206,7 @@ function build_tree(clens, cmap, MAX/*:number*/)/*:number*/ {
return maxlen;
}
/* Fixed Huffman */
var fix_lmap = use_typed_arrays ? new Uint16Array(512) : zero_fill_array(512);
var fix_dmap = use_typed_arrays ? new Uint16Array(32) : zero_fill_array(32);
if(!use_typed_arrays) {
@ -1212,8 +1226,124 @@ if(!use_typed_arrays) {
for(; i<=279; i++) clens.push(7);
for(; i<=287; i++) clens.push(8);
build_tree(clens, fix_lmap, 288);
})();var _deflateRaw = (function() {
var DST_LN_RE = use_typed_arrays ? new Uint8Array(0x8000) : [];
for(var j = 0, k = 0; j < DST_LN.length; ++j) {
for(; k < DST_LN[j+1]; ++k) DST_LN_RE[k] = j;
}
for(;k < 32768; ++k) DST_LN_RE[k] = 29;
var LEN_LN_RE = use_typed_arrays ? new Uint8Array(0x102) : [];
for(j = 0, k = 0; j < LEN_LN.length; ++j) {
for(; k < LEN_LN[j+1]; ++k) LEN_LN_RE[k] = j;
}
function write_stored(data, out) {
var boff = 0;
while(boff < data.length) {
var L = Math.min(0xFFFF, data.length - boff);
var h = boff + L == data.length;
out.write_shift(1, +h);
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
}
return out.l;
}
/* Fixed Huffman */
function write_huff_fixed(data, out) {
var bl = 0;
var boff = 0;
var addrs = use_typed_arrays ? new Uint16Array(0x8000) : [];
while(boff < data.length) {
var L = /* data.length - boff; */ Math.min(0xFFFF, data.length - boff);
/* write a stored block for short data */
if(L < 10) {
bl = write_bits_3(out, bl, +!!(boff + L == data.length)); // jshint ignore:line
if(bl & 7) bl += 8 - (bl & 7);
out.l = (bl / 8) | 0;
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
bl = out.l * 8;
continue;
}
bl = write_bits_3(out, bl, +!!(boff + L == data.length) + 2); // jshint ignore:line
var hash = 0;
while(L-- > 0) {
var d = data[boff];
hash = ((hash << 5) ^ d) & 0x7FFF;
var match = -1, mlen = 0;
if((match = addrs[hash])) {
match |= boff & ~0x7FFF;
if(match > boff) match -= 0x8000;
if(match < boff) while(data[match + mlen] == data[boff + mlen] && mlen < 250) ++mlen;
}
if(mlen > 2) {
/* Copy Token */
d = LEN_LN_RE[mlen];
if(d <= 22) bl = write_bits_8(out, bl, bitswap8[d+1]>>1) - 1;
else {
write_bits_8(out, bl, 3);
bl += 5;
write_bits_8(out, bl, bitswap8[d-23]>>5);
bl += 3;
}
var len_eb = (d < 8) ? 0 : ((d - 4)>>2);
if(len_eb > 0) {
write_bits_16(out, bl, mlen - LEN_LN[d]);
bl += len_eb;
}
d = DST_LN_RE[boff - match];
bl = write_bits_8(out, bl, bitswap8[d]>>3);
bl -= 3;
var dst_eb = d < 4 ? 0 : (d-2)>>1;
if(dst_eb > 0) {
write_bits_16(out, bl, boff - match - DST_LN[d]);
bl += dst_eb;
}
for(var q = 0; q < mlen; ++q) {
addrs[hash] = boff & 0x7FFF;
hash = ((hash << 5) ^ data[boff]) & 0x7FFF;
++boff;
}
L-= mlen - 1;
} else {
/* Literal Token */
if(d <= 143) d = d + 48;
else bl = write_bits_1(out, bl, 1);
bl = write_bits_8(out, bl, bitswap8[d]);
addrs[hash] = boff & 0x7FFF;
++boff;
}
}
bl = write_bits_8(out, bl, 0) - 1;
}
out.l = ((bl + 7)/8)|0;
return out.l;
}
return function _deflateRaw(data, out) {
if(data.length < 8) return write_stored(data, out);
return write_huff_fixed(data, out);
};
})();
function _deflate(data) {
var buf = new_buf(50+Math.floor(data.length*1.1));
var off = _deflateRaw(data, buf);
return buf.slice(0, off);
}
/* modified inflate function also moves original read head */
var dyn_lmap = use_typed_arrays ? new Uint16Array(32768) : zero_fill_array(32768);
var dyn_dmap = use_typed_arrays ? new Uint16Array(32768) : zero_fill_array(32768);
var dyn_cmap = use_typed_arrays ? new Uint16Array(128) : zero_fill_array(128);
@ -1332,8 +1462,8 @@ function inflate(data, usz/*:number*/) {
boff = dyn(data, boff);
max_len_1 = dyn_len_1; max_len_2 = dyn_len_2;
}
if(!usz && (OL < woff + 32767)) { outbuf = realloc(outbuf, woff + 32767); OL = outbuf.length; }
for(;;) { // while(true) is apparently out of vogue in modern JS circles
if(!usz && (OL < woff + 32767)) { outbuf = realloc(outbuf, woff + 32767); OL = outbuf.length; }
/* ingest code and move read head */
var bits = read_bits_n(data, boff, max_len_1);
var code = (header>>>1) == 1 ? fix_lmap[bits] : dyn_lmap[bits];
@ -1364,7 +1494,7 @@ function inflate(data, usz/*:number*/) {
}
/* in the common case, manual byte copy is faster than TA set / Buffer copy */
if(!usz && OL < tgt) { outbuf = realloc(outbuf, tgt); OL = outbuf.length; }
if(!usz && OL < tgt) { outbuf = realloc(outbuf, tgt + 100); OL = outbuf.length; }
while(woff < tgt) { outbuf[woff] = outbuf[woff - dst]; ++woff; }
}
}
@ -1463,7 +1593,7 @@ function parse_local_file(blob/*:CFBlob*/, csz/*:number*/, usz/*:number*/, o/*:C
var data = blob.slice(blob.l, blob.l + _csz);
switch(meth) {
case 8: data = _inflateRawSync(blob, _usz); break;
case 0: break;
case 0: break; // TODO: scan for magic number
default: throw new Error("Unsupported ZIP Compression method " + meth);
}
@ -1530,7 +1660,10 @@ function write_zip(cfb/*:CFBContainer*/, options/*:CFBWriteOpts*/)/*:RawBytes*/
start_cd += namebuf.length;
out.push(namebuf);
/* TODO: extra fields? */
/* TODO: encryption header ? */
start_cd += outbuf.length;
out.push(outbuf);
@ -1789,7 +1922,8 @@ function write_mad(cfb/*:CFBContainer*/, options/*:CFBWriteOpts*/)/*:string*/ {
}
out.push(boundary + '--\r\n');
return out.join("\r\n");
}function cfb_new(opts/*:?any*/)/*:CFBContainer*/ {
}
function cfb_new(opts/*:?any*/)/*:CFBContainer*/ {
var o/*:CFBContainer*/ = ({}/*:any*/);
init_cfb(o, opts);
return o;

