libretic/etherpad-lite
6
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Normalize inserted text using UNorm

Browse source 
For some reason, the client was sending the server a Unicode-normalized
version of inserted strings. So if for example we inserted the string
'ä' (i.e. \x61\xCC\x88) into the document, what would be sent to the
server would be 'ä' (i.e. \xC3\xA4).

This wouldn't be a problem on its own. BUT JavaScript reports that the
length of the first string is 2, while the length of the second one is
1.

So the command that was being sent to the server was 'Z:1>2*0+1$ä', when
it should really be 'Z:1>1*0+1$ä'. When the `checkRep` method checks the
length of the inserted string, it finds an inconsistency, and
disconnects the client.

We now normalize the inserted string before the command is generated, so
the length is always correct.
This commit is contained in:
Daniel Perez Alvarez 2012-08-13 17:09:02 +01:00
parent 48f1545777
commit 699aa299f8
1 changed files with 2 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

3
src/static/js/contentcollector.js
View file

@ -25,13 +25,14 @@
var _MAX_LIST_LEVEL = 8; var _MAX_LIST_LEVEL = 8;
var UNorm = require('./unorm');
var Changeset = require('./Changeset'); var Changeset = require('./Changeset');
var hooks = require('./pluginfw/hooks'); var hooks = require('./pluginfw/hooks');
var _ = require('./underscore'); var _ = require('./underscore');
function sanitizeUnicode(s) function sanitizeUnicode(s)
{ {
return s.replace(/[\uffff\ufffe\ufeff\ufdd0-\ufdef\ud800-\udfff]/g, '?'); return UNorm.nfc(s).replace(/[\uffff\ufffe\ufeff\ufdd0-\ufdef\ud800-\udfff]/g, '?');
} }
function makeContentCollector(collectStyles, browser, apool, domInterface, className2Author) function makeContentCollector(collectStyles, browser, apool, domInterface, className2Author)