node-gyp/node_modules/tar
2012-07-09 22:44:54 -07:00
..
examples Commit node_modules. 2012-02-10 23:44:09 -08:00
lib Update bundled node-tar to v0.1.13 2012-03-19 19:14:12 -07:00
node_modules update the bundled "block-stream" inside "tar" to v0.0.6. 2012-07-09 22:44:54 -07:00
test Commit node_modules. 2012-02-10 23:44:09 -08:00
.npmignore Commit node_modules. 2012-02-10 23:44:09 -08:00
.travis.yml Commit node_modules. 2012-02-10 23:44:09 -08:00
package.json re-commit all the npm dependencies 2012-03-19 19:17:46 -07:00
README.md Commit node_modules. 2012-02-10 23:44:09 -08:00
tar.js Commit node_modules. 2012-02-10 23:44:09 -08:00

node-tar

Tar for Node.js.

Goals of this project

  1. Be able to parse and reasonably extract the contents of any tar file created by any program that creates tar files, period.

    At least, this includes every version of:

    • bsdtar
    • gnutar
    • solaris posix tar
    • Joerg Schilling's star ("Schilly tar")
  2. Create tar files that can be extracted by any of the following tar programs:

    • bsdtar/libarchive version 2.6.2
    • gnutar 1.15 and above
    • SunOS Posix tar
    • Joerg Schilling's star ("Schilly tar")
  3. 100% test coverage. Speed is important. Correctness is slightly more important.

  4. Create the kind of tar interface that Node users would want to use.

  5. Satisfy npm's needs for a portable tar implementation with a JavaScript interface.

  6. No excuses. No complaining. No tolerance for failure.

But isn't there already a tar.js?

Yes, there are a few. This one is going to be better, and it will be fanatically maintained, because npm will depend on it.

That's why I need to write it from scratch. Creating and extracting tarballs is such a large part of what npm does, I simply can't have it be a black box any longer.

Didn't you have something already? Where'd it go?

It's in the "old" folder. It's not functional. Don't use it.

It was a useful exploration to learn the issues involved, but like most software of any reasonable complexity, node-tar won't be useful until it's been written at least 3 times.