196
cfb.js
View File

@ -277,7 +277,7 @@ CRC32.str = crc32_str;
/* [MS-CFB] v20171201 */
var CFB = (function _CFB(){
var exports = {};
exports.version = '1.2.0';
exports.version = '1.2.1';
/* [MS-CFB] 2.6.4 */
function namecmp(l, r) {
var L = l.split("/"), R = r.split("/");
@ -689,7 +689,11 @@ function read_file(filename, options) {
}
function read(blob, options) {
switch(options && options.type || "base64") {
var type = options && options.type;
if(!type) {
if(has_buf && Buffer.isBuffer(blob)) type = "buffer";
}
switch(type || "base64") {
case "file": return read_file(blob, options);
case "base64": return parse(s2a(Base64.decode(blob)), options);
case "binary": return parse(s2a(blob), options);
@ -1081,6 +1085,38 @@ function read_bits_n(buf, bl, n) {
return v & f;
}
/* helpers for unaligned bit writes */
function write_bits_3(buf, bl, v) { var w = bl & 7, h = bl >>> 3;
if(w <= 5) buf[h] |= (v & 7) << w;
else {
buf[h] |= (v << w) & 0xFF;
buf[h+1] = (v&7) >> (8-w);
}
return bl + 3;
}
function write_bits_1(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v = (v&1) << w;
buf[h] |= v;
return bl + 1;
}
function write_bits_8(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v <<= w;
buf[h] |= v & 0xFF; v >>>= 8;
buf[h+1] = v;
return bl + 8;
}
function write_bits_16(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v <<= w;
buf[h] |= v & 0xFF; v >>>= 8;
buf[h+1] = v & 0xFF;
buf[h+2] = v >>> 8;
return bl + 16;
}
/* until ArrayBuffer#realloc is a thing, fake a realloc */
function realloc(b, sz) {
var L = b.length, M = 2*L > sz ? 2*L : sz + 5, i = 0;
@ -1094,7 +1130,7 @@ function realloc(b, sz) {
} else if(use_typed_arrays) {
var a = new Uint8Array(M);
if(a.set) a.set(b);
else for(; i < b.length; ++i) a[i] = b[i];
else for(; i < L; ++i) a[i] = b[i];
return a;
}
b.length = M;
@ -1106,30 +1142,7 @@ function zero_fill_array(n) {
var o = new Array(n);
for(var i = 0; i < n; ++i) o[i] = 0;
return o;
}var _deflate = (function() {
var _deflateRaw = (function() {
return function deflateRaw(data, out) {
var boff = 0;
while(boff < data.length) {
var L = Math.min(0xFFFF, data.length - boff);
var h = boff + L == data.length;
/* TODO: this is only type 0 stored */
out.write_shift(1, +h);
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
}
return out.l;
};
})();
return function(data) {
var buf = new_buf(50+Math.floor(data.length*1.1));
var off = _deflateRaw(data, buf);
return buf.slice(0, off);
};
})();
/* modified inflate function also moves original read head */
}
/* build tree (used for literals and lengths) */
function build_tree(clens, cmap, MAX) {
@ -1169,6 +1182,7 @@ function build_tree(clens, cmap, MAX) {
return maxlen;
}
/* Fixed Huffman */
var fix_lmap = use_typed_arrays ? new Uint16Array(512) : zero_fill_array(512);
var fix_dmap = use_typed_arrays ? new Uint16Array(32) : zero_fill_array(32);
if(!use_typed_arrays) {
@ -1188,8 +1202,124 @@ if(!use_typed_arrays) {
for(; i<=279; i++) clens.push(7);
for(; i<=287; i++) clens.push(8);
build_tree(clens, fix_lmap, 288);
})();var _deflateRaw = (function() {
var DST_LN_RE = use_typed_arrays ? new Uint8Array(0x8000) : [];
for(var j = 0, k = 0; j < DST_LN.length; ++j) {
for(; k < DST_LN[j+1]; ++k) DST_LN_RE[k] = j;
}
for(;k < 32768; ++k) DST_LN_RE[k] = 29;
var LEN_LN_RE = use_typed_arrays ? new Uint8Array(0x102) : [];
for(j = 0, k = 0; j < LEN_LN.length; ++j) {
for(; k < LEN_LN[j+1]; ++k) LEN_LN_RE[k] = j;
}
function write_stored(data, out) {
var boff = 0;
while(boff < data.length) {
var L = Math.min(0xFFFF, data.length - boff);
var h = boff + L == data.length;
out.write_shift(1, +h);
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
}
return out.l;
}
/* Fixed Huffman */
function write_huff_fixed(data, out) {
var bl = 0;
var boff = 0;
var addrs = use_typed_arrays ? new Uint16Array(0x8000) : [];
while(boff < data.length) {
var L = /* data.length - boff; */ Math.min(0xFFFF, data.length - boff);
/* write a stored block for short data */
if(L < 10) {
bl = write_bits_3(out, bl, +!!(boff + L == data.length)); // jshint ignore:line
if(bl & 7) bl += 8 - (bl & 7);
out.l = (bl / 8) | 0;
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
bl = out.l * 8;
continue;
}
bl = write_bits_3(out, bl, +!!(boff + L == data.length) + 2); // jshint ignore:line
var hash = 0;
while(L-- > 0) {
var d = data[boff];
hash = ((hash << 5) ^ d) & 0x7FFF;
var match = -1, mlen = 0;
if((match = addrs[hash])) {
match |= boff & ~0x7FFF;
if(match > boff) match -= 0x8000;
if(match < boff) while(data[match + mlen] == data[boff + mlen] && mlen < 250) ++mlen;
}
if(mlen > 2) {
/* Copy Token */
d = LEN_LN_RE[mlen];
if(d <= 22) bl = write_bits_8(out, bl, bitswap8[d+1]>>1) - 1;
else {
write_bits_8(out, bl, 3);
bl += 5;
write_bits_8(out, bl, bitswap8[d-23]>>5);
bl += 3;
}
var len_eb = (d < 8) ? 0 : ((d - 4)>>2);
if(len_eb > 0) {
write_bits_16(out, bl, mlen - LEN_LN[d]);
bl += len_eb;
}
d = DST_LN_RE[boff - match];
bl = write_bits_8(out, bl, bitswap8[d]>>3);
bl -= 3;
var dst_eb = d < 4 ? 0 : (d-2)>>1;
if(dst_eb > 0) {
write_bits_16(out, bl, boff - match - DST_LN[d]);
bl += dst_eb;
}
for(var q = 0; q < mlen; ++q) {
addrs[hash] = boff & 0x7FFF;
hash = ((hash << 5) ^ data[boff]) & 0x7FFF;
++boff;
}
L-= mlen - 1;
} else {
/* Literal Token */
if(d <= 143) d = d + 48;
else bl = write_bits_1(out, bl, 1);
bl = write_bits_8(out, bl, bitswap8[d]);
addrs[hash] = boff & 0x7FFF;
++boff;
}
}
bl = write_bits_8(out, bl, 0) - 1;
}
out.l = ((bl + 7)/8)|0;
return out.l;
}
return function _deflateRaw(data, out) {
if(data.length < 8) return write_stored(data, out);
return write_huff_fixed(data, out);
};
})();
function _deflate(data) {
var buf = new_buf(50+Math.floor(data.length*1.1));
var off = _deflateRaw(data, buf);
return buf.slice(0, off);
}
/* modified inflate function also moves original read head */
var dyn_lmap = use_typed_arrays ? new Uint16Array(32768) : zero_fill_array(32768);
var dyn_dmap = use_typed_arrays ? new Uint16Array(32768) : zero_fill_array(32768);
var dyn_cmap = use_typed_arrays ? new Uint16Array(128) : zero_fill_array(128);
@ -1308,8 +1438,8 @@ function inflate(data, usz) {
boff = dyn(data, boff);
max_len_1 = dyn_len_1; max_len_2 = dyn_len_2;
}
if(!usz && (OL < woff + 32767)) { outbuf = realloc(outbuf, woff + 32767); OL = outbuf.length; }
for(;;) { // while(true) is apparently out of vogue in modern JS circles
if(!usz && (OL < woff + 32767)) { outbuf = realloc(outbuf, woff + 32767); OL = outbuf.length; }
/* ingest code and move read head */
var bits = read_bits_n(data, boff, max_len_1);
var code = (header>>>1) == 1 ? fix_lmap[bits] : dyn_lmap[bits];
@ -1340,7 +1470,7 @@ function inflate(data, usz) {
}
/* in the common case, manual byte copy is faster than TA set / Buffer copy */
if(!usz && OL < tgt) { outbuf = realloc(outbuf, tgt); OL = outbuf.length; }
if(!usz && OL < tgt) { outbuf = realloc(outbuf, tgt + 100); OL = outbuf.length; }
while(woff < tgt) { outbuf[woff] = outbuf[woff - dst]; ++woff; }
}
}
@ -1439,7 +1569,7 @@ function parse_local_file(blob, csz, usz, o, EF) {
var data = blob.slice(blob.l, blob.l + _csz);
switch(meth) {
case 8: data = _inflateRawSync(blob, _usz); break;
case 0: break;
case 0: break; // TODO: scan for magic number
default: throw new Error("Unsupported ZIP Compression method " + meth);
}
@ -1506,7 +1636,10 @@ function write_zip(cfb, options) {
start_cd += namebuf.length;
out.push(namebuf);
/* TODO: extra fields? */
/* TODO: encryption header ? */
start_cd += outbuf.length;
out.push(outbuf);
@ -1765,7 +1898,8 @@ function write_mad(cfb, options) {
}
out.push(boundary + '--\r\n');
return out.join("\r\n");
}function cfb_new(opts) {
}
function cfb_new(opts) {
var o = ({});
init_cfb(o, opts);
return o;

