6858 - get rid of PA_HackTable in parser

Status: VERIFIED FIXED

(Core :: DOM: HTML Parser, defect, P3)

Description

[Erik van der Poel]

The PA_HackTable in the parser, which converts CP 1252 to Unicode, is redundant
and in fact buggy. It is missing some of the new 1252 characters such as the
euro. Might be a good idea to get rid of this table in the parser, and make
the parser use the Unicode converter instead (mozilla/intl/uconv).

Comment 1

[Erik van der Poel]

From: Brendan Eich <[e-mail]>

Look at http://lxr.mozilla.org/mozilla/search?string=PA_HackTable --
PA_HackTable is duplicated not only with respect to mozilla/intl's
table, it's duplicated in nsString.cpp and nsString2.cpp.  The nsString
definition and uses were the ones I was questioning: if NCRs and Windows
page compatibility are handled by the parser, why should nsString have
to do anything?

Comment 2

[rickg]

Please notify me when this is fixed, and I'll strip UCS2 stuff out of string.

Comment 3

[brendan-obsolete]

Should the duplication of PA_HackTable in nsString.cpp (or nsString2.cpp, or
both, whatever is in use) be a separate bug?  It can't be [e-mail]'s
bug because the solution there is not to switch nsString to use mozilla/intl's
table -- it's to eliminate the use of any table.  The CP 1252 hacking should be
confined to htmlparser, and not done in a generic string class.

/be

Comment 4

[rickg]

Removed hack table from nsString, and removed ToUCS2() method as well.
Note that the hacktable must remain in the token class in order to correctly map
NCR's. Also removed call to ToUCS() in scanner.

Comment 5

[bsharma]

updated qa contact.

Comment 6

[bsharma]

Verified on:
build: 2001-04-02-09-Mtrunk
platform: WinNT

Marking verified as per above developer comments.