[Webkit-unassigned] [Bug 18994] LANG/LC_ALL influences the result of element.style.opacity

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=18994





--- Comment #54 from Evan Martin <evan at chromium.org>  2009-10-08 12:42:36 PDT ---
Blah, I keep going in circles on this.

UString.cpp's concatenate() is almost exactly what I want (it's where I got the
dtoa handling from) except that it does exponents and works with UStrings.  I
could maybe factor that out into something that writes into a char* buffer
(which is what it does, then convert to a wide buffer as the last step) but
then I'd also need to add special-casing for whether we want exponents or not. 
To handle the TextStream case fully it also needs to behave differently for
significant digits -- it's basically the difference between how %f and %g in
printf handle the precision argument.

I think I bit off more than I should have in my last patch; I will write a new
one that *only* touches the CSS floats (what this bug was initially about) and
not float handling in general.  Since those floats need special handling (no
exponents), it seems reasonable to me to have a special function for it.  I
think even in that case I'll need new baselines for a bunch of SVG tests where
the 10e-6th digit of a float changes by one due to rounding.  :(

Regarding reserving the correct buffer from the start: I think it could be
done, at the cost of uglifying the code a bit.  I will write it and see what
you think.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.