249
dist/cfb.js vendored
View File

@ -57,7 +57,7 @@ if(typeof Buffer !== 'undefined') {
if(!nbfs) try { Buffer.from("foo", "utf8"); } catch(e) { nbfs = true; }
Buffer_from = nbfs ? function(buf, enc) { return (enc) ? new Buffer(buf, enc) : new Buffer(buf); } : Buffer.from.bind(Buffer);
// $FlowIgnore
if(!Buffer.alloc) Buffer.alloc = function(n) { return new Buffer(n); };
if(!Buffer.alloc) Buffer.alloc = function(n) { var b = new Buffer(n); b.fill(0); return b; };
// $FlowIgnore
if(!Buffer.allocUnsafe) Buffer.allocUnsafe = function(n) { return new Buffer(n); };
}
@ -277,7 +277,7 @@ CRC32.str = crc32_str;
/* [MS-CFB] v20171201 */
var CFB = (function _CFB(){
var exports = {};
exports.version = '1.2.0';
exports.version = '1.2.1';
/* [MS-CFB] 2.6.4 */
function namecmp(l, r) {
var L = l.split("/"), R = r.split("/");
@ -689,7 +689,11 @@ function read_file(filename, options) {
}
function read(blob, options) {
switch(options && options.type || "base64") {
var type = options && options.type;
if(!type) {
if(has_buf && Buffer.isBuffer(blob)) type = "buffer";
}
switch(type || "base64") {
case "file": return read_file(blob, options);
case "base64": return parse(s2a(Base64.decode(blob)), options);
case "binary": return parse(s2a(blob), options);
@ -738,22 +742,28 @@ function rebuild_cfb(cfb, f) {
if(!gc && !f) return;
var now = new Date(1987, 1, 19), j = 0;
// Track which names exist
var fullPaths = Object.create ? Object.create(null) : {};
var data = [];
for(i = 0; i < cfb.FullPaths.length; ++i) {
fullPaths[cfb.FullPaths[i]] = true;
if(cfb.FileIndex[i].type === 0) continue;
data.push([cfb.FullPaths[i], cfb.FileIndex[i]]);
}
for(i = 0; i < data.length; ++i) {
var dad = dirname(data[i][0]);
s = false;
for(j = 0; j < data.length; ++j) if(data[j][0] === dad) s = true;
if(!s) data.push([dad, ({
name: filename(dad).replace("/",""),
type: 1,
clsid: HEADER_CLSID,
ct: now, mt: now,
content: null
})]);
s = fullPaths[dad];
if(!s) {
data.push([dad, ({
name: filename(dad).replace("/",""),
type: 1,
clsid: HEADER_CLSID,
ct: now, mt: now,
content: null
})]);
// Add name to set
fullPaths[dad] = true;
}
}
data.sort(function(x,y) { return namecmp(x[0], y[0]); });
@ -908,18 +918,35 @@ flen = file.content.length;
file = cfb.FileIndex[i];
if(file.size >= 0x1000) {
o.l = (file.start+1) << 9;
for(j = 0; j < file.size; ++j) o.write_shift(1, file.content[j]);
for(; j & 0x1FF; ++j) o.write_shift(1, 0);
if (has_buf && Buffer.isBuffer(file.content)) {
file.content.copy(o, o.l, 0, file.size);
// o is a 0-filled Buffer so just set next offset
o.l += (file.size + 511) & -512;
} else {
for(j = 0; j < file.size; ++j) o.write_shift(1, file.content[j]);
for(; j & 0x1FF; ++j) o.write_shift(1, 0);
}
}
}
for(i = 1; i < cfb.FileIndex.length; ++i) {
file = cfb.FileIndex[i];
if(file.size > 0 && file.size < 0x1000) {
for(j = 0; j < file.size; ++j) o.write_shift(1, file.content[j]);
for(; j & 0x3F; ++j) o.write_shift(1, 0);
if (has_buf && Buffer.isBuffer(file.content)) {
file.content.copy(o, o.l, 0, file.size);
// o is a 0-filled Buffer so just set next offset
o.l += (file.size + 63) & -64;
} else {
for(j = 0; j < file.size; ++j) o.write_shift(1, file.content[j]);
for(; j & 0x3F; ++j) o.write_shift(1, 0);
}
}
}
while(o.l < o.length) o.write_shift(1, 0);
if (has_buf) {
o.l = o.length;
} else {
// When using Buffer, already 0-filled
while(o.l < o.length) o.write_shift(1, 0);
}
return o;
}
/* [MS-CFB] 2.6.4 (Unicode 3.0.1 case conversion) */
@ -1058,6 +1085,38 @@ function read_bits_n(buf, bl, n) {
return v & f;
}
/* helpers for unaligned bit writes */
function write_bits_3(buf, bl, v) { var w = bl & 7, h = bl >>> 3;
if(w <= 5) buf[h] |= (v & 7) << w;
else {
buf[h] |= (v << w) & 0xFF;
buf[h+1] = (v&7) >> (8-w);
}
return bl + 3;
}
function write_bits_1(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v = (v&1) << w;
buf[h] |= v;
return bl + 1;
}
function write_bits_8(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v <<= w;
buf[h] |= v & 0xFF; v >>>= 8;
buf[h+1] = v;
return bl + 8;
}
function write_bits_16(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v <<= w;
buf[h] |= v & 0xFF; v >>>= 8;
buf[h+1] = v & 0xFF;
buf[h+2] = v >>> 8;
return bl + 16;
}
/* until ArrayBuffer#realloc is a thing, fake a realloc */
function realloc(b, sz) {
var L = b.length, M = 2*L > sz ? 2*L : sz + 5, i = 0;
@ -1071,7 +1130,7 @@ function realloc(b, sz) {
} else if(use_typed_arrays) {
var a = new Uint8Array(M);
if(a.set) a.set(b);
else for(; i < b.length; ++i) a[i] = b[i];
else for(; i < L; ++i) a[i] = b[i];
return a;
}
b.length = M;
@ -1083,30 +1142,7 @@ function zero_fill_array(n) {
var o = new Array(n);
for(var i = 0; i < n; ++i) o[i] = 0;
return o;
}var _deflate = (function() {
var _deflateRaw = (function() {
return function deflateRaw(data, out) {
var boff = 0;
while(boff < data.length) {
var L = Math.min(0xFFFF, data.length - boff);
var h = boff + L == data.length;
/* TODO: this is only type 0 stored */
out.write_shift(1, +h);
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
}
return out.l;
};
})();
return function(data) {
var buf = new_buf(50+Math.floor(data.length*1.1));
var off = _deflateRaw(data, buf);
return buf.slice(0, off);
};
})();
/* modified inflate function also moves original read head */
}
/* build tree (used for literals and lengths) */
function build_tree(clens, cmap, MAX) {
@ -1146,6 +1182,7 @@ function build_tree(clens, cmap, MAX) {
return maxlen;
}
/* Fixed Huffman */
var fix_lmap = use_typed_arrays ? new Uint16Array(512) : zero_fill_array(512);
var fix_dmap = use_typed_arrays ? new Uint16Array(32) : zero_fill_array(32);
if(!use_typed_arrays) {
@ -1165,8 +1202,124 @@ if(!use_typed_arrays) {
for(; i<=279; i++) clens.push(7);
for(; i<=287; i++) clens.push(8);
build_tree(clens, fix_lmap, 288);
})();var _deflateRaw = (function() {
var DST_LN_RE = use_typed_arrays ? new Uint8Array(0x8000) : [];
for(var j = 0, k = 0; j < DST_LN.length; ++j) {
for(; k < DST_LN[j+1]; ++k) DST_LN_RE[k] = j;
}
for(;k < 32768; ++k) DST_LN_RE[k] = 29;
var LEN_LN_RE = use_typed_arrays ? new Uint8Array(0x102) : [];
for(j = 0, k = 0; j < LEN_LN.length; ++j) {
for(; k < LEN_LN[j+1]; ++k) LEN_LN_RE[k] = j;
}
function write_stored(data, out) {
var boff = 0;
while(boff < data.length) {
var L = Math.min(0xFFFF, data.length - boff);
var h = boff + L == data.length;
out.write_shift(1, +h);
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
}
return out.l;
}
/* Fixed Huffman */
function write_huff_fixed(data, out) {
var bl = 0;
var boff = 0;
var addrs = use_typed_arrays ? new Uint16Array(0x8000) : [];
while(boff < data.length) {
var L = /* data.length - boff; */ Math.min(0xFFFF, data.length - boff);
/* write a stored block for short data */
if(L < 10) {
bl = write_bits_3(out, bl, +!!(boff + L == data.length)); // jshint ignore:line
if(bl & 7) bl += 8 - (bl & 7);
out.l = (bl / 8) | 0;
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
bl = out.l * 8;
continue;
}
bl = write_bits_3(out, bl, +!!(boff + L == data.length) + 2); // jshint ignore:line
var hash = 0;
while(L-- > 0) {
var d = data[boff];
hash = ((hash << 5) ^ d) & 0x7FFF;
var match = -1, mlen = 0;
if((match = addrs[hash])) {
match |= boff & ~0x7FFF;
if(match > boff) match -= 0x8000;
if(match < boff) while(data[match + mlen] == data[boff + mlen] && mlen < 250) ++mlen;
}
if(mlen > 2) {
/* Copy Token */
d = LEN_LN_RE[mlen];
if(d <= 22) bl = write_bits_8(out, bl, bitswap8[d+1]>>1) - 1;
else {
write_bits_8(out, bl, 3);
bl += 5;
write_bits_8(out, bl, bitswap8[d-23]>>5);
bl += 3;
}
var len_eb = (d < 8) ? 0 : ((d - 4)>>2);
if(len_eb > 0) {
write_bits_16(out, bl, mlen - LEN_LN[d]);
bl += len_eb;
}
d = DST_LN_RE[boff - match];
bl = write_bits_8(out, bl, bitswap8[d]>>3);
bl -= 3;
var dst_eb = d < 4 ? 0 : (d-2)>>1;
if(dst_eb > 0) {
write_bits_16(out, bl, boff - match - DST_LN[d]);
bl += dst_eb;
}
for(var q = 0; q < mlen; ++q) {
addrs[hash] = boff & 0x7FFF;
hash = ((hash << 5) ^ data[boff]) & 0x7FFF;
++boff;
}
L-= mlen - 1;
} else {
/* Literal Token */
if(d <= 143) d = d + 48;
else bl = write_bits_1(out, bl, 1);
bl = write_bits_8(out, bl, bitswap8[d]);
addrs[hash] = boff & 0x7FFF;
++boff;
}
}
bl = write_bits_8(out, bl, 0) - 1;
}
out.l = ((bl + 7)/8)|0;
return out.l;
}
return function _deflateRaw(data, out) {
if(data.length < 8) return write_stored(data, out);
return write_huff_fixed(data, out);
};
})();
function _deflate(data) {
var buf = new_buf(50+Math.floor(data.length*1.1));
var off = _deflateRaw(data, buf);
return buf.slice(0, off);
}
/* modified inflate function also moves original read head */
var dyn_lmap = use_typed_arrays ? new Uint16Array(32768) : zero_fill_array(32768);
var dyn_dmap = use_typed_arrays ? new Uint16Array(32768) : zero_fill_array(32768);
var dyn_cmap = use_typed_arrays ? new Uint16Array(128) : zero_fill_array(128);
@ -1285,8 +1438,8 @@ function inflate(data, usz) {
boff = dyn(data, boff);
max_len_1 = dyn_len_1; max_len_2 = dyn_len_2;
}
if(!usz && (OL < woff + 32767)) { outbuf = realloc(outbuf, woff + 32767); OL = outbuf.length; }
for(;;) { // while(true) is apparently out of vogue in modern JS circles
if(!usz && (OL < woff + 32767)) { outbuf = realloc(outbuf, woff + 32767); OL = outbuf.length; }
/* ingest code and move read head */
var bits = read_bits_n(data, boff, max_len_1);
var code = (header>>>1) == 1 ? fix_lmap[bits] : dyn_lmap[bits];
@ -1317,7 +1470,7 @@ function inflate(data, usz) {
}
/* in the common case, manual byte copy is faster than TA set / Buffer copy */
if(!usz && OL < tgt) { outbuf = realloc(outbuf, tgt); OL = outbuf.length; }
if(!usz && OL < tgt) { outbuf = realloc(outbuf, tgt + 100); OL = outbuf.length; }
while(woff < tgt) { outbuf[woff] = outbuf[woff - dst]; ++woff; }
}
}
@ -1416,7 +1569,7 @@ function parse_local_file(blob, csz, usz, o, EF) {
var data = blob.slice(blob.l, blob.l + _csz);
switch(meth) {
case 8: data = _inflateRawSync(blob, _usz); break;
case 0: break;
case 0: break; // TODO: scan for magic number
default: throw new Error("Unsupported ZIP Compression method " + meth);
}
@ -1483,7 +1636,10 @@ function write_zip(cfb, options) {
start_cd += namebuf.length;
out.push(namebuf);
/* TODO: extra fields? */
/* TODO: encryption header ? */
start_cd += outbuf.length;
out.push(outbuf);
@ -1742,7 +1898,8 @@ function write_mad(cfb, options) {
}
out.push(boundary + '--\r\n');
return out.join("\r\n");
}function cfb_new(opts) {
}
function cfb_new(opts) {
var o = ({});
init_cfb(o, opts);
return o;

