version bump 0.10.2: proper directory/FAT analysis
- reprocess directory sector lists if not available - search now strips out OLE metacharacters \0-\5 - strip sourceMappingURL references from minified scripts (h/t @vinin3) - added xlscfb.js for js-xls
This commit is contained in:
parent
597de28cf5
commit
fcc05e3567
|
|
@ -1,6 +1,8 @@
|
|||
test_files
|
||||
node_modules/
|
||||
*.xls
|
||||
node_modules
|
||||
misc/coverage.html
|
||||
prof.js
|
||||
v8.log
|
||||
test_files
|
||||
test_files_pres
|
||||
*.xls
|
||||
*.sheetjs
|
||||
|
|
|
|||
50
Makefile
50
Makefile
|
|
@ -1,7 +1,16 @@
|
|||
LIB=cfb
|
||||
FMT=xls doc ppt misc full
|
||||
REQS=
|
||||
ADDONS=
|
||||
AUXTARGETS=xlscfb.js
|
||||
|
||||
ULIB=$(shell echo $(LIB) | tr a-z A-Z)
|
||||
DEPS=$(sort $(wildcard bits/*.js))
|
||||
TARGET=$(LIB).js
|
||||
|
||||
.PHONY: all
|
||||
all: $(TARGET) $(AUXTARGETS)
|
||||
|
||||
$(TARGET): $(DEPS)
|
||||
cat $^ | tr -d '\15\32' > $@
|
||||
|
||||
|
|
@ -11,32 +20,48 @@ bits/31_version.js: package.json
|
|||
.PHONY: clean
|
||||
clean:
|
||||
rm -f $(TARGET)
|
||||
rm -rf ./test_files/
|
||||
|
||||
.PHONY: clean-data
|
||||
clean-data:
|
||||
rm -fr ./test_files/ ./test_files_pres/
|
||||
|
||||
.PHONY: init
|
||||
init:
|
||||
if [ ! -e test_files ]; then git clone https://github.com/SheetJS/test_files; fi
|
||||
cd test_files; git pull; make
|
||||
if [ ! -e test_files_pres ]; then git clone https://github.com/SheetJS/test_files_pres; fi
|
||||
cd test_files_pres; git pull
|
||||
|
||||
.PHONY: test mocha
|
||||
test mocha: test.js
|
||||
mocha -R spec
|
||||
mocha -R spec -t 20000
|
||||
|
||||
.PHONY: prof
|
||||
prof:
|
||||
cat misc/prof.js test.js > prof.js
|
||||
node --prof prof.js
|
||||
|
||||
TESTFMT=$(patsubst %,test_%,$(FMT))
|
||||
.PHONY: $(TESTFMT)
|
||||
$(TESTFMT): test_%:
|
||||
FMTS=$* make test
|
||||
|
||||
|
||||
.PHONY: lint
|
||||
lint: $(TARGET)
|
||||
jshint --show-non-errors $(TARGET)
|
||||
jscs $(TARGET)
|
||||
jshint --show-non-errors $(TARGET) $(AUXTARGETS)
|
||||
jscs $(TARGET) $(AUXTARGETS)
|
||||
|
||||
.PHONY: cov cov-spin
|
||||
cov: misc/coverage.html
|
||||
cov-spin:
|
||||
make cov & bash misc/spin.sh $$!
|
||||
|
||||
COVFMT=$(patsubst %,cov_%,$(FMT))
|
||||
.PHONY: $(COVFMT)
|
||||
$(COVFMT): cov_%:
|
||||
FMTS=$* make cov
|
||||
|
||||
misc/coverage.html: $(TARGET) test.js
|
||||
mocha --require blanket -R html-cov > $@
|
||||
|
||||
|
|
@ -48,7 +73,22 @@ coveralls-spin:
|
|||
make coveralls & bash misc/spin.sh $$!
|
||||
|
||||
.PHONY: dist
|
||||
dist: $(TARGET)
|
||||
dist: dist-deps $(TARGET)
|
||||
cp $(TARGET) dist/
|
||||
cp LICENSE dist/
|
||||
uglifyjs $(TARGET) -o dist/$(LIB).min.js --source-map dist/$(LIB).min.map --preamble "$$(head -n 1 bits/00_header.js)"
|
||||
misc/strip_sourcemap.sh dist/$(LIB).min.js
|
||||
|
||||
.PHONY: aux
|
||||
aux: $(AUXTARGETS)
|
||||
|
||||
.PHONY: xls
|
||||
xls: xlscfb.js
|
||||
|
||||
XLSDEPS=misc/suppress_export.js $(filter-out bits/08_blob.js,$(DEPS))
|
||||
xlscfb.js: $(XLSDEPS)
|
||||
cat $^ | tr -d '\15\32' > $@
|
||||
|
||||
.PHONY: dist-deps
|
||||
dist-deps: xlscfb.js
|
||||
cp xlscfb.js dist/xlscfb.js
|
||||
|
|
|
|||
23
README.md
23
README.md
|
|
@ -1,8 +1,8 @@
|
|||
# Compound File Binary Format
|
||||
# Compound File Binary Format
|
||||
|
||||
This is a Pure-JS implementation of MS-CFB: Compound File Binary File Format, a
|
||||
format used in many Microsoft file types (such as XLS, DOC, and other Microsoft
|
||||
Office file types).
|
||||
Office file types).
|
||||
|
||||
# Utility Installation and Usage
|
||||
|
||||
|
|
@ -14,7 +14,7 @@ $ cfb path/to/CFB/file
|
|||
```
|
||||
|
||||
The command will extract the storages and streams in the container, generating
|
||||
files that line up with the tree-based structure of the storage. Metadata
|
||||
files that line up with the tree-based structure of the storage. Metadata
|
||||
such as the red-black tree are discarded (and in the future, new CFB containers
|
||||
will exclusively use black nodes)
|
||||
|
||||
|
|
@ -55,9 +55,9 @@ It has the following properties and methods:
|
|||
|
||||
- `.find(path)` performs a case-insensitive match for the path (or file name, if
|
||||
there are no slashes) and returns an entry object (described later) or null if
|
||||
not found
|
||||
not found
|
||||
|
||||
- `.FullPaths` is an array of the names of all of the streams (files) and
|
||||
- `.FullPaths` is an array of the names of all of the streams (files) and
|
||||
storages (directories) in the container. The paths are properly prefixed from
|
||||
the root entry (so the entries are unique)
|
||||
|
||||
|
|
@ -67,24 +67,27 @@ It has the following properties and methods:
|
|||
- `.FileIndex` is an array of the objects from `.FullPathDir`, in the same order
|
||||
as `.FullPaths`.
|
||||
|
||||
- `.raw` contains the raw header and sectors
|
||||
- `.raw` contains the raw header and sectors
|
||||
|
||||
## Entry Object Description
|
||||
|
||||
The entry objects are available from `FullPathDir` and `FileIndex` elements of the
|
||||
container object.
|
||||
The entry objects are available from `FullPathDir` and `FileIndex` elements of
|
||||
the container object.
|
||||
|
||||
- `.name` is the (case sensitive) internal name
|
||||
- `.type` is the type (`2 (stream)` for files, `1 (storage)` for dirs, `5 (root)` for root)
|
||||
- `.type` is the type as defined in "Object Type" in [MS-CFB] 2.6.1:
|
||||
`2 (stream)` for files, `1 (storage)` for dirs, `5 (root)` for root)
|
||||
- `.content` is a Buffer/Array with the raw content
|
||||
- `.ct`/`.mt` are the creation and modification time (if provided in file)
|
||||
|
||||
# Notes
|
||||
|
||||
Case comparison has not been verified for non-ASCII character
|
||||
Case comparison has not been verified for non-ASCII characters
|
||||
|
||||
Writing is not supported. It is in the works, but it has not yet been released.
