Somehow I’m not surprised. You guys take the good stuff and throw away the rest!

Shlomi

In Drizzle we made utf8 the default and optimized around it (the default collatin utf8_general_ci). For anything else? Just use binary.

Cheers,
-Brian

Personally I use case insensitive collations more often (for user supplied data at least).

Just as another example, we can define a VARCHAR, utf8 column on a MEMORY table.
I wasn’t asking for fixed width – but MySQL/MEMORY made it so.

Regards

Thanks, I think we both agree here.
I saw need to mention that because the misconception that utf8 columns will always require only as much storage as needed – is widespread.
So the notion of “you asked for a fixed size column” is not clear to some.

I hope this clarifies.
Regards

well, you asked for a fixed size column, so you got a fixed size column, and as it is fixed size it needs to be big enough to store 10 3 byte utf8 sequences up front

I have the opinion that collations should be case sensitive by default; this makes for faster comparisons.

utf8 encodes ASCII as single character – true; by MySQL and its engines do not necessarily follow. For example, if you have CHAR(10) CHARSET utf8, then each such value will take exactly 30 bytes, regardless of content. See also: MySQL’s character sets and collations demystified

I agree though, utf8 should be introduced as a default encoding, and utf8_general_ci as default collation. AFAIK utf8 stores ASCII characters as single byte values.