4
dist/cfb.min.js vendored
View File

File diff suppressed because one or more lines are too long

2
dist/cfb.min.map vendored
View File

File diff suppressed because one or more lines are too long

247
dist/xlscfb.js vendored
View File

@ -143,7 +143,7 @@ CRC32.str = crc32_str;
/* [MS-CFB] v20171201 */
var CFB = (function _CFB(){
var exports/*:CFBModule*/ = /*::(*/{}/*:: :any)*/;
exports.version = '1.2.0';
exports.version = '1.2.1';
/* [MS-CFB] 2.6.4 */
function namecmp(l/*:string*/, r/*:string*/)/*:number*/ {
var L = l.split("/"), R = r.split("/");
@ -555,7 +555,11 @@ function read_file(filename/*:string*/, options/*:CFBReadOpts*/) {
}
function read(blob/*:RawBytes|string*/, options/*:CFBReadOpts*/) {
switch(options && options.type || "base64") {
var type = options && options.type;
if(!type) {
if(has_buf && Buffer.isBuffer(blob)) type = "buffer";
}
switch(type || "base64") {
case "file": /*:: if(typeof blob !== 'string') throw "Must pass a filename when type='file'"; */return read_file(blob, options);
case "base64": /*:: if(typeof blob !== 'string') throw "Must pass a base64-encoded binary string when type='file'"; */return parse(s2a(Base64.decode(blob)), options);
case "binary": /*:: if(typeof blob !== 'string') throw "Must pass a binary string when type='file'"; */return parse(s2a(blob), options);
@ -604,22 +608,28 @@ function rebuild_cfb(cfb/*:CFBContainer*/, f/*:?boolean*/)/*:void*/ {
if(!gc && !f) return;
var now = new Date(1987, 1, 19), j = 0;
// Track which names exist
var fullPaths = Object.create ? Object.create(null) : {};
var data/*:Array<[string, CFBEntry]>*/ = [];
for(i = 0; i < cfb.FullPaths.length; ++i) {
fullPaths[cfb.FullPaths[i]] = true;
if(cfb.FileIndex[i].type === 0) continue;
data.push([cfb.FullPaths[i], cfb.FileIndex[i]]);
}
for(i = 0; i < data.length; ++i) {
var dad = dirname(data[i][0]);
s = false;
for(j = 0; j < data.length; ++j) if(data[j][0] === dad) s = true;
if(!s) data.push([dad, ({
name: filename(dad).replace("/",""),
type: 1,
clsid: HEADER_CLSID,
ct: now, mt: now,
content: null
}/*:any*/)]);
s = fullPaths[dad];
if(!s) {
data.push([dad, ({
name: filename(dad).replace("/",""),
type: 1,
clsid: HEADER_CLSID,
ct: now, mt: now,
content: null
}/*:any*/)]);
// Add name to set
fullPaths[dad] = true;
}
}
data.sort(function(x,y) { return namecmp(x[0], y[0]); });
@ -778,19 +788,36 @@ function _write(cfb/*:CFBContainer*/, options/*:CFBWriteOpts*/)/*:RawBytes|strin
/*:: if(!file.content) throw new Error("unreachable"); */
if(file.size >= 0x1000) {
o.l = (file.start+1) << 9;
for(j = 0; j < file.size; ++j) o.write_shift(1, file.content[j]);
for(; j & 0x1FF; ++j) o.write_shift(1, 0);
if (has_buf && Buffer.isBuffer(file.content)) {
file.content.copy(o, o.l, 0, file.size);
// o is a 0-filled Buffer so just set next offset
o.l += (file.size + 511) & -512;
} else {
for(j = 0; j < file.size; ++j) o.write_shift(1, file.content[j]);
for(; j & 0x1FF; ++j) o.write_shift(1, 0);
}
}
}
for(i = 1; i < cfb.FileIndex.length; ++i) {
file = cfb.FileIndex[i];
/*:: if(!file.content) throw new Error("unreachable"); */
if(file.size > 0 && file.size < 0x1000) {
for(j = 0; j < file.size; ++j) o.write_shift(1, file.content[j]);
for(; j & 0x3F; ++j) o.write_shift(1, 0);
if (has_buf && Buffer.isBuffer(file.content)) {
file.content.copy(o, o.l, 0, file.size);
// o is a 0-filled Buffer so just set next offset
o.l += (file.size + 63) & -64;
} else {
for(j = 0; j < file.size; ++j) o.write_shift(1, file.content[j]);
for(; j & 0x3F; ++j) o.write_shift(1, 0);
}
}
}
while(o.l < o.length) o.write_shift(1, 0);
if (has_buf) {
o.l = o.length;
} else {
// When using Buffer, already 0-filled
while(o.l < o.length) o.write_shift(1, 0);
}
return o;
}
/* [MS-CFB] 2.6.4 (Unicode 3.0.1 case conversion) */
@ -930,6 +957,38 @@ function read_bits_n(buf, bl, n) {
return v & f;
}
/* helpers for unaligned bit writes */
function write_bits_3(buf, bl, v) { var w = bl & 7, h = bl >>> 3;
if(w <= 5) buf[h] |= (v & 7) << w;
else {
buf[h] |= (v << w) & 0xFF;
buf[h+1] = (v&7) >> (8-w);
}
return bl + 3;
}
function write_bits_1(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v = (v&1) << w;
buf[h] |= v;
return bl + 1;
}
function write_bits_8(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v <<= w;
buf[h] |= v & 0xFF; v >>>= 8;
buf[h+1] = v;
return bl + 8;
}
function write_bits_16(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v <<= w;
buf[h] |= v & 0xFF; v >>>= 8;
buf[h+1] = v & 0xFF;
buf[h+2] = v >>> 8;
return bl + 16;
}
/* until ArrayBuffer#realloc is a thing, fake a realloc */
function realloc(b, sz/*:number*/) {
var L = b.length, M = 2*L > sz ? 2*L : sz + 5, i = 0;
@ -943,7 +1002,7 @@ function realloc(b, sz/*:number*/) {
} else if(use_typed_arrays) {
var a = new Uint8Array(M);
if(a.set) a.set(b);
else for(; i < b.length; ++i) a[i] = b[i];
else for(; i < L; ++i) a[i] = b[i];
return a;
}
b.length = M;
@ -955,30 +1014,7 @@ function zero_fill_array(n) {
var o = new Array(n);
for(var i = 0; i < n; ++i) o[i] = 0;
return o;
}var _deflate = (function() {
var _deflateRaw = (function() {
return function deflateRaw(data, out) {
var boff = 0;
while(boff < data.length) {
var L = Math.min(0xFFFF, data.length - boff);
var h = boff + L == data.length;
/* TODO: this is only type 0 stored */
out.write_shift(1, +h);
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
}
return out.l;
};
})();
return function(data) {
var buf = new_buf(50+Math.floor(data.length*1.1));
var off = _deflateRaw(data, buf);
return buf.slice(0, off);
};
})();