|
||||
|
||||
The `xlscfb.js` file is designed to be embedded in [js-xls](http://git.io/xls)
|
||||
|
||||
# License
|
||||
|
||||
This implementation is covered under Apache 2.0 license. It complies with the
|
||||
|
|
|
|||
|
|
@ -5,7 +5,9 @@ var fs = require('fs'), program = require('commander');
|
|||
program
|
||||
.version(CFB.version)
|
||||
.usage('[options] <file>')
|
||||
.option('-q, --quiet', 'print but do not extract')
|
||||
.option('-q, --quiet', 'process but do not report')
|
||||
.option('-d, --dump', 'dump internal representation but do not extract')
|
||||
.option('--dev', 'development mode')
|
||||
.parse(process.argv);
|
||||
|
||||
if(program.args.length === 0 || !fs.existsSync(program.args[0])) {
|
||||
|
|
@ -13,15 +15,17 @@ if(program.args.length === 0 || !fs.existsSync(program.args[0])) {
|
|||
process.exit(1);
|
||||
}
|
||||
|
||||
var cfb = CFB.read(program.args[0], {type:'file'});
|
||||
if(program.quiet) {
|
||||
var opts = {type:'file'};
|
||||
if(program.dev) opts.WTF = true;
|
||||
|
||||
var cfb = CFB.read(program.args[0], opts);
|
||||
if(program.dump) {
|
||||
console.log("Full Paths:")
|
||||
console.log(cfb.FullPaths.map(function(x) { return " " + x; }).join("\n"));
|
||||
console.log("Full Path Directory:")
|
||||
console.log(cfb.FullPathDir);
|
||||
return;
|
||||
}
|
||||
for(var i=0; i != cfb.FullPaths.length; ++i) {
|
||||
if(!program.quiet && !program.dump) for(var i=0; i!=cfb.FullPaths.length; ++i) {
|
||||
if(cfb.FullPaths[i].slice(-1) === "/") {
|
||||
console.error("mkdir " + cfb.FullPaths[i]);
|
||||
fs.mkdirSync(cfb.FullPaths[i]);
|
||||
|
|
@ -1 +1 @@
|
|||
exports.version = '0.10.1';
|
||||
exports.version = '0.10.2';
|
||||
|
|
|
|||
|
|
@ -14,7 +14,8 @@ var blob = file.slice(0,512);
|
|||
prep_blob(blob, 0);
|
||||
|
||||
/* major version */
|
||||
mver = check_get_mver(blob);
|
||||
var mv = check_get_mver(blob);
|
||||
mver = mv[0];
|
||||
switch(mver) {
|
||||
case 3: ssz = 512; break; case 4: ssz = 4096; break;
|
||||
default: throw "Major Version: Expected 3 or 4 saw " + mver;
|
||||
|
|
@ -74,6 +75,8 @@ var sector_list = make_sector_list(sectors, dir_start, fat_addrs, ssz);
|
|||
sector_list[dir_start].name = "!Directory";
|
||||
if(nmfs > 0 && minifat_start !== ENDOFCHAIN) sector_list[minifat_start].name = "!MiniFAT";
|
||||
sector_list[fat_addrs[0]].name = "!FAT";
|
||||
sector_list.fat_addrs = fat_addrs;
|
||||
sector_list.ssz = ssz;
|
||||
|
||||
/* [MS-CFB] 2.6.1 Compound File Directory Entry */
|
||||
var files = {}, Paths = [], FileIndex = [], FullPaths = [], FullPathDir = {};
|
||||
|
|
|
|||
|
|
@ -7,9 +7,9 @@ function check_get_mver(blob) {
|
|||
blob.chk(HEADER_CLSID, 'CLSID: ');
|
||||
|
||||
// minor version 2
|
||||
blob.l += 2;
|
||||
var mver = blob.read_shift(2, 'u');
|
||||
|
||||
return blob.read_shift(2,'u');
|
||||
return [blob.read_shift(2,'u'), mver];
|
||||
}
|
||||
function check_shifts(blob, mver) {
|
||||
var shift = 0x09;
|
||||
|
|
|
|||
|
|
@ -2,13 +2,13 @@
|
|||
function make_find_path(FullPaths, Paths, FileIndex, files, root_name) {
|
||||
var UCFullPaths = new Array(FullPaths.length);
|
||||
var UCPaths = new Array(Paths.length), i;
|
||||
for(i = 0; i < FullPaths.length; ++i) UCFullPaths[i] = FullPaths[i].toUpperCase();
|
||||
for(i = 0; i < Paths.length; ++i) UCPaths[i] = Paths[i].toUpperCase();
|
||||
for(i = 0; i < FullPaths.length; ++i) UCFullPaths[i] = FullPaths[i].toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
for(i = 0; i < Paths.length; ++i) UCPaths[i] = Paths[i].toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
return function find_path(path) {
|
||||
var k;
|
||||
if(path.charCodeAt(0) === 47 /* "/" */) { k=true; path = root_name + path; }
|
||||
else k = path.indexOf("/") !== -1;
|
||||
var UCPath = path.toUpperCase();
|
||||
var UCPath = path.toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
var w = k === true ? UCFullPaths.indexOf(UCPath) : UCPaths.indexOf(UCPath);
|
||||
if(w === -1) return null;
|
||||
return k === true ? FileIndex[w] : files[Paths[w]];
|
||||
|
|
|
|||
|
|
@ -14,6 +14,26 @@ function sleuth_fat(idx, cnt, sectors, ssz, fat_addrs) {
|
|||
}
|
||||
}
|
||||
|
||||
/** Follow the linked list of sectors for a given starting point */
|
||||
function get_sector_list(sectors, start, fat_addrs, ssz, chkd) {
|
||||
var sl = sectors.length;
|
||||
var buf, buf_chain;
|
||||
if(!chkd) chkd = new Array(sl);
|
||||
var modulus = ssz - 1, j, jj;
|
||||
buf = [];
|
||||
buf_chain = [];
|
||||
for(j=start; j>=0;) {
|
||||
chkd[j] = true;
|
||||
buf[buf.length] = j;
|
||||
buf_chain.push(sectors[j]);
|
||||
var addr = fat_addrs[Math.floor(j*4/ssz)];
|
||||
jj = ((j*4) & modulus);
|
||||
if(ssz < 4 + jj) throw "FAT boundary crossed: " + j + " 4 "+ssz;
|
||||
j = __readInt32LE(sectors[addr], jj);
|
||||
}
|
||||
return {nodes: buf, data:__toBuffer([buf_chain])};
|
||||
}
|
||||
|
||||
/** Chase down the sector linked lists */
|
||||
function make_sector_list(sectors, dir_start, fat_addrs, ssz) {
|
||||
var sl = sectors.length, sector_list = new Array(sl);
|
||||
|
|
|
|||
|
|
@ -37,7 +37,7 @@ function read_directory(dir_start, sector_list, sectors, Paths, nmfs, files, Fil
|
|||
/*minifat_size = o.size;*/
|
||||
} else if(o.size >= 4096 /* MSCSZ */) {
|
||||
o.storage = 'fat';
|
||||
if(sector_list[o.start] === undefined) if((o.start+=dir_start)>=sectors.length) o.start-=sectors.length;
|
||||
if(sector_list[o.start] === undefined) sector_list[o.start] = get_sector_list(sectors, o.start, sector_list.fat_addrs, sector_list.ssz);
|
||||
sector_list[o.start].name = o.name;
|
||||
o.content = sector_list[o.start].data.slice(0,o.size);
|
||||
prep_blob(o.content, 0);
|
||||
|
|
|
|||
|
|
@ -1,14 +1,14 @@
|
|||
var fs;
|
||||
function readFileSync(filename) {
|
||||
function readFileSync(filename, options) {
|
||||
if(fs === undefined) fs = require('fs');
|
||||
return parse(fs.readFileSync(filename));
|
||||
return parse(fs.readFileSync(filename), options);
|
||||
}
|
||||
|
||||
function readSync(blob, options) {
|
||||
switch(options !== undefined && options.type !== undefined ? options.type : "base64") {
|
||||
case "file": return readFileSync(blob);
|
||||
case "base64": return parse(s2a(Base64.decode(blob)));
|
||||
case "binary": return parse(s2a(blob));
|
||||
case "file": return readFileSync(blob, options);
|
||||
case "base64": return parse(s2a(Base64.decode(blob)), options);
|
||||
case "binary": return parse(s2a(blob), options);
|
||||
}
|
||||
return parse(blob);
|
||||
}
|
||||
|
|
|
|||
49
cfb.js
49
cfb.js
|
|
@ -86,7 +86,7 @@ function prep_blob(blob, pos) {
|
|||
/* [MS-CFB] v20130118 */
|
||||
var CFB = (function _CFB(){
|
||||
var exports = {};
|
||||
exports.version = '0.10.1';
|
||||
exports.version = '0.10.2';
|
||||
function parse(file) {
|
||||
var mver = 3; // major version
|
||||
var ssz = 512; // sector size
|
||||
|
|
@ -103,7 +103,8 @@ var blob = file.slice(0,512);
|
|||
