18012, 16800, maybe 17565 and 13278:
Integrate contributions from Ben Bucksch (mozilla@bucksch.org)
and Daniel Brattell (bratell@lysator.liu.se) to improve formatted
plaintext output, including adding text style indicators, improving
table output and wrapping, and fixing several wrapping/whitespace bugs.
Also update the auto tests to reflect some of the bugs recently fixed.
r=myself,mozilla@bucksch.org,bratell@lysator.liu.se
format=flowed for mail plaintext output. Restructures the plaintext
output wrapping code, wholly or partially fixes numerous bugs in plaintext
output (13278, 13442, 17823, 17824, 13753, 12551), makes the automated
tests work on Windows and adds some new automated tests. r=akkana.
- Rewrite wrapping logic for plaintext output sink,
and run all output through the wrapping/formatting Write method
(which now has logic to determine when we should be wrapping).
- Handle blockquote type=cite mail quotes.
- Write several new regression tests for plaintext output bugs,
and include those tests and their expected output into the
automated output test.
r=sfraser,harishd
to add new flags (pass a flag in the constructor, rather than specific
booleans. This in order to fix:
11249: Add flag to write only the body in html output.
When prettyprinting is on, we will attempt to make the plaintext
mimic the look of the html; when off (as for copy/paste), we just
dump the text with no formatting information.
In the case of the stream, the document is encoded using the document defined charset
or the user may override the charset.
Added constructors for creating a sink using a nsIOutputStream or an nsString.
Removed all HTML and TEXT encoding which relied on references to ostream. Using John McMullen's
platform independent streaming code in its place. :)
document to HTML to Text. The charset information is first encoded
in XIF and then that information is used when interpretting the unicode
for output.
Added support to output character entity information which should address
bug 4709