/* modified inflate function also moves original read head */
}
/* build tree (used for literals and lengths) */
function build_tree(clens, cmap, MAX/*:number*/)/*:number*/ {
@ -1018,6 +1054,7 @@ function build_tree(clens, cmap, MAX/*:number*/)/*:number*/ {
return maxlen;
}
/* Fixed Huffman */
var fix_lmap = use_typed_arrays ? new Uint16Array(512) : zero_fill_array(512);
var fix_dmap = use_typed_arrays ? new Uint16Array(32) : zero_fill_array(32);
if(!use_typed_arrays) {
@ -1037,8 +1074,124 @@ if(!use_typed_arrays) {
for(; i<=279; i++) clens.push(7);
for(; i<=287; i++) clens.push(8);
build_tree(clens, fix_lmap, 288);
})();var _deflateRaw = (function() {
var DST_LN_RE = use_typed_arrays ? new Uint8Array(0x8000) : [];
for(var j = 0, k = 0; j < DST_LN.length; ++j) {
for(; k < DST_LN[j+1]; ++k) DST_LN_RE[k] = j;
}
for(;k < 32768; ++k) DST_LN_RE[k] = 29;
var LEN_LN_RE = use_typed_arrays ? new Uint8Array(0x102) : [];
for(j = 0, k = 0; j < LEN_LN.length; ++j) {
for(; k < LEN_LN[j+1]; ++k) LEN_LN_RE[k] = j;
}
function write_stored(data, out) {
var boff = 0;
while(boff < data.length) {
var L = Math.min(0xFFFF, data.length - boff);
var h = boff + L == data.length;
out.write_shift(1, +h);
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
}
return out.l;
}
/* Fixed Huffman */
function write_huff_fixed(data, out) {
var bl = 0;
var boff = 0;
var addrs = use_typed_arrays ? new Uint16Array(0x8000) : [];
while(boff < data.length) {
var L = /* data.length - boff; */ Math.min(0xFFFF, data.length - boff);
/* write a stored block for short data */
if(L < 10) {
bl = write_bits_3(out, bl, +!!(boff + L == data.length)); // jshint ignore:line
if(bl & 7) bl += 8 - (bl & 7);
out.l = (bl / 8) | 0;
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
bl = out.l * 8;
continue;
}
bl = write_bits_3(out, bl, +!!(boff + L == data.length) + 2); // jshint ignore:line
var hash = 0;
while(L-- > 0) {
var d = data[boff];
hash = ((hash << 5) ^ d) & 0x7FFF;
var match = -1, mlen = 0;
if((match = addrs[hash])) {
match |= boff & ~0x7FFF;
if(match > boff) match -= 0x8000;
if(match < boff) while(data[match + mlen] == data[boff + mlen] && mlen < 250) ++mlen;
}
if(mlen > 2) {
/* Copy Token */
d = LEN_LN_RE[mlen];
if(d <= 22) bl = write_bits_8(out, bl, bitswap8[d+1]>>1) - 1;
else {
write_bits_8(out, bl, 3);
bl += 5;
write_bits_8(out, bl, bitswap8[d-23]>>5);
bl += 3;
}
var len_eb = (d < 8) ? 0 : ((d - 4)>>2);
if(len_eb > 0) {
write_bits_16(out, bl, mlen - LEN_LN[d]);
bl += len_eb;
}
d = DST_LN_RE[boff - match];
bl = write_bits_8(out, bl, bitswap8[d]>>3);
bl -= 3;
var dst_eb = d < 4 ? 0 : (d-2)>>1;
if(dst_eb > 0) {
write_bits_16(out, bl, boff - match - DST_LN[d]);
bl += dst_eb;
}
for(var q = 0; q < mlen; ++q) {
addrs[hash] = boff & 0x7FFF;
hash = ((hash << 5) ^ data[boff]) & 0x7FFF;
++boff;
}
L-= mlen - 1;
} else {
/* Literal Token */
if(d <= 143) d = d + 48;
else bl = write_bits_1(out, bl, 1);
bl = write_bits_8(out, bl, bitswap8[d]);
addrs[hash] = boff & 0x7FFF;
++boff;
}
}
bl = write_bits_8(out, bl, 0) - 1;
}
out.l = ((bl + 7)/8)|0;
return out.l;
}
return function _deflateRaw(data, out) {
if(data.length < 8) return write_stored(data, out);
return write_huff_fixed(data, out);
};
})();
function _deflate(data) {
var buf = new_buf(50+Math.floor(data.length*1.1));
var off = _deflateRaw(data, buf);
return buf.slice(0, off);
}
/* modified inflate function also moves original read head */
var dyn_lmap = use_typed_arrays ? new Uint16Array(32768) : zero_fill_array(32768);
var dyn_dmap = use_typed_arrays ? new Uint16Array(32768) : zero_fill_array(32768);
var dyn_cmap = use_typed_arrays ? new Uint16Array(128) : zero_fill_array(128);
@ -1157,8 +1310,8 @@ function inflate(data, usz/*:number*/) {
boff = dyn(data, boff);
max_len_1 = dyn_len_1; max_len_2 = dyn_len_2;
}
if(!usz && (OL < woff + 32767)) { outbuf = realloc(outbuf, woff + 32767); OL = outbuf.length; }
for(;;) { // while(true) is apparently out of vogue in modern JS circles
if(!usz && (OL < woff + 32767)) { outbuf = realloc(outbuf, woff + 32767); OL = outbuf.length; }
/* ingest code and move read head */
var bits = read_bits_n(data, boff, max_len_1);
var code = (header>>>1) == 1 ? fix_lmap[bits] : dyn_lmap[bits];
@ -1189,7 +1342,7 @@ function inflate(data, usz/*:number*/) {
}
/* in the common case, manual byte copy is faster than TA set / Buffer copy */
if(!usz && OL < tgt) { outbuf = realloc(outbuf, tgt); OL = outbuf.length; }
if(!usz && OL < tgt) { outbuf = realloc(outbuf, tgt + 100); OL = outbuf.length; }
while(woff < tgt) { outbuf[woff] = outbuf[woff - dst]; ++woff; }
}
}
@ -1288,7 +1441,7 @@ function parse_local_file(blob/*:CFBlob*/, csz/*:number*/, usz/*:number*/, o/*:C
var data = blob.slice(blob.l, blob.l + _csz);
switch(meth) {
case 8: data = _inflateRawSync(blob, _usz); break;
case 0: break;
case 0: break; // TODO: scan for magic number
default: throw new Error("Unsupported ZIP Compression method " + meth);
}
@ -1355,7 +1508,10 @@ function write_zip(cfb/*:CFBContainer*/, options/*:CFBWriteOpts*/)/*:RawBytes*/
start_cd += namebuf.length;
out.push(namebuf);
/* TODO: extra fields? */
/* TODO: encryption header ? */
start_cd += outbuf.length;
out.push(outbuf);
@ -1614,7 +1770,8 @@ function write_mad(cfb/*:CFBContainer*/, options/*:CFBWriteOpts*/)/*:string*/ {
}
out.push(boundary + '--\r\n');
return out.join("\r\n");
}function cfb_new(opts/*:?any*/)/*:CFBContainer*/ {
}
function cfb_new(opts/*:?any*/)/*:CFBContainer*/ {
var o/*:CFBContainer*/ = ({}/*:any*/);
init_cfb(o, opts);
return o;

21
index.html
View File

@ -53,6 +53,12 @@ Use readAsBinaryString: (when available) <input type="checkbox" name="userabs" c
/* eslint no-use-before-define:0 */
var global_cfb;
if(!String.prototype.repeat) String.prototype.repeat = function(count) {
var o = "";
for(var i = 0; i < count; ++i) o += this;
return o;
};
var get_manifest = (function() {
var sprintf = PRINTJ.sprintf;
function fix_string(x/*:string*/)/*:string*/ { return x.replace(/[\u0000-\u001f]/, function($$) { return sprintf("\\u%04X", $$.charCodeAt(0)); }); }
@ -99,20 +105,13 @@ var do_file = (function() {
var domrabs = document.getElementsByName("userabs")[0];
if(!rABS) domrabs.disabled = !(domrabs.checked = false);
function fixdata(data) {
var o = "", l = 0, w = 10240;
for(; l<data.byteLength/w; ++l) o+=String.fromCharCode.apply(null,new Uint8Array(data.slice(l*w,l*w+w)));
o+=String.fromCharCode.apply(null, new Uint8Array(data.slice(l*w)));
return o;
}
return function do_file(files) {
rABS = domrabs.checked;
var f = files[0];
var reader = new FileReader();
reader.onload = function(e) {
var data = e.target.result;
var cfb = CFB.read(rABS ? data : btoa(fixdata(data)), {type: rABS ? 'binary' : 'base64'});
var cfb = CFB.read(rABS ? data : new Uint8Array(data), {type: rABS ? 'binary' : 'buffer'});
process_data(cfb);
};
if(rABS) reader.readAsBinaryString(f);
@ -161,7 +160,11 @@ var savefile = (function() {
if(typeof console !== 'undefined') console.log(global_cfb);
var data = CFB.write(global_cfb, {type:'binary', fileType: type});
if(typeof console !== 'undefined') console.log(data);
saveAs(new Blob([s2ab(data)],{type:"application/octet-stream"}), "SheetJSCFBDemo." + type);
var fname = "SheetJSCFBDemo." + type;
var blob = new Blob([s2ab(data)],{type:"application/octet-stream"});
if(typeof navigator !== 'undefined' && navigator.msSaveBlob) return navigator.msSaveBlob(blob, fname);
saveAs(blob, fname);
};
})();

6
package.json
View File

@ -1,6 +1,6 @@
{
"name": "cfb",
"version": "1.2.0",
"version": "1.2.1",
"author": "sheetjs",
"description": "Compound File Binary File Format extractor",
"keywords": [
@ -16,9 +16,9 @@
"fs": false
},
"dependencies": {
"adler-32": "~1.2.0",
"adler-32": "~1.3.0",
"crc-32": "~1.2.0",
"printj": "~1.1.2"
"printj": "~1.3.0"
},
"devDependencies": {
"@sheetjs/uglify-js": "~2.7.3",

