Bug#403275: OOo bug: ZWJ: The zero width joiner shouldn't be filtered out
Hi Rene,
On Fri, 2006-12-15 at 23:00 +0100, Rene Engelhard wrote:
> Didn't you write in the issue that you have objections against cmcs
> patch?
"reservations". The reasons were:
1) Was the original bug/patch that caused this regression really valid?
Hence should we just yank the patch that caused the regression instead
of applying another patch.
2) Could the function be written slightly differently to get a
performance benefit at the same time as fixing this bug.
3) The new patch doesn't address control chars required for valid BiDi
analysis in ICU. However, it looks like OOo does the BiDI analysis
itself rather than allow ICU to do it.
cmc's patch is a perfectly reasonable conservative approach to fixing
the problem at hand. My "reservations" were more in-line with whether we
could/should do more.
> Why should I include it in 2.0.4 (which is frozen and I need to get
> approval for any change...) then?
This patch fixes a regression caused by a patch from November 2005.
Without this patch, all scripts which employ the Zero Width Joiner (U
+200D) and Zero Width Non Joiner (U+200C) will be broken. This includes,
all South Asian scripts and most probably Middle Eastern scripts. i.e.
any script that requires ZWJ and/or ZWNJ for the correct rendering of
the script.
Essentially, do you want to ship a version of OOo which is broken for
regions of the world where OOo is a very attractive alternative. FC6
ships with this patch in OOo.
cya,
#
Reply to: