Re: CJK line breaking

Hi Roland!
>
First of all, totally great work! I hope Tomas can have a better look at
it soon, so this can be integrated!

>No problem. I need a CJK capable word processor more than most. :-)
>
This is just a nitpicking mail, as I don't understand your patch at all
(the functionaly aspect that is). On the other hand, it would be great
to include your work, since proper CJK breaking would be loved by a lot
of people :-)

We of course expect more great patches from you in the future, and
therefor, it's better to 'teach' you Abi's coding/patching style now
that later on; hence this whining mail :-)

>1, moves the functionality into xap_EncodingManager::can_break_at(). The only
>difference now in GR_Graphics::canBreak is that it looks up the characters on
>either side of requested chracter to pass to can_break_at().

Could you s/san_break_at/canBreakAt ? The underscores are very C-like in
our cool C++ code. I see that you made it a static void function. Why
not place it conveniently in a GR_Graphics::canBreakAt() function?

Furthermore, I see that you are #if 0-ing or commenting out existing
code. Just remove the old crap it if it is not needed anymore (ie.
reworked by your patch)! ;-)

>2, Uses UT_return_val_if_fail on pointers instead of UT_ASSERT. There's still a
>call to UT_ASSERT(UT_SHOULD_NOT_HAPPEN) in xap_EncodingManager::can_break_at().
>Is that OK?

This should be OK, as it will not result in a crash if the assert is
reached.

>3, Uses UT_DEBUGMSG.
>
Great!

>4, Tries to use Hugarian Notation for variables, although I'm not totally clear
>on exactly what I should be calling everything.

Looks OK to me.

Another thingy: don't use functions like snprintf on fixed arrays (not
even for debugging code). We've writting classes like UT_UTF8String and
corresponding UT_UTF8String_snprintf functions that will magically
prevent any possible overflowing from occuring.

Keep up the totally great work! It's really appreciated!

  Marc
Received on Thu Mar 10 23:21:07 2005