196
xlscfb.flow.js
View File

@ -143,7 +143,7 @@ CRC32.str = crc32_str;
/* [MS-CFB] v20171201 */
var CFB = (function _CFB(){
var exports/*:CFBModule*/ = /*::(*/{}/*:: :any)*/;
exports.version = '1.2.0';
exports.version = '1.2.1';
/* [MS-CFB] 2.6.4 */
function namecmp(l/*:string*/, r/*:string*/)/*:number*/ {
var L = l.split("/"), R = r.split("/");
@ -555,7 +555,11 @@ function read_file(filename/*:string*/, options/*:CFBReadOpts*/) {
}
function read(blob/*:RawBytes|string*/, options/*:CFBReadOpts*/) {
switch(options && options.type || "base64") {
var type = options && options.type;
if(!type) {
if(has_buf && Buffer.isBuffer(blob)) type = "buffer";
}
switch(type || "base64") {
case "file": /*:: if(typeof blob !== 'string') throw "Must pass a filename when type='file'"; */return read_file(blob, options);
case "base64": /*:: if(typeof blob !== 'string') throw "Must pass a base64-encoded binary string when type='file'"; */return parse(s2a(Base64.decode(blob)), options);
case "binary": /*:: if(typeof blob !== 'string') throw "Must pass a binary string when type='file'"; */return parse(s2a(blob), options);
@ -953,6 +957,38 @@ function read_bits_n(buf, bl, n) {
return v & f;
}
/* helpers for unaligned bit writes */
function write_bits_3(buf, bl, v) { var w = bl & 7, h = bl >>> 3;
if(w <= 5) buf[h] |= (v & 7) << w;
else {
buf[h] |= (v << w) & 0xFF;
buf[h+1] = (v&7) >> (8-w);
}
return bl + 3;
}
function write_bits_1(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v = (v&1) << w;
buf[h] |= v;
return bl + 1;
}
function write_bits_8(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v <<= w;
buf[h] |= v & 0xFF; v >>>= 8;
buf[h+1] = v;
return bl + 8;
}
function write_bits_16(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v <<= w;
buf[h] |= v & 0xFF; v >>>= 8;
buf[h+1] = v & 0xFF;
buf[h+2] = v >>> 8;
return bl + 16;
}
/* until ArrayBuffer#realloc is a thing, fake a realloc */
function realloc(b, sz/*:number*/) {
var L = b.length, M = 2*L > sz ? 2*L : sz + 5, i = 0;
@ -966,7 +1002,7 @@ function realloc(b, sz/*:number*/) {
} else if(use_typed_arrays) {
var a = new Uint8Array(M);
if(a.set) a.set(b);
else for(; i < b.length; ++i) a[i] = b[i];
else for(; i < L; ++i) a[i] = b[i];
return a;
}
b.length = M;
@ -978,30 +1014,7 @@ function zero_fill_array(n) {
var o = new Array(n);
for(var i = 0; i < n; ++i) o[i] = 0;
return o;
}var _deflate = (function() {
var _deflateRaw = (function() {
return function deflateRaw(data, out) {
var boff = 0;
while(boff < data.length) {
var L = Math.min(0xFFFF, data.length - boff);
var h = boff + L == data.length;
/* TODO: this is only type 0 stored */
out.write_shift(1, +h);
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
}
return out.l;
};
})();
return function(data) {
var buf = new_buf(50+Math.floor(data.length*1.1));
var off = _deflateRaw(data, buf);
return buf.slice(0, off);
};
})();
/* modified inflate function also moves original read head */
}
/* build tree (used for literals and lengths) */
function build_tree(clens, cmap, MAX/*:number*/)/*:number*/ {
@ -1041,6 +1054,7 @@ function build_tree(clens, cmap, MAX/*:number*/)/*:number*/ {
return maxlen;
}
/* Fixed Huffman */
var fix_lmap = use_typed_arrays ? new Uint16Array(512) : zero_fill_array(512);
var fix_dmap = use_typed_arrays ? new Uint16Array(32) : zero_fill_array(32);
if(!use_typed_arrays) {
@ -1060,8 +1074,124 @@ if(!use_typed_arrays) {
for(; i<=279; i++) clens.push(7);
for(; i<=287; i++) clens.push(8);
build_tree(clens, fix_lmap, 288);
})();var _deflateRaw = (function() {
var DST_LN_RE = use_typed_arrays ? new Uint8Array(0x8000) : [];
for(var j = 0, k = 0; j < DST_LN.length; ++j) {
for(; k < DST_LN[j+1]; ++k) DST_LN_RE[k] = j;
}
for(;k < 32768; ++k) DST_LN_RE[k] = 29;
var LEN_LN_RE = use_typed_arrays ? new Uint8Array(0x102) : [];
for(j = 0, k = 0; j < LEN_LN.length; ++j) {
for(; k < LEN_LN[j+1]; ++k) LEN_LN_RE[k] = j;
}
function write_stored(data, out) {
var boff = 0;
while(boff < data.length) {
var L = Math.min(0xFFFF, data.length - boff);
var h = boff + L == data.length;
out.write_shift(1, +h);
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
}
return out.l;
}
/* Fixed Huffman */
function write_huff_fixed(data, out) {
var bl = 0;
var boff = 0;
var addrs = use_typed_arrays ? new Uint16Array(0x8000) : [];
while(boff < data.length) {
var L = /* data.length - boff; */ Math.min(0xFFFF, data.length - boff);
/* write a stored block for short data */
if(L < 10) {
bl = write_bits_3(out, bl, +!!(boff + L == data.length)); // jshint ignore:line
if(bl & 7) bl += 8 - (bl & 7);
out.l = (bl / 8) | 0;
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
bl = out.l * 8;
continue;
}
bl = write_bits_3(out, bl, +!!(boff + L == data.length) + 2); // jshint ignore:line
var hash = 0;
while(L-- > 0) {
var d = data[boff];
hash = ((hash << 5) ^ d) & 0x7FFF;
var match = -1, mlen = 0;
if((match = addrs[hash])) {
match |= boff & ~0x7FFF;
if(match > boff) match -= 0x8000;
if(match < boff) while(data[match + mlen] == data[boff + mlen] && mlen < 250) ++mlen;
}
if(mlen > 2) {
/* Copy Token */
d = LEN_LN_RE[mlen];
if(d <= 22) bl = write_bits_8(out, bl, bitswap8[d+1]>>1) - 1;
else {
write_bits_8(out, bl, 3);
bl += 5;
write_bits_8(out, bl, bitswap8[d-23]>>5);
bl += 3;
}
var len_eb = (d < 8) ? 0 : ((d - 4)>>2);
if(len_eb > 0) {
write_bits_16(out, bl, mlen - LEN_LN[d]);
bl += len_eb;
}
d = DST_LN_RE[boff - match];
bl = write_bits_8(out, bl, bitswap8[d]>>3);
bl -= 3;
var dst_eb = d < 4 ? 0 : (d-2)>>1;
if(dst_eb > 0) {
write_bits_16(out, bl, boff - match - DST_LN[d]);
bl += dst_eb;
}
for(var q = 0; q < mlen; ++q) {
addrs[hash] = boff & 0x7FFF;
hash = ((hash << 5) ^ data[boff]) & 0x7FFF;
++boff;
}
L-= mlen - 1;
} else {
/* Literal Token */
if(d <= 143) d = d + 48;
else bl = write_bits_1(out, bl, 1);
bl = write_bits_8(out, bl, bitswap8[d]);
addrs[hash] = boff & 0x7FFF;
++boff;
}
}
bl = write_bits_8(out, bl, 0) - 1;
}
out.l = ((bl + 7)/8)|0;
return out.l;
}
return function _deflateRaw(data, out) {
if(data.length < 8) return write_stored(data, out);
return write_huff_fixed(data, out);
};
})();
function _deflate(data) {
var buf = new_buf(50+Math.floor(data.length*1.1));
var off = _deflateRaw(data, buf);
return buf.slice(0, off);
}
/* modified inflate function also moves original read head */
var dyn_lmap = use_typed_arrays ? new Uint16Array(32768) : zero_fill_array(32768);
var dyn_dmap = use_typed_arrays ? new Uint16Array(32768) : zero_fill_array(32768);
var dyn_cmap = use_typed_arrays ? new Uint16Array(128) : zero_fill_array(128);
@ -1180,8 +1310,8 @@ function inflate(data, usz/*:number*/) {
boff = dyn(data, boff);
max_len_1 = dyn_len_1; max_len_2 = dyn_len_2;
}
if(!usz && (OL < woff + 32767)) { outbuf = realloc(outbuf, woff + 32767); OL = outbuf.length; }
for(;;) { // while(true) is apparently out of vogue in modern JS circles
if(!usz && (OL < woff + 32767)) { outbuf = realloc(outbuf, woff + 32767); OL = outbuf.length; }
/* ingest code and move read head */
var bits = read_bits_n(data, boff, max_len_1);
var code = (header>>>1) == 1 ? fix_lmap[bits] : dyn_lmap[bits];
@ -1212,7 +1342,7 @@ function inflate(data, usz/*:number*/) {
}
/* in the common case, manual byte copy is faster than TA set / Buffer copy */
if(!usz && OL < tgt) { outbuf = realloc(outbuf, tgt); OL = outbuf.length; }
if(!usz && OL < tgt) { outbuf = realloc(outbuf, tgt + 100); OL = outbuf.length; }
while(woff < tgt) { outbuf[woff] = outbuf[woff - dst]; ++woff; }
}
}
@ -1311,7 +1441,7 @@ function parse_local_file(blob/*:CFBlob*/, csz/*:number*/, usz/*:number*/, o/*:C
var data = blob.slice(blob.l, blob.l + _csz);
switch(meth) {
case 8: data = _inflateRawSync(blob, _usz); break;
case 0: break;
case 0: break; // TODO: scan for magic number
default: throw new Error("Unsupported ZIP Compression method " + meth);
}
@ -1378,7 +1508,10 @@ function write_zip(cfb/*:CFBContainer*/, options/*:CFBWriteOpts*/)/*:RawBytes*/
start_cd += namebuf.length;
out.push(namebuf);
/* TODO: extra fields? */
/* TODO: encryption header ? */
start_cd += outbuf.length;
out.push(outbuf);
@ -1637,7 +1770,8 @@ function write_mad(cfb/*:CFBContainer*/, options/*:CFBWriteOpts*/)/*:string*/ {
}
out.push(boundary + '--\r\n');
return out.join("\r\n");
}function cfb_new(opts/*:?any*/)/*:CFBContainer*/ {
}
function cfb_new(opts/*:?any*/)/*:CFBContainer*/ {
var o/*:CFBContainer*/ = ({}/*:any*/);
init_cfb(o, opts);
return o;