prep_blob(blob, 0);
|
||||
|
||||
/* major version */
|
||||
mver = check_get_mver(blob);
|
||||
var mv = check_get_mver(blob);
|
||||
mver = mv[0];
|
||||
switch(mver) {
|
||||
case 3: ssz = 512; break; case 4: ssz = 4096; break;
|
||||
default: throw "Major Version: Expected 3 or 4 saw " + mver;
|
||||
|
|
@ -163,6 +164,8 @@ var sector_list = make_sector_list(sectors, dir_start, fat_addrs, ssz);
|
|||
sector_list[dir_start].name = "!Directory";
|
||||
if(nmfs > 0 && minifat_start !== ENDOFCHAIN) sector_list[minifat_start].name = "!MiniFAT";
|
||||
sector_list[fat_addrs[0]].name = "!FAT";
|
||||
sector_list.fat_addrs = fat_addrs;
|
||||
sector_list.ssz = ssz;
|
||||
|
||||
/* [MS-CFB] 2.6.1 Compound File Directory Entry */
|
||||
var files = {}, Paths = [], FileIndex = [], FullPaths = [], FullPathDir = {};
|
||||
|
|
@ -194,9 +197,9 @@ function check_get_mver(blob) {
|
|||
blob.chk(HEADER_CLSID, 'CLSID: ');
|
||||
|
||||
// minor version 2
|
||||
blob.l += 2;
|
||||
var mver = blob.read_shift(2, 'u');
|
||||
|
||||
return blob.read_shift(2,'u');
|
||||
return [blob.read_shift(2,'u'), mver];
|
||||
}
|
||||
function check_shifts(blob, mver) {
|
||||
var shift = 0x09;
|
||||
|
|
@ -272,13 +275,13 @@ function build_full_paths(FI, FPD, FP, Paths) {
|
|||
function make_find_path(FullPaths, Paths, FileIndex, files, root_name) {
|
||||
var UCFullPaths = new Array(FullPaths.length);
|
||||
var UCPaths = new Array(Paths.length), i;
|
||||
for(i = 0; i < FullPaths.length; ++i) UCFullPaths[i] = FullPaths[i].toUpperCase();
|
||||
for(i = 0; i < Paths.length; ++i) UCPaths[i] = Paths[i].toUpperCase();
|
||||
for(i = 0; i < FullPaths.length; ++i) UCFullPaths[i] = FullPaths[i].toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
for(i = 0; i < Paths.length; ++i) UCPaths[i] = Paths[i].toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
return function find_path(path) {
|
||||
var k;
|
||||
if(path.charCodeAt(0) === 47 /* "/" */) { k=true; path = root_name + path; }
|
||||
else k = path.indexOf("/") !== -1;
|
||||
var UCPath = path.toUpperCase();
|
||||
var UCPath = path.toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
var w = k === true ? UCFullPaths.indexOf(UCPath) : UCPaths.indexOf(UCPath);
|
||||
if(w === -1) return null;
|
||||
return k === true ? FileIndex[w] : files[Paths[w]];
|
||||
|
|
@ -301,6 +304,26 @@ function sleuth_fat(idx, cnt, sectors, ssz, fat_addrs) {
|
|||
}
|
||||
}
|
||||
|
||||
/** Follow the linked list of sectors for a given starting point */
|
||||
function get_sector_list(sectors, start, fat_addrs, ssz, chkd) {
|
||||
var sl = sectors.length;
|
||||
var buf, buf_chain;
|
||||
if(!chkd) chkd = new Array(sl);
|
||||
var modulus = ssz - 1, j, jj;
|
||||
buf = [];
|
||||
buf_chain = [];
|
||||
for(j=start; j>=0;) {
|
||||
chkd[j] = true;
|
||||
buf[buf.length] = j;
|
||||
buf_chain.push(sectors[j]);
|
||||
var addr = fat_addrs[Math.floor(j*4/ssz)];
|
||||
jj = ((j*4) & modulus);
|
||||
if(ssz < 4 + jj) throw "FAT boundary crossed: " + j + " 4 "+ssz;
|
||||
j = __readInt32LE(sectors[addr], jj);
|
||||
}
|
||||
return {nodes: buf, data:__toBuffer([buf_chain])};
|
||||
}
|
||||
|
||||
/** Chase down the sector linked lists */
|
||||
function make_sector_list(sectors, dir_start, fat_addrs, ssz) {
|
||||
var sl = sectors.length, sector_list = new Array(sl);
|
||||
|
|
@ -364,7 +387,7 @@ function read_directory(dir_start, sector_list, sectors, Paths, nmfs, files, Fil
|
|||
/*minifat_size = o.size;*/
|
||||
} else if(o.size >= 4096 /* MSCSZ */) {
|
||||
o.storage = 'fat';
|
||||
if(sector_list[o.start] === undefined) if((o.start+=dir_start)>=sectors.length) o.start-=sectors.length;
|
||||
if(sector_list[o.start] === undefined) sector_list[o.start] = get_sector_list(sectors, o.start, sector_list.fat_addrs, sector_list.ssz);
|
||||
sector_list[o.start].name = o.name;
|
||||
o.content = sector_list[o.start].data.slice(0,o.size);
|
||||
prep_blob(o.content, 0);
|
||||
|
|
@ -385,16 +408,16 @@ function read_date(blob, offset) {
|
|||
}
|
||||
|
||||
var fs;
|
||||
function readFileSync(filename) {
|
||||
function readFileSync(filename, options) {
|
||||
if(fs === undefined) fs = require('fs');
|
||||
return parse(fs.readFileSync(filename));
|
||||
return parse(fs.readFileSync(filename), options);
|
||||
}
|
||||
|
||||
function readSync(blob, options) {
|
||||
switch(options !== undefined && options.type !== undefined ? options.type : "base64") {
|
||||
case "file": return readFileSync(blob);
|
||||
case "base64": return parse(s2a(Base64.decode(blob)));
|
||||
case "binary": return parse(s2a(blob));
|
||||
case "file": return readFileSync(blob, options);
|
||||
case "base64": return parse(s2a(Base64.decode(blob)), options);
|
||||
case "binary": return parse(s2a(blob), options);
|
||||
}
|
||||
return parse(blob);
|
||||
}
|
||||
|
|
|
|||
|
|
@ -86,7 +86,7 @@ function prep_blob(blob, pos) {
|
|||
/* [MS-CFB] v20130118 */
|
||||
var CFB = (function _CFB(){
|
||||
var exports = {};
|
||||
exports.version = '0.10.1';
|
||||
exports.version = '0.10.2';
|
||||
function parse(file) {
|
||||
var mver = 3; // major version
|
||||
var ssz = 512; // sector size
|
||||
|
|
@ -103,7 +103,8 @@ var blob = file.slice(0,512);
|
|||
prep_blob(blob, 0);
|
||||
|
||||
/* major version */
|
||||
mver = check_get_mver(blob);
|
||||
var mv = check_get_mver(blob);
|
||||
mver = mv[0];
|
||||
switch(mver) {
|
||||
case 3: ssz = 512; break; case 4: ssz = 4096; break;
|
||||
default: throw "Major Version: Expected 3 or 4 saw " + mver;
|
||||
|
|
@ -163,6 +164,8 @@ var sector_list = make_sector_list(sectors, dir_start, fat_addrs, ssz);
|
|||
sector_list[dir_start].name = "!Directory";
|
||||
if(nmfs > 0 && minifat_start !== ENDOFCHAIN) sector_list[minifat_start].name = "!MiniFAT";
|
||||
sector_list[fat_addrs[0]].name = "!FAT";
|
||||
sector_list.fat_addrs = fat_addrs;
|
||||
sector_list.ssz = ssz;
|
||||
|
||||
/* [MS-CFB] 2.6.1 Compound File Directory Entry */
|
||||
var files = {}, Paths = [], FileIndex = [], FullPaths = [], FullPathDir = {};
|
||||
|
|
@ -194,9 +197,9 @@ function check_get_mver(blob) {
|
|||
blob.chk(HEADER_CLSID, 'CLSID: ');
|
||||
|
||||
// minor version 2
|
||||
blob.l += 2;
|
||||
var mver = blob.read_shift(2, 'u');
|
||||
|
||||
return blob.read_shift(2,'u');
|
||||
return [blob.read_shift(2,'u'), mver];
|
||||
}
|
||||
function check_shifts(blob, mver) {
|
||||
var shift = 0x09;
|
||||
|
|
@ -272,13 +275,13 @@ function build_full_paths(FI, FPD, FP, Paths) {
|
|||
function make_find_path(FullPaths, Paths, FileIndex, files, root_name) {
|
||||
var UCFullPaths = new Array(FullPaths.length);
|
||||
var UCPaths = new Array(Paths.length), i;
|
||||
for(i = 0; i < FullPaths.length; ++i) UCFullPaths[i] = FullPaths[i].toUpperCase();
|
||||
for(i = 0; i < Paths.length; ++i) UCPaths[i] = Paths[i].toUpperCase();
|
||||
for(i = 0; i < FullPaths.length; ++i) UCFullPaths[i] = FullPaths[i].toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
