libdot: wc: update to Unicode 10.0.0

We first have to create a script to programatically generate the lookup
tables.  Markus Kuhn's uniset table works for combining characters, but
it has a hardcoded set for the unambiguous wide characters, and requires
a non-standard preformatted text file for the ambiguous wide characters.
So rather than continuing to patch those pieces together by hand (which
is pretty error prone), write a new python helper that works off of all
the official Unicode database files.  Also, uniset is perl ;).

With that in hand, we can document the procedure for merging Unicode
updates, and then we follow it to import the 10.0.0 release.  There's a
lot of changes here because the last time we updated was Unicode 5.0.0
from ~11 years ago, and because there's a bunch of new single codepoints
sprinkled in different blocks.  The astute reader might also notice that
the previous unambiguous table included some codepoints that this update
does not, but that is by design: the old version included unallocated
blocks under the assumption they'd all be wide in the future (which ended
up not being the case).

Since we're basically throwing out the old tables with completely new,
take the opportunity to reformat to better align with our style guides.

The tests take a hit because we were using an emoji codepoint that under
the older standard had a width of 1, but the new one sets it to 2.

The new helper was sanity checked by running it against the new and old
releases and comparing them against the output of Markus's uniset tool.
Furthermore, many of the new single elements that showed up were spot
checked against the Unicode database to make sure they should be there.

Change-Id: Ifab96c8531be6f8ca927d13f2faebe2c9f1e3001
Reviewed-on: https://chromium-review.googlesource.com/646797
Reviewed-by: Brandon Gilmore <varz@google.com>
Tested-by: Mike Frysinger <vapier@chromium.org>
5 files changed
tree: 5f4d6f5387be8e613a3b0ab559e47191b42c168f
  1. hterm/
  2. libdot/
  3. nassh/
  4. saltpig/
  5. ssh_client/
  6. wam/
  7. wash/
  8. .gitignore
  9. HACK.md
  10. LICENSE
  11. package.json
  12. README.md
README.md

Hello

This repository contains the libdot JavaScript library and some web applications that make use of it.

The official copy of this repository is hosted at https://chromium.googlesource.com/apps/libapps.

There is also a mirror on github at https://github.com/libapps/libapps-mirror. Keep in mind that this mirror may occasionally be behind the official repository.

All changes must go through the Gerrit code review server on https://chromium-review.googlesource.com. Github pull requests cannot be accepted. Please see the HACK.md document in this directory for the details.

Top level directories

  • libdot/ is a small set of JS libraries initially developed as part of hterm, now available as shared code. It provides a base layer for web applications. The code is intended to work in any modern browser, in either a plain web page or a “privileged” environment such as a Chrome platform application or Firefox extension. In practice, it's only been put to use in Chrome platform applications so far.

  • hterm/ is a JS library that provides a terminal emulator. It is reasonably fast, reasonably correct, and reasonably portable across browsers.

  • nassh/ is the Secure Shell Chrome App (currently a “v1.5” app, soon to become a “v2” or platform app) that combines hterm with a NaCl build of OpenSSH to provide a PuTTY-like app for Chrome users.

  • ssh_client/ is the NaCl port of OpenSSH. It is used by nassh to create the Secure Shell App.

  • wash/ is a library for cross-origin virtual filesystems, similar to the Plan 9 filesystem. This directory also contains a simple bash-like shell environment for exploring these filesystems. The code in this directory is a work-in-progress.