196
xlscfb.js
View File

@ -109,7 +109,7 @@ CRC32.str = crc32_str;
/* [MS-CFB] v20171201 */
var CFB = (function _CFB(){
var exports = {};
exports.version = '1.2.0';
exports.version = '1.2.1';
/* [MS-CFB] 2.6.4 */
function namecmp(l, r) {
var L = l.split("/"), R = r.split("/");
@ -521,7 +521,11 @@ function read_file(filename, options) {
}
function read(blob, options) {
switch(options && options.type || "base64") {
var type = options && options.type;
if(!type) {
if(has_buf && Buffer.isBuffer(blob)) type = "buffer";
}
switch(type || "base64") {
case "file": return read_file(blob, options);
case "base64": return parse(s2a(Base64.decode(blob)), options);
case "binary": return parse(s2a(blob), options);
@ -913,6 +917,38 @@ function read_bits_n(buf, bl, n) {
return v & f;
}
/* helpers for unaligned bit writes */
function write_bits_3(buf, bl, v) { var w = bl & 7, h = bl >>> 3;
if(w <= 5) buf[h] |= (v & 7) << w;
else {
buf[h] |= (v << w) & 0xFF;
buf[h+1] = (v&7) >> (8-w);
}
return bl + 3;
}
function write_bits_1(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v = (v&1) << w;
buf[h] |= v;
return bl + 1;
}
function write_bits_8(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v <<= w;
buf[h] |= v & 0xFF; v >>>= 8;
buf[h+1] = v;
return bl + 8;
}
function write_bits_16(buf, bl, v) {
var w = bl & 7, h = bl >>> 3;
v <<= w;
buf[h] |= v & 0xFF; v >>>= 8;
buf[h+1] = v & 0xFF;
buf[h+2] = v >>> 8;
return bl + 16;
}
/* until ArrayBuffer#realloc is a thing, fake a realloc */
function realloc(b, sz) {
var L = b.length, M = 2*L > sz ? 2*L : sz + 5, i = 0;
@ -926,7 +962,7 @@ function realloc(b, sz) {
} else if(use_typed_arrays) {
var a = new Uint8Array(M);
if(a.set) a.set(b);
else for(; i < b.length; ++i) a[i] = b[i];
else for(; i < L; ++i) a[i] = b[i];
return a;
}
b.length = M;
@ -938,30 +974,7 @@ function zero_fill_array(n) {
var o = new Array(n);
for(var i = 0; i < n; ++i) o[i] = 0;
return o;
}var _deflate = (function() {
var _deflateRaw = (function() {
return function deflateRaw(data, out) {
var boff = 0;
while(boff < data.length) {
var L = Math.min(0xFFFF, data.length - boff);
var h = boff + L == data.length;
/* TODO: this is only type 0 stored */
out.write_shift(1, +h);
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
}
return out.l;
};
})();
return function(data) {
var buf = new_buf(50+Math.floor(data.length*1.1));
var off = _deflateRaw(data, buf);
return buf.slice(0, off);
};
})();
/* modified inflate function also moves original read head */
}
/* build tree (used for literals and lengths) */
function build_tree(clens, cmap, MAX) {
@ -1001,6 +1014,7 @@ function build_tree(clens, cmap, MAX) {
return maxlen;
}
/* Fixed Huffman */
var fix_lmap = use_typed_arrays ? new Uint16Array(512) : zero_fill_array(512);
var fix_dmap = use_typed_arrays ? new Uint16Array(32) : zero_fill_array(32);
if(!use_typed_arrays) {
@ -1020,8 +1034,124 @@ if(!use_typed_arrays) {
for(; i<=279; i++) clens.push(7);
for(; i<=287; i++) clens.push(8);
build_tree(clens, fix_lmap, 288);
})();var _deflateRaw = (function() {
var DST_LN_RE = use_typed_arrays ? new Uint8Array(0x8000) : [];
for(var j = 0, k = 0; j < DST_LN.length; ++j) {
for(; k < DST_LN[j+1]; ++k) DST_LN_RE[k] = j;
}
for(;k < 32768; ++k) DST_LN_RE[k] = 29;
var LEN_LN_RE = use_typed_arrays ? new Uint8Array(0x102) : [];
for(j = 0, k = 0; j < LEN_LN.length; ++j) {
for(; k < LEN_LN[j+1]; ++k) LEN_LN_RE[k] = j;
}
function write_stored(data, out) {
var boff = 0;
while(boff < data.length) {
var L = Math.min(0xFFFF, data.length - boff);
var h = boff + L == data.length;
out.write_shift(1, +h);
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
}
return out.l;
}
/* Fixed Huffman */
function write_huff_fixed(data, out) {
var bl = 0;
var boff = 0;
var addrs = use_typed_arrays ? new Uint16Array(0x8000) : [];
while(boff < data.length) {
var L = /* data.length - boff; */ Math.min(0xFFFF, data.length - boff);
/* write a stored block for short data */
if(L < 10) {
bl = write_bits_3(out, bl, +!!(boff + L == data.length)); // jshint ignore:line
if(bl & 7) bl += 8 - (bl & 7);
out.l = (bl / 8) | 0;
out.write_shift(2, L);
out.write_shift(2, (~L) & 0xFFFF);
while(L-- > 0) out[out.l++] = data[boff++];
bl = out.l * 8;
continue;
}
bl = write_bits_3(out, bl, +!!(boff + L == data.length) + 2); // jshint ignore:line
var hash = 0;
while(L-- > 0) {
var d = data[boff];
hash = ((hash << 5) ^ d) & 0x7FFF;
var match = -1, mlen = 0;
if((match = addrs[hash])) {
match |= boff & ~0x7FFF;
if(match > boff) match -= 0x8000;
if(match < boff) while(data[match + mlen] == data[boff + mlen] && mlen < 250) ++mlen;
}
if(mlen > 2) {
/* Copy Token */
d = LEN_LN_RE[mlen];
if(d <= 22) bl = write_bits_8(out, bl, bitswap8[d+1]>>1) - 1;
else {
write_bits_8(out, bl, 3);
bl += 5;
write_bits_8(out, bl, bitswap8[d-23]>>5);
bl += 3;
}
var len_eb = (d < 8) ? 0 : ((d - 4)>>2);
if(len_eb > 0) {
write_bits_16(out, bl, mlen - LEN_LN[d]);
bl += len_eb;
}
d = DST_LN_RE[boff - match];
bl = write_bits_8(out, bl, bitswap8[d]>>3);
bl -= 3;
var dst_eb = d < 4 ? 0 : (d-2)>>1;
if(dst_eb > 0) {
write_bits_16(out, bl, boff - match - DST_LN[d]);
bl += dst_eb;
}
for(var q = 0; q < mlen; ++q) {
addrs[hash] = boff & 0x7FFF;
hash = ((hash << 5) ^ data[boff]) & 0x7FFF;
++boff;
}
L-= mlen - 1;
} else {
/* Literal Token */
if(d <= 143) d = d + 48;
else bl = write_bits_1(out, bl, 1);
bl = write_bits_8(out, bl, bitswap8[d]);
addrs[hash] = boff & 0x7FFF;
++boff;
}
}
bl = write_bits_8(out, bl, 0) - 1;
}
out.l = ((bl + 7)/8)|0;
return out.l;
}
return function _deflateRaw(data, out) {
if(data.length < 8) return write_stored(data, out);
return write_huff_fixed(data, out);
};
})();
function _deflate(data) {
var buf = new_buf(50+Math.floor(data.length*1.1));
var off = _deflateRaw(data, buf);
return buf.slice(0, off);
}
/* modified inflate function also moves original read head */
var dyn_lmap = use_typed_arrays ? new Uint16Array(32768) : zero_fill_array(32768);
var dyn_dmap = use_typed_arrays ? new Uint16Array(32768) : zero_fill_array(32768);
var dyn_cmap = use_typed_arrays ? new Uint16Array(128) : zero_fill_array(128);
@ -1140,8 +1270,8 @@ function inflate(data, usz) {
boff = dyn(data, boff);
max_len_1 = dyn_len_1; max_len_2 = dyn_len_2;
}
if(!usz && (OL < woff + 32767)) { outbuf = realloc(outbuf, woff + 32767); OL = outbuf.length; }
for(;;) { // while(true) is apparently out of vogue in modern JS circles
if(!usz && (OL < woff + 32767)) { outbuf = realloc(outbuf, woff + 32767); OL = outbuf.length; }
/* ingest code and move read head */
var bits = read_bits_n(data, boff, max_len_1);
var code = (header>>>1) == 1 ? fix_lmap[bits] : dyn_lmap[bits];
@ -1172,7 +1302,7 @@ function inflate(data, usz) {
}
/* in the common case, manual byte copy is faster than TA set / Buffer copy */
if(!usz && OL < tgt) { outbuf = realloc(outbuf, tgt); OL = outbuf.length; }
if(!usz && OL < tgt) { outbuf = realloc(outbuf, tgt + 100); OL = outbuf.length; }
while(woff < tgt) { outbuf[woff] = outbuf[woff - dst]; ++woff; }
}
}
@ -1271,7 +1401,7 @@ function parse_local_file(blob, csz, usz, o, EF) {
var data = blob.slice(blob.l, blob.l + _csz);
switch(meth) {
case 8: data = _inflateRawSync(blob, _usz); break;
case 0: break;
case 0: break; // TODO: scan for magic number
default: throw new Error("Unsupported ZIP Compression method " + meth);
}
@ -1338,7 +1468,10 @@ function write_zip(cfb, options) {
start_cd += namebuf.length;
out.push(namebuf);
/* TODO: extra fields? */
/* TODO: encryption header ? */
start_cd += outbuf.length;
out.push(outbuf);
@ -1597,7 +1730,8 @@ function write_mad(cfb, options) {
}
out.push(boundary + '--\r\n');
return out.join("\r\n");
}function cfb_new(opts) {
}
function cfb_new(opts) {
var o = ({});
init_cfb(o, opts);
return o;