for(i = 0; i < Paths.length; ++i) UCPaths[i] = Paths[i].toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
return function find_path(path) {
|
||||
var k;
|
||||
if(path.charCodeAt(0) === 47 /* "/" */) { k=true; path = root_name + path; }
|
||||
else k = path.indexOf("/") !== -1;
|
||||
var UCPath = path.toUpperCase();
|
||||
var UCPath = path.toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
var w = k === true ? UCFullPaths.indexOf(UCPath) : UCPaths.indexOf(UCPath);
|
||||
if(w === -1) return null;
|
||||
return k === true ? FileIndex[w] : files[Paths[w]];
|
||||
|
|
@ -301,6 +304,26 @@ function sleuth_fat(idx, cnt, sectors, ssz, fat_addrs) {
|
|||
}
|
||||
}
|
||||
|
||||
/** Follow the linked list of sectors for a given starting point */
|
||||
function get_sector_list(sectors, start, fat_addrs, ssz, chkd) {
|
||||
var sl = sectors.length;
|
||||
var buf, buf_chain;
|
||||
if(!chkd) chkd = new Array(sl);
|
||||
var modulus = ssz - 1, j, jj;
|
||||
buf = [];
|
||||
buf_chain = [];
|
||||
for(j=start; j>=0;) {
|
||||
chkd[j] = true;
|
||||
buf[buf.length] = j;
|
||||
buf_chain.push(sectors[j]);
|
||||
var addr = fat_addrs[Math.floor(j*4/ssz)];
|
||||
jj = ((j*4) & modulus);
|
||||
if(ssz < 4 + jj) throw "FAT boundary crossed: " + j + " 4 "+ssz;
|
||||
j = __readInt32LE(sectors[addr], jj);
|
||||
}
|
||||
return {nodes: buf, data:__toBuffer([buf_chain])};
|
||||
}
|
||||
|
||||
/** Chase down the sector linked lists */
|
||||
function make_sector_list(sectors, dir_start, fat_addrs, ssz) {
|
||||
var sl = sectors.length, sector_list = new Array(sl);
|
||||
|
|
@ -364,7 +387,7 @@ function read_directory(dir_start, sector_list, sectors, Paths, nmfs, files, Fil
|
|||
/*minifat_size = o.size;*/
|
||||
} else if(o.size >= 4096 /* MSCSZ */) {
|
||||
o.storage = 'fat';
|
||||
if(sector_list[o.start] === undefined) if((o.start+=dir_start)>=sectors.length) o.start-=sectors.length;
|
||||
if(sector_list[o.start] === undefined) sector_list[o.start] = get_sector_list(sectors, o.start, sector_list.fat_addrs, sector_list.ssz);
|
||||
sector_list[o.start].name = o.name;
|
||||
o.content = sector_list[o.start].data.slice(0,o.size);
|
||||
prep_blob(o.content, 0);
|
||||
|
|
@ -385,16 +408,16 @@ function read_date(blob, offset) {
|
|||
}
|
||||
|
||||
var fs;
|
||||
function readFileSync(filename) {
|
||||
function readFileSync(filename, options) {
|
||||
if(fs === undefined) fs = require('fs');
|
||||
return parse(fs.readFileSync(filename));
|
||||
return parse(fs.readFileSync(filename), options);
|
||||
}
|
||||
|
||||
function readSync(blob, options) {
|
||||
switch(options !== undefined && options.type !== undefined ? options.type : "base64") {
|
||||
case "file": return readFileSync(blob);
|
||||
case "base64": return parse(s2a(Base64.decode(blob)));
|
||||
case "binary": return parse(s2a(blob));
|
||||
case "file": return readFileSync(blob, options);
|
||||
case "base64": return parse(s2a(Base64.decode(blob)), options);
|
||||
case "binary": return parse(s2a(blob), options);
|
||||
}
|
||||
return parse(blob);
|
||||
}
|
||||
|
|
|
|||
File diff suppressed because one or more lines are too long
File diff suppressed because one or more lines are too long
|
|
@ -0,0 +1,383 @@
|
|||
var DO_NOT_EXPORT_CFB = true;
|
||||
/* cfb.js (C) 2013-2014 SheetJS -- http://sheetjs.com */
|
||||
/* vim: set ts=2: */
|
||||
/*jshint eqnull:true */
|
||||
|
||||
/* [MS-CFB] v20130118 */
|
||||
var CFB = (function _CFB(){
|
||||
var exports = {};
|
||||
exports.version = '0.10.2';
|
||||
function parse(file) {
|
||||
var mver = 3; // major version
|
||||
var ssz = 512; // sector size
|
||||
var nmfs = 0; // number of mini FAT sectors
|
||||
var ndfs = 0; // number of DIFAT sectors
|
||||
var dir_start = 0; // first directory sector location
|
||||
var minifat_start = 0; // first mini FAT sector location
|
||||
var difat_start = 0; // first mini FAT sector location
|
||||
|
||||
var fat_addrs = []; // locations of FAT sectors
|
||||
|
||||
/* [MS-CFB] 2.2 Compound File Header */
|
||||
var blob = file.slice(0,512);
|
||||
prep_blob(blob, 0);
|
||||
|
||||
/* major version */
|
||||
var mv = check_get_mver(blob);
|
||||
mver = mv[0];
|
||||
switch(mver) {
|
||||
case 3: ssz = 512; break; case 4: ssz = 4096; break;
|
||||
default: throw "Major Version: Expected 3 or 4 saw " + mver;
|
||||
}
|
||||
|
||||
/* reprocess header */
|
||||
if(ssz !== 512) { blob = file.slice(0,ssz); prep_blob(blob, 28 /* blob.l */); }
|
||||
/* Save header for final object */
|
||||
var header = file.slice(0,ssz);
|
||||
|
||||
check_shifts(blob, mver);
|
||||
|
||||
// Number of Directory Sectors
|
||||
var nds = blob.read_shift(4, 'i');
|
||||
if(mver === 3 && nds !== 0) throw '# Directory Sectors: Expected 0 saw ' + nds;
|
||||
|
||||
// Number of FAT Sectors
|
||||
//var nfs = blob.read_shift(4, 'i');
|
||||
blob.l += 4;
|
||||
|
||||
// First Directory Sector Location
|
||||
dir_start = blob.read_shift(4, 'i');
|
||||
|
||||
// Transaction Signature
|
||||
blob.l += 4;
|
||||
|
||||
// Mini Stream Cutoff Size
|
||||
blob.chk('00100000', 'Mini Stream Cutoff Size: ');
|
||||
|
||||
// First Mini FAT Sector Location
|
||||
minifat_start = blob.read_shift(4, 'i');
|
||||
|
||||
// Number of Mini FAT Sectors
|
||||
nmfs = blob.read_shift(4, 'i');
|
||||
|
||||
// First DIFAT sector location
|
||||
difat_start = blob.read_shift(4, 'i');
|
||||
|
||||
// Number of DIFAT Sectors
|
||||
ndfs = blob.read_shift(4, 'i');
|
||||
|
||||
// Grab FAT Sector Locations
|
||||
for(var q, j = 0; j < 109; ++j) { /* 109 = (512 - blob.l)>>>2; */
|
||||
q = blob.read_shift(4, 'i');
|
||||
if(q<0) break;
|
||||
fat_addrs[j] = q;
|
||||
}
|
||||
|
||||
/** Break the file up into sectors */
|
||||
var sectors = sectorify(file, ssz);
|
||||
|
||||
sleuth_fat(difat_start, ndfs, sectors, ssz, fat_addrs);
|
||||
|
||||
/** Chains */
|
||||
var sector_list = make_sector_list(sectors, dir_start, fat_addrs, ssz);
|
||||
|
||||
sector_list[dir_start].name = "!Directory";
|
||||
if(nmfs > 0 && minifat_start !== ENDOFCHAIN) sector_list[minifat_start].name = "!MiniFAT";
|
||||
sector_list[fat_addrs[0]].name = "!FAT";
|
||||
sector_list.fat_addrs = fat_addrs;
|
||||
sector_list.ssz = ssz;
|
||||
|
||||
/* [MS-CFB] 2.6.1 Compound File Directory Entry */
|
||||
var files = {}, Paths = [], FileIndex = [], FullPaths = [], FullPathDir = {};
|
||||
read_directory(dir_start, sector_list, sectors, Paths, nmfs, files, FileIndex);
|
||||
|
||||
build_full_paths(FileIndex, FullPathDir, FullPaths, Paths);
|
||||
|
||||
var root_name = Paths.shift();
|
||||
Paths.root = root_name;
|
||||
|
||||
/* [MS-CFB] 2.6.4 (Unicode 3.0.1 case conversion) */
|
||||
var find_path = make_find_path(FullPaths, Paths, FileIndex, files, root_name);
|
||||
|
||||
return {
|
||||
raw: {header: header, sectors: sectors},
|
||||
FileIndex: FileIndex,
|
||||
FullPaths: FullPaths,
|
||||
FullPathDir: FullPathDir,
|
||||
find: find_path
