[Seaside-dev] Re: what happened to WASqueakUtf8Codec?

[Lukas Renggli]

> Oh and something else:
> TextConverter newForEncoding: 'ISO-8859-1'
> returns nil. That's why the old code had a special case for that one.

Why would you want to instantiate a codec by name? Normally one of the
instances from the list returned by "WACodex allCodecs" is chosen.

Of course that can be fixed, but is it worth it? I wanted to avoid
forcing platforms to understand all variations of codec names like
'UTF8', 'UTF-8', 'utf-8', 'utf8', #utf8, #'utf-8', #UTF8, #UTF8', ...
but instead just deal with a list of possible names. The new
implementation does not requires a platform to do encoding/decoding.
This makes porting much simpler.

> You see there was a reason the old code was like this. It was tweaked
> and tested over a long period. If you don't understand why something
> is there please ask. At the very least just test whether it works.

I honestly don't know any project that uses anything but the
WANullCodec in Squeak. The encoded Squeak streams are so incredibly
slow, its not believable.

Lukas

-- 
Lukas Renggli
http://www.lukas-renggli.ch