Re: [v9.2] make_greater_string() does not return a string in some cases

Thank you for reviewing.

The new version of this patch is attached to this message.

> > But it seems to me that if the datatype is BYTEAOID then
> > there's no need to restore anything at all, because we're not
> > going to call pg_mbcliplen() in that case anyway.  So I think
> > the logic here can be simplified.

I agree with you. The original code does not restore the changed
byte. I removed the lastchar preservation from
make_greater_string().

> > Also, you haven't completely fixed the style issues.  Function
> > definitions should look like this:
..
> > Opening curly braces should be on a line by themselves, not at the end
> > of the preceding if, while, etc. line.
> >
> > "finnaly" is spelled incorrectly.

I'm very sorry for left mixed defferent style a lot. I think
I've done the correction of the styles for function definition
and braces. The misspelled word is removed with whole sentense
because re-cheking just before return had been removed.

> Oh, and there's this:
>
> wchar.c: In function ‘pg_utf8_increment’:
> wchar.c:1376: warning: unused variable ‘success’
> wchar.c: In function ‘pg_eucjp_increment’:
> wchar.c:1433: warning: unused variable ‘success’

Oops... I have rebased the patch and removed all warnings.
make check has been passed all-ok.

I confirmed that the planner decides to use index with proper
boundaries for like expression with the certain characters on
problematic code point, on the database encodings both UTF-8 and
EUC-JP with the database locale is C, and database locale is
ja_JP.UTF8. And also for bytea ends with 0xff and 0x00, 0x01.

This is the third version of the patch.

Regards,

--
Kyotaro Horiguchi
NTT Open Source Software Center