|
||||
};
|
||||
} // parse
|
||||
|
||||
/* [MS-CFB] 2.2 Compound File Header -- read up to major version */
|
||||
function check_get_mver(blob) {
|
||||
// header signature 8
|
||||
blob.chk(HEADER_SIGNATURE, 'Header Signature: ');
|
||||
|
||||
// clsid 16
|
||||
blob.chk(HEADER_CLSID, 'CLSID: ');
|
||||
|
||||
// minor version 2
|
||||
var mver = blob.read_shift(2, 'u');
|
||||
|
||||
return [blob.read_shift(2,'u'), mver];
|
||||
}
|
||||
function check_shifts(blob, mver) {
|
||||
var shift = 0x09;
|
||||
|
||||
// Byte Order
|
||||
blob.chk('feff', 'Byte Order: ');
|
||||
|
||||
// Sector Shift
|
||||
switch((shift = blob.read_shift(2))) {
|
||||
case 0x09: if(mver !== 3) throw 'MajorVersion/SectorShift Mismatch'; break;
|
||||
case 0x0c: if(mver !== 4) throw 'MajorVersion/SectorShift Mismatch'; break;
|
||||
default: throw 'Sector Shift: Expected 9 or 12 saw ' + shift;
|
||||
}
|
||||
|
||||
// Mini Sector Shift
|
||||
blob.chk('0600', 'Mini Sector Shift: ');
|
||||
|
||||
// Reserved
|
||||
blob.chk('000000000000', 'Reserved: ');
|
||||
}
|
||||
|
||||
/** Break the file up into sectors */
|
||||
function sectorify(file, ssz) {
|
||||
var nsectors = Math.ceil(file.length/ssz)-1;
|
||||
var sectors = new Array(nsectors);
|
||||
for(var i=1; i < nsectors; ++i) sectors[i-1] = file.slice(i*ssz,(i+1)*ssz);
|
||||
sectors[nsectors-1] = file.slice(nsectors*ssz);
|
||||
return sectors;
|
||||
}
|
||||
|
||||
/* [MS-CFB] 2.6.4 Red-Black Tree */
|
||||
function build_full_paths(FI, FPD, FP, Paths) {
|
||||
var i = 0, L = 0, R = 0, C = 0, j = 0, pl = Paths.length;
|
||||
var dad = new Array(pl), q = new Array(pl);
|
||||
|
||||
for(; i < pl; ++i) { dad[i]=q[i]=i; FP[i]=Paths[i]; }
|
||||
|
||||
for(; j < q.length; ++j) {
|
||||
i = q[j];
|
||||
L = FI[i].L; R = FI[i].R; C = FI[i].C;
|
||||
if(dad[i] === i) {
|
||||
if(L !== -1 /*NOSTREAM*/ && dad[L] !== L) dad[i] = dad[L];
|
||||
if(R !== -1 && dad[R] !== R) dad[i] = dad[R];
|
||||
}
|
||||
if(C !== -1 /*NOSTREAM*/) dad[C] = i;
|
||||
if(L !== -1) { dad[L] = dad[i]; q.push(L); }
|
||||
if(R !== -1) { dad[R] = dad[i]; q.push(R); }
|
||||
}
|
||||
for(i=1; i !== pl; ++i) if(dad[i] === i) {
|
||||
if(R !== -1 /*NOSTREAM*/ && dad[R] !== R) dad[i] = dad[R];
|
||||
else if(L !== -1 && dad[L] !== L) dad[i] = dad[L];
|
||||
}
|
||||
|
||||
for(i=1; i < pl; ++i) {
|
||||
if(FI[i].type === 0 /* unknown */) continue;
|
||||
j = dad[i];
|
||||
if(j === 0) FP[i] = FP[0] + "/" + FP[i];
|
||||
else while(j !== 0) {
|
||||
FP[i] = FP[j] + "/" + FP[i];
|
||||
j = dad[j];
|
||||
}
|
||||
dad[i] = 0;
|
||||
}
|
||||
|
||||
FP[0] += "/";
|
||||
for(i=1; i < pl; ++i) {
|
||||
if(FI[i].type !== 2 /* stream */) FP[i] += "/";
|
||||
FPD[FP[i]] = FI[i];
|
||||
}
|
||||
}
|
||||
|
||||
/* [MS-CFB] 2.6.4 */
|
||||
function make_find_path(FullPaths, Paths, FileIndex, files, root_name) {
|
||||
var UCFullPaths = new Array(FullPaths.length);
|
||||
var UCPaths = new Array(Paths.length), i;
|
||||
for(i = 0; i < FullPaths.length; ++i) UCFullPaths[i] = FullPaths[i].toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
for(i = 0; i < Paths.length; ++i) UCPaths[i] = Paths[i].toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
return function find_path(path) {
|
||||
var k;
|
||||
if(path.charCodeAt(0) === 47 /* "/" */) { k=true; path = root_name + path; }
|
||||
else k = path.indexOf("/") !== -1;
|
||||
var UCPath = path.toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
var w = k === true ? UCFullPaths.indexOf(UCPath) : UCPaths.indexOf(UCPath);
|
||||
if(w === -1) return null;
|
||||
return k === true ? FileIndex[w] : files[Paths[w]];
|
||||
};
|
||||
}
|
||||
|
||||
/** Chase down the rest of the DIFAT chain to build a comprehensive list
|
||||
DIFAT chains by storing the next sector number as the last 32 bytes */
|
||||
function sleuth_fat(idx, cnt, sectors, ssz, fat_addrs) {
|
||||
var q;
|
||||
if(idx === ENDOFCHAIN) {
|
||||
if(cnt !== 0) throw "DIFAT chain shorter than expected";
|
||||
} else if(idx !== -1 /*FREESECT*/) {
|
||||
var sector = sectors[idx], m = (ssz>>>2)-1;
|
||||
for(var i = 0; i < m; ++i) {
|
||||
if((q = __readInt32LE(sector,i*4)) === ENDOFCHAIN) break;
|
||||
fat_addrs.push(q);
|
||||
}
|
||||
sleuth_fat(__readInt32LE(sector,ssz-4),cnt - 1, sectors, ssz, fat_addrs);
|
||||
}
|
||||
}
|
||||
|
||||
/** Follow the linked list of sectors for a given starting point */
|
||||
function get_sector_list(sectors, start, fat_addrs, ssz, chkd) {
|
||||
var sl = sectors.length;
|
||||
var buf, buf_chain;
|
||||
if(!chkd) chkd = new Array(sl);
|
||||
var modulus = ssz - 1, j, jj;
|
||||
buf = [];
|
||||
buf_chain = [];
|
||||
for(j=start; j>=0;) {
|
||||
chkd[j] = true;
|
||||
buf[buf.length] = j;
|
||||
buf_chain.push(sectors[j]);
|
||||
var addr = fat_addrs[Math.floor(j*4/ssz)];
|
||||
jj = ((j*4) & modulus);
|
||||
if(ssz < 4 + jj) throw "FAT boundary crossed: " + j + " 4 "+ssz;
|
||||
j = __readInt32LE(sectors[addr], jj);
|
||||
}
|
||||
return {nodes: buf, data:__toBuffer([buf_chain])};
|
||||
}
|
||||
|
||||
/** Chase down the sector linked lists */
|
||||
function make_sector_list(sectors, dir_start, fat_addrs, ssz) {
|
||||
var sl = sectors.length, sector_list = new Array(sl);
|
||||
var chkd = new Array(sl), buf, buf_chain;
|
||||
var modulus = ssz - 1, i, j, k, jj;
|
||||
for(i=0; i < sl; ++i) {
|
||||
buf = [];
|
||||
k = (i + dir_start); if(k >= sl) k-=sl;
|
||||
if(chkd[k] === true) continue;
|
||||
buf_chain = [];
|
||||
for(j=k; j>=0;) {
|
||||
chkd[j] = true;
|
||||
buf[buf.length] = j;
|
||||
buf_chain.push(sectors[j]);
|
||||
var addr = fat_addrs[Math.floor(j*4/ssz)];
|
||||
jj = ((j*4) & modulus);
|
||||
if(ssz < 4 + jj) throw "FAT boundary crossed: " + j + " 4 "+ssz;
|
||||
j = __readInt32LE(sectors[addr], jj);
|
||||
}
|
||||
sector_list[k] = {nodes: buf, data:__toBuffer([buf_chain])};
|
||||
}
|
||||
return sector_list;
|
||||
}
|
||||
|
||||
/* [MS-CFB] 2.6.1 Compound File Directory Entry */
|
||||
function read_directory(dir_start, sector_list, sectors, Paths, nmfs, files, FileIndex) {
|
||||
var blob;
|
||||
var minifat_store = 0, pl = (Paths.length?2:0);
|
||||
var sector = sector_list[dir_start].data;
|
||||
var i = 0, namelen = 0, name, o, ctime, mtime;
|
||||
for(; i < sector.length; i+= 128) {
|
||||
blob = sector.slice(i, i+128);
|
||||
prep_blob(blob, 64);
|
||||
namelen = blob.read_shift(2);
|
||||
if(namelen === 0) continue;
|
||||
name = __utf16le(blob,0,namelen-pl);
|
||||
Paths.push(name);
|
||||
o = {
|
||||
name: name,
|
||||
type: blob.read_shift(1),
|
||||
color: blob.read_shift(1),
|
||||
L: blob.read_shift(4, 'i'),
|
||||
R: blob.read_shift(4, 'i'),
|
||||
C: blob.read_shift(4, 'i'),
|
||||
clsid: blob.read_shift(16),
|
||||
state: blob.read_shift(4, 'i')
|
||||
};
|
||||
ctime = blob.read_shift(2) + blob.read_shift(2) + blob.read_shift(2) + blob.read_shift(2);
|
||||
if(ctime !== 0) {
|
||||
o.ctime = ctime; o.ct = read_date(blob, blob.l-8);
|
||||
}
|
||||
mtime = blob.read_shift(2) + blob.read_shift(2) + blob.read_shift(2) + blob.read_shift(2);
|
||||
if(mtime !== 0) {
|
||||
o.mtime = mtime; o.mt = read_date(blob, blob.l-8);
|
||||
}
|
||||
o.start = blob.read_shift(4, 'i');
|
||||
o.size = blob.read_shift(4, 'i');
|
||||
if(o.type === 5) { /* root */
|
||||
minifat_store = o.start;
|
||||
if(nmfs > 0 && minifat_store !== ENDOFCHAIN) sector_list[minifat_store].name = "!StreamData";
|
||||
/*minifat_size = o.size;*/
|
||||
} else if(o.size >= 4096 /* MSCSZ */) {
|
||||
o.storage = 'fat';
|
||||
if(sector_list[o.start] === undefined) sector_list[o.start] = get_sector_list(sectors, o.start, sector_list.fat_addrs, sector_list.ssz);
|
||||
sector_list[o.start].name = o.name;
|
||||
o.content = sector_list[o.start].data.slice(0,o.size);
|
||||
prep_blob(o.content, 0);
|
||||
} else {
|
||||
o.storage = 'minifat';
|
||||
if(minifat_store !== ENDOFCHAIN && o.start !== ENDOFCHAIN) {
|
||||
o.content = sector_list[minifat_store].data.slice(o.start*MSSZ,o.start*MSSZ+o.size);
|
||||
prep_blob(o.content, 0);
|
||||
}
|
||||
}
|
||||
files[name] = o;
|
||||
FileIndex.push(o);
|
||||
}
|
||||
}
|
||||
|
||||
function read_date(blob, offset) {
|
||||
return new Date(( ( (__readUInt32LE(blob,offset+4)/1e7)*Math.pow(2,32)+__readUInt32LE(blob,offset)/1e7 ) - 11644473600)*1000);
|
||||
}
|
||||
|
||||
var fs;
|
||||
function readFileSync(filename, options) {
|
||||
if(fs === undefined) fs = require('fs');
|
||||
return parse(fs.readFileSync(filename), options);
|
||||
}
|
||||
|
||||
function readSync(blob, options) {
|
||||
switch(options !== undefined && options.type !== undefined ? options.type : "base64") {
|
||||
case "file": return readFileSync(blob, options);
|
||||
case "base64": return parse(s2a(Base64.decode(blob)), options);
|
||||
case "binary": return parse(s2a(blob), options);
|
||||
}
|
||||
return parse(blob);
|
||||
}
|
||||
|
||||
/** CFB Constants */
|
||||
var MSSZ = 64; /* Mini Sector Size = 1<<6 */
|
||||
//var MSCSZ = 4096; /* Mini Stream Cutoff Size */
|
||||
/* 2.1 Compound File Sector Numbers and Types */
|
||||
var ENDOFCHAIN = -2;
|
||||
/* 2.2 Compound File Header */
|
||||
var HEADER_SIGNATURE = 'd0cf11e0a1b11ae1';
|
||||
var HEADER_CLSID = '00000000000000000000000000000000';
|
||||
var consts = {
|
||||
/* 2.1 Compund File Sector Numbers and Types */
|
||||
MAXREGSECT: -6,
|
||||
DIFSECT: -4,
|
||||
FATSECT: -3,
|
||||
ENDOFCHAIN: ENDOFCHAIN,
|
||||
FREESECT: -1,
|
||||
/* 2.2 Compound File Header */
|
||||
HEADER_SIGNATURE: HEADER_SIGNATURE,
|
||||
HEADER_MINOR_VERSION: '3e00',
|
||||
MAXREGSID: -6,
|
||||
NOSTREAM: -1,
|
||||
HEADER_CLSID: HEADER_CLSID,
|
||||
/* 2.6.1 Compound File Directory Entry */
|
||||
EntryTypes: ['unknown','storage','stream','lockbytes','property','root']
|
||||
};
|
||||
|
||||
exports.read = readSync;
|
||||
exports.parse = parse;
|
||||
exports.utils = {
|
||||
ReadShift: ReadShift,
|
||||
CheckField: CheckField,
|
||||
prep_blob: prep_blob,
|
||||
bconcat: bconcat,
|
||||
consts: consts
|
||||
};
|
||||
|
||||
return exports;
|
||||
})();
|
||||
|
||||
if(typeof require !== 'undefined' && typeof module !== 'undefined' && typeof DO_NOT_EXPORT_CFB === 'undefined') { module.exports = CFB; }
|
||||
|
|
@ -0,0 +1,11 @@
|
|||
#!/bin/bash
|
||||
# strip_sourcemap.sh -- strip sourcemaps from a JS file (missing from uglifyjs)
|
||||
# Copyright (C) 2014 SheetJS
|
||||
|
||||
if [ $# -gt 0 ]; then
|
||||
if [ -e "$1" ]; then
|
||||
sed -i .sheetjs '/sourceMappingURL/d' "$1"
|
||||
fi
|
||||
else
|
||||
cat - | sed '/sourceMappingURL/d'
|
||||
fi
|
||||
|
|
@ -0,0 +1 @@
|
|||
var DO_NOT_EXPORT_CFB = true;
|
||||
|
|
@ -1,6 +1,6 @@
|
|||
{
|
||||
"name": "cfb",
|
||||
"version": "0.10.1",
|
||||
"version": "0.10.2",
|
||||
"author": "sheetjs",
|
||||
"description": "Compound File Binary File Format extractor",
|
||||
"keywords": [ "cfb", "compression", "office" ],
|
||||
|
|
|
|||
26
test.js
26
test.js
|
|
@ -3,21 +3,35 @@ var CFB;
|
|||
var fs = require('fs');
|
||||
describe('source', function() { it('should load', function() { CFB = require('./'); }); });
|
||||
|
||||
var ffunc = function(x){return x.substr(-4)==".xls" && fails.indexOf(x) === -1;};
|
||||
var ex = [".xls",".doc",".ppt"];
|
||||
if(process.env.FMTS) ex=process.env.FMTS.split(":").map(function(x){return x[0]==="."?x:"."+x;});
|
||||
if(process.env.FMTS === "full") process.env.FMTS = ex.join(":");
|
||||
if(process.env.FMTS) ex=process.env.FMTS.split(":").map(function(x){return x[0]==="."?x:"."+x;});
|
||||
|
||||
var ffunc = function(x){return (ex.indexOf(x.substr(-4))>=0 || ex.indexOf(x.substr(-3))>=0) && fails.indexOf(x) === -1;};
|
||||
var fails = fs.existsSync('./fails.lst') ? fs.readFileSync('./fails.lst', 'utf-8').split("\n") : [];
|
||||
var files = fs.readdirSync('test_files').filter(ffunc);
|
||||
var f2011 = fs.readdirSync('test_files/2011').filter(ffunc);
|
||||
var f2013 = fs.readdirSync('test_files/2013').filter(ffunc);
|
||||
var fpres = fs.readdirSync('test_files_pres').filter(ffunc);
|
||||
|
||||
var dir = "./test_files/";
|
||||
|
||||
function parsetest(x, cfb) {
|
||||
describe(x + ' should have basic parts', function() {
|
||||
it('should find relative path', function() {
|
||||
if(!cfb.find('Workbook') && !cfb.find('Book')) throw new Error("Cannot find workbook for " + x);
|
||||
switch(x.substr(-4)) {
|
||||
case '.xls': if(!cfb.find('Workbook') && !cfb.find('Book')) throw new Error("Cannot find workbook for " + x); break;
|
||||
case '.ppt': if(!cfb.find('PowerPoint Document')) throw new Error("Cannot find presentation for " + x); break;
|
||||
case '.doc': if(!cfb.find('WordDocument') && !cfb.find('Word Document')) throw new Error("Cannot find doc for " + x); break;
|
||||
}
|
||||
});
|
||||
it('should find absolute path', function() {
|
||||
if(!cfb.find('/Workbook') && !cfb.find('/Book')) throw new Error("Cannot find workbook for " + x);
|
||||
switch(x.substr(-4)) {
|
||||
case '.xls': if(!cfb.find('/Workbook') && !cfb.find('/Book')) throw new Error("Cannot find workbook for " + x); break;
|
||||
case '.ppt': if(!cfb.find('/PowerPoint Document')) throw new Error("Cannot find presentation for " + x); break;
|
||||
case '.doc': if(!cfb.find('/WordDocument') && !cfb.find('/Word Document')) throw new Error("Cannot find doc for " + x); break;
|
||||
}
|
||||
});
|
||||
});
|
||||
}
|
||||
|
|
@ -29,6 +43,12 @@ describe('should parse test files', function() {
|
|||
parsetest(x, cfb);
|
||||
});
|
||||
});
|
||||
fpres.forEach(function(x) {
|
||||
it('should parse ' + x, function() {
|
||||
var cfb = CFB.read('./test_files_pres/' + x, {type: "file"});
|
||||
parsetest(x, cfb);
|
||||
});
|
||||
});
|
||||
f2011.forEach(function(x) {
|
||||
it('should parse ' + x, function() {
|
||||
var cfb = CFB.read('./test_files/2011/' + x, {type: "file"});
|
||||
|
|
|
|||
|
|
@ -0,0 +1,383 @@
|
|||
var DO_NOT_EXPORT_CFB = true;
|
||||
/* cfb.js (C) 2013-2014 SheetJS -- http://sheetjs.com */
|
||||
/* vim: set ts=2: */
|
||||
/*jshint eqnull:true */
|
||||
|
||||
/* [MS-CFB] v20130118 */
|
||||
var CFB = (function _CFB(){
|
||||
var exports = {};
|
||||
exports.version = '0.10.2';
|
||||
function parse(file) {
|
||||
var mver = 3; // major version
|
||||
var ssz = 512; // sector size
|
||||
var nmfs = 0; // number of mini FAT sectors
|
||||
var ndfs = 0; // number of DIFAT sectors
|
||||
var dir_start = 0; // first directory sector location
|
||||
var minifat_start = 0; // first mini FAT sector location
|
||||
var difat_start = 0; // first mini FAT sector location
|
||||
|
||||
var fat_addrs = []; // locations of FAT sectors
|
||||
|
||||
/* [MS-CFB] 2.2 Compound File Header */
|
||||
var blob = file.slice(0,512);
|
||||
prep_blob(blob, 0);
|
||||
|
||||
/* major version */
|
||||
var mv = check_get_mver(blob);
|
||||
mver = mv[0];
|
||||
switch(mver) {
|
||||
case 3: ssz = 512; break; case 4: ssz = 4096; break;
|
||||
default: throw "Major Version: Expected 3 or 4 saw " + mver;
|
||||
}
|
||||
|
||||
/* reprocess header */
|
||||
if(ssz !== 512) { blob = file.slice(0,ssz); prep_blob(blob, 28 /* blob.l */); }
|
||||
/* Save header for final object */
|
||||
var header = file.slice(0,ssz);
|
||||
|
||||
check_shifts(blob, mver);
|
||||
|
||||
// Number of Directory Sectors
|
||||
var nds = blob.read_shift(4, 'i');
|
||||
if(mver === 3 && nds !== 0) throw '# Directory Sectors: Expected 0 saw ' + nds;
|
||||
|
||||
// Number of FAT Sectors
|
||||
//var nfs = blob.read_shift(4, 'i');
|
||||
blob.l += 4;
|
||||
|
||||
// First Directory Sector Location
|
||||
dir_start = blob.read_shift(4, 'i');
|
||||
|
||||
// Transaction Signature
|
||||
blob.l += 4;
|
||||
|
||||
// Mini Stream Cutoff Size
|
||||
blob.chk('00100000', 'Mini Stream Cutoff Size: ');
|
||||
|
||||
// First Mini FAT Sector Location
|
||||
minifat_start = blob.read_shift(4, 'i');
|
||||
|
||||
// Number of Mini FAT Sectors
|
||||
nmfs = blob.read_shift(4, 'i');
|
||||
|
||||
// First DIFAT sector location
|
||||
difat_start = blob.read_shift(4, 'i');
|
||||
|
||||
// Number of DIFAT Sectors
|
||||
ndfs = blob.read_shift(4, 'i');
|
||||
|
||||
// Grab FAT Sector Locations
|
||||
for(var q, j = 0; j < 109; ++j) { /* 109 = (512 - blob.l)>>>2; */
|
||||
q = blob.read_shift(4, 'i');
|
||||
if(q<0) break;
|
||||
fat_addrs[j] = q;
|
||||
}
|
||||
|
||||
/** Break the file up into sectors */
|
||||
var sectors = sectorify(file, ssz);
|
||||
|
||||
sleuth_fat(difat_start, ndfs, sectors, ssz, fat_addrs);
|
||||
|
||||
/** Chains */
|
||||
var sector_list = make_sector_list(sectors, dir_start, fat_addrs, ssz);
|
||||
|
||||
sector_list[dir_start].name = "!Directory";
|
||||
if(nmfs > 0 && minifat_start !== ENDOFCHAIN) sector_list[minifat_start].name = "!MiniFAT";
|
||||
sector_list[fat_addrs[0]].name = "!FAT";
|
||||
sector_list.fat_addrs = fat_addrs;
|
||||
sector_list.ssz = ssz;
|
||||
|
||||
/* [MS-CFB] 2.6.1 Compound File Directory Entry */
|
||||
var files = {}, Paths = [], FileIndex = [], FullPaths = [], FullPathDir = {};
|
||||
read_directory(dir_start, sector_list, sectors, Paths, nmfs, files, FileIndex);
|
||||
|
||||
build_full_paths(FileIndex, FullPathDir, FullPaths, Paths);
|
||||
|
||||
var root_name = Paths.shift();
|
||||
Paths.root = root_name;
|
||||
|
||||
/* [MS-CFB] 2.6.4 (Unicode 3.0.1 case conversion) */
|
||||
var find_path = make_find_path(FullPaths, Paths, FileIndex, files, root_name);
|
||||
|
||||
return {
|
||||
raw: {header: header, sectors: sectors},
|
||||
FileIndex: FileIndex,
|
||||
FullPaths: FullPaths,
|
||||
FullPathDir: FullPathDir,
|
||||
find: find_path
|
||||
};
|
||||
} // parse
|
||||
|
||||
/* [MS-CFB] 2.2 Compound File Header -- read up to major version */
|
||||
function check_get_mver(blob) {
|
||||
// header signature 8
|
||||
blob.chk(HEADER_SIGNATURE, 'Header Signature: ');
|
||||
|
||||
// clsid 16
|
||||
blob.chk(HEADER_CLSID, 'CLSID: ');
|
||||
|
||||
// minor version 2
|
||||
var mver = blob.read_shift(2, 'u');
|
||||
|
||||
return [blob.read_shift(2,'u'), mver];
|
||||
}
|
||||
function check_shifts(blob, mver) {
|
||||
var shift = 0x09;
|
||||
|
||||
// Byte Order
|
||||
blob.chk('feff', 'Byte Order: ');
|
||||
|
||||
// Sector Shift
|
||||
switch((shift = blob.read_shift(2))) {
|
||||
case 0x09: if(mver !== 3) throw 'MajorVersion/SectorShift Mismatch'; break;
|
||||
case 0x0c: if(mver !== 4) throw 'MajorVersion/SectorShift Mismatch'; break;
|
||||
default: throw 'Sector Shift: Expected 9 or 12 saw ' + shift;
|
||||
}
|
||||
|
||||
// Mini Sector Shift
|
||||
blob.chk('0600', 'Mini Sector Shift: ');
|
||||
|
||||
// Reserved
|
||||
blob.chk('000000000000', 'Reserved: ');
|
||||
}
|
||||
|
||||
/** Break the file up into sectors */
|
||||
function sectorify(file, ssz) {
|
||||
var nsectors = Math.ceil(file.length/ssz)-1;
|
||||
var sectors = new Array(nsectors);
|
||||
for(var i=1; i < nsectors; ++i) sectors[i-1] = file.slice(i*ssz,(i+1)*ssz);
|
||||
sectors[nsectors-1] = file.slice(nsectors*ssz);
|
||||
return sectors;
|
||||
}
|
||||
|
||||
/* [MS-CFB] 2.6.4 Red-Black Tree */
|
||||
function build_full_paths(FI, FPD, FP, Paths) {
|
||||
var i = 0, L = 0, R = 0, C = 0, j = 0, pl = Paths.length;
|
||||
var dad = new Array(pl), q = new Array(pl);
|
||||
|
||||
for(; i < pl; ++i) { dad[i]=q[i]=i; FP[i]=Paths[i]; }
|
||||
|
||||
for(; j < q.length; ++j) {
|
||||
i = q[j];
|
||||
L = FI[i].L; R = FI[i].R; C = FI[i].C;
|
||||
if(dad[i] === i) {
|
||||
if(L !== -1 /*NOSTREAM*/ && dad[L] !== L) dad[i] = dad[L];
|
||||
if(R !== -1 && dad[R] !== R) dad[i] = dad[R];
|
||||
}
|
||||
if(C !== -1 /*NOSTREAM*/) dad[C] = i;
|
||||
if(L !== -1) { dad[L] = dad[i]; q.push(L); }
|
||||
if(R !== -1) { dad[R] = dad[i]; q.push(R); }
|
||||
}
|
||||
for(i=1; i !== pl; ++i) if(dad[i] === i) {
|
||||
if(R !== -1 /*NOSTREAM*/ && dad[R] !== R) dad[i] = dad[R];
|
||||
else if(L !== -1 && dad[L] !== L) dad[i] = dad[L];
|
||||
}
|
||||
|
||||
for(i=1; i < pl; ++i) {
|
||||
if(FI[i].type === 0 /* unknown */) continue;
|
||||
j = dad[i];
|
||||
if(j === 0) FP[i] = FP[0] + "/" + FP[i];
|
||||
else while(j !== 0) {
|
||||
FP[i] = FP[j] + "/" + FP[i];
|
||||
j = dad[j];
|
||||
}
|
||||
dad[i] = 0;
|
||||
}
|
||||
|
||||
FP[0] += "/";
|
||||
for(i=1; i < pl; ++i) {
|
||||
if(FI[i].type !== 2 /* stream */) FP[i] += "/";
|
||||
FPD[FP[i]] = FI[i];
|
||||
}
|
||||
}
|
||||
|
||||
/* [MS-CFB] 2.6.4 */
|
||||
function make_find_path(FullPaths, Paths, FileIndex, files, root_name) {
|
||||
var UCFullPaths = new Array(FullPaths.length);
|
||||
var UCPaths = new Array(Paths.length), i;
|
||||
for(i = 0; i < FullPaths.length; ++i) UCFullPaths[i] = FullPaths[i].toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
for(i = 0; i < Paths.length; ++i) UCPaths[i] = Paths[i].toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
return function find_path(path) {
|
||||
var k;
|
||||
if(path.charCodeAt(0) === 47 /* "/" */) { k=true; path = root_name + path; }
|
||||
else k = path.indexOf("/") !== -1;
|
||||
var UCPath = path.toUpperCase().replace(chr0,'').replace(chr1,'!');
|
||||
var w = k === true ? UCFullPaths.indexOf(UCPath) : UCPaths.indexOf(UCPath);
|
||||
if(w === -1) return null;
|
||||
return k === true ? FileIndex[w] : files[Paths[w]];
|
||||
};
|
||||
}
|
||||
|
||||
/** Chase down the rest of the DIFAT chain to build a comprehensive list
|
||||
DIFAT chains by storing the next sector number as the last 32 bytes */
|
||||
function sleuth_fat(idx, cnt, sectors, ssz, fat_addrs) {
|
||||
var q;
|
||||
if(idx === ENDOFCHAIN) {
|
||||
if(cnt !== 0) throw "DIFAT chain shorter than expected";
|
||||
} else if(idx !== -1 /*FREESECT*/) {
|
||||
var sector = sectors[idx], m = (ssz>>>2)-1;
|
||||
for(var i = 0; i < m; ++i) {
|
||||
if((q = __readInt32LE(sector,i*4)) === ENDOFCHAIN) break;
|
||||
fat_addrs.push(q);
|
||||
}
|
||||
sleuth_fat(__readInt32LE(sector,ssz-4),cnt - 1, sectors, ssz, fat_addrs);
|
||||
}
|
||||
}
|
||||
|
||||
/** Follow the linked list of sectors for a given starting point */
|
||||
function get_sector_list(sectors, start, fat_addrs, ssz, chkd) {
|
||||
var sl = sectors.length;
|
||||
var buf, buf_chain;
|
||||
if(!chkd) chkd = new Array(sl);
|
||||
var modulus = ssz - 1, j, jj;
|
||||
buf = [];
|
||||
buf_chain = [];
|
||||
for(j=start; j>=0;) {
|
||||
chkd[j] = true;
|
||||
buf[buf.length] = j;
|
||||
buf_chain.push(sectors[j]);
|
||||
var addr = fat_addrs[Math.floor(j*4/ssz)];
|
||||
jj = ((j*4) & modulus);
|
||||
if(ssz < 4 + jj) throw "FAT boundary crossed: " + j + " 4 "+ssz;
|
||||
j = __readInt32LE(sectors[addr], jj);
|
||||
}
|
||||
return {nodes: buf, data:__toBuffer([buf_chain])};
|
||||
}
|
||||
|
||||
/** Chase down the sector linked lists */
|
||||
function make_sector_list(sectors, dir_start, fat_addrs, ssz) {
|
||||
var sl = sectors.length, sector_list = new Array(sl);
|
||||
var chkd = new Array(sl), buf, buf_chain;
|
||||
var modulus = ssz - 1, i, j, k, jj;
|
||||
for(i=0; i < sl; ++i) {
|
||||
buf = [];
|
||||
k = (i + dir_start); if(k >= sl) k-=sl;
|
||||
if(chkd[k] === true) continue;
|
||||
buf_chain = [];
|
||||
for(j=k; j>=0;) {
|
||||
chkd[j] = true;
|
||||
buf[buf.length] = j;
|
||||
buf_chain.push(sectors[j]);
|
||||
var addr = fat_addrs[Math.floor(j*4/ssz)];
|
||||
jj = ((j*4) & modulus);
|
||||
if(ssz < 4 + jj) throw "FAT boundary crossed: " + j + " 4 "+ssz;
|
||||
j = __readInt32LE(sectors[addr], jj);
|
||||
}
|
||||
sector_list[k] = {nodes: buf, data:__toBuffer([buf_chain])};
|
||||
}
|
||||
return sector_list;
|
||||
}
|
||||
|
||||
/* [MS-CFB] 2.6.1 Compound File Directory Entry */
|
||||
function read_directory(dir_start, sector_list, sectors, Paths, nmfs, files, FileIndex) {
|
||||
var blob;
|
||||
var minifat_store = 0, pl = (Paths.length?2:0);
|
||||
var sector = sector_list[dir_start].data;
|
||||
var i = 0, namelen = 0, name, o, ctime, mtime;
|
||||
for(; i < sector.length; i+= 128) {
|
||||
blob = sector.slice(i, i+128);
|
||||
prep_blob(blob, 64);
|
||||
namelen = blob.read_shift(2);
|
||||
if(namelen === 0) continue;
|
||||
name = __utf16le(blob,0,namelen-pl);
|
||||
Paths.push(name);
|
||||
o = {
|
||||
name: name,
|
||||
type: blob.read_shift(1),
|
||||
color: blob.read_shift(1),
|
||||
L: blob.read_shift(4, 'i'),
|
||||
R: blob.read_shift(4, 'i'),
|
||||
C: blob.read_shift(4, 'i'),
|
||||
clsid: blob.read_shift(16),
|
||||
state: blob.read_shift(4, 'i')
|
||||
};
|
||||
ctime = blob.read_shift(2) + blob.read_shift(2) + blob.read_shift(2) + blob.read_shift(2);
|
||||
if(ctime !== 0) {
|
||||
o.ctime = ctime; o.ct = read_date(blob, blob.l-8);
|
||||
}
|
||||
mtime = blob.read_shift(2) + blob.read_shift(2) + blob.read_shift(2) + blob.read_shift(2);
|
||||
if(mtime !== 0) {
|
||||
o.mtime = mtime; o.mt = read_date(blob, blob.l-8);
|
||||
}
|
||||
o.start = blob.read_shift(4, 'i');
|
||||
o.size = blob.read_shift(4, 'i');
|
||||
if(o.type === 5) { /* root */
|
||||
minifat_store = o.start;
|
||||
if(nmfs > 0 && minifat_store !== ENDOFCHAIN) sector_list[minifat_store].name = "!StreamData";
|
||||
/*minifat_size = o.size;*/
|
||||
} else if(o.size >= 4096 /* MSCSZ */) {
|
||||
o.storage = 'fat';
|
||||
if(sector_list[o.start] === undefined) sector_list[o.start] = get_sector_list(sectors, o.start, sector_list.fat_addrs, sector_list.ssz);
|
||||
sector_list[o.start].name = o.name;
|
||||
o.content = sector_list[o.start].data.slice(0,o.size);
|
||||
prep_blob(o.content, 0);
|
||||
} else {
|
||||
o.storage = 'minifat';
|
||||
if(minifat_store !== ENDOFCHAIN && o.start !== ENDOFCHAIN) {
|
||||
o.content = sector_list[minifat_store].data.slice(o.start*MSSZ,o.start*MSSZ+o.size);
|
||||
prep_blob(o.content, 0);
|
||||
}
|
||||
}
|
||||
files[name] = o;
|
||||
FileIndex.push(o);
|
||||
}
|
||||
}
|
||||
|
||||
function read_date(blob, offset) {
|
||||
return new Date(( ( (__readUInt32LE(blob,offset+4)/1e7)*Math.pow(2,32)+__readUInt32LE(blob,offset)/1e7 ) - 11644473600)*1000);
|
||||
}
|
||||
|
||||
var fs;
|
||||
function readFileSync(filename, options) {
|
||||
if(fs === undefined) fs = require('fs');
|
||||
return parse(fs.readFileSync(filename), options);
|
||||
}
|
||||
|
||||
function readSync(blob, options) {
|
||||
switch(options !== undefined && options.type !== undefined ? options.type : "base64") {
|
||||
case "file": return readFileSync(blob, options);
|
||||
case "base64": return parse(s2a(Base64.decode(blob)), options);
|
||||
case "binary": return parse(s2a(blob), options);
|
||||
}
|
||||
return parse(blob);
|
||||
}
|
||||
|
||||
/** CFB Constants */
|
||||
var MSSZ = 64; /* Mini Sector Size = 1<<6 */
|
||||
//var MSCSZ = 4096; /* Mini Stream Cutoff Size */
|
||||
/* 2.1 Compound File Sector Numbers and Types */
|
||||
var ENDOFCHAIN = -2;
|
||||
/* 2.2 Compound File Header */
|
||||
var HEADER_SIGNATURE = 'd0cf11e0a1b11ae1';
|
||||
var HEADER_CLSID = '00000000000000000000000000000000';
|
||||
var consts = {
|
||||
/* 2.1 Compund File Sector Numbers and Types */
|
||||
MAXREGSECT: -6,
|
||||
DIFSECT: -4,
|
||||
FATSECT: -3,
|
||||
ENDOFCHAIN: ENDOFCHAIN,
|
||||
FREESECT: -1,
|
||||
/* 2.2 Compound File Header */
|
||||
HEADER_SIGNATURE: HEADER_SIGNATURE,
|
||||
HEADER_MINOR_VERSION: '3e00',
|
||||
MAXREGSID: -6,
|
||||
NOSTREAM: -1,
|
||||
HEADER_CLSID: HEADER_CLSID,
|
||||
/* 2.6.1 Compound File Directory Entry */
|
||||
EntryTypes: ['unknown','storage','stream','lockbytes','property','root']
|
||||
};
|
||||
|
||||
exports.read = readSync;
|
||||
exports.parse = parse;
|
||||
exports.utils = {
|
||||
ReadShift: ReadShift,
|
||||
CheckField: CheckField,
|
||||
prep_blob: prep_blob,
|
||||
bconcat: bconcat,
|
||||
consts: consts
|
||||
};
|
||||
|
||||
return exports;
|
||||
})();
|
||||
|
||||
if(typeof require !== 'undefined' && typeof module !== 'undefined' && typeof DO_NOT_EXPORT_CFB === 'undefined') { module.exports = CFB; }
|
||||
Loading…
Reference in New Issue