Re: [PATCH] gitweb: use decode_utf8 directly
- From
- Ismail Dönmez <ismail@pardus.org.tr>
- Date
- Apr 27, 2007, 09:22 UTC
- Message-ID
- <200704271223.03468.ismail@pardus.org.tr>
- In-Reply-To
- <7v1wi6p4lt.fsf@assigned-by-dhcp.cox.net>
On Friday 27 April 2007 12:07:58 you wrote:
Show 35 quoted lines
> Ismail Dönmez <ismail@pardus.org.tr> writes:
> >> I Google'd a bit but the relevant information seems to be missing about
> >> this error. Anyhow there is no need for a wrapper at all as Encode class
> >> has a decode_utf8 function which fixes the problem I am experiencing too
> >> and chops off the unneeded wrapper.
> >>
> >> Patch against git 1.5.1.2 is attached. Comments welcome.
> >>
> >> P.S: I am using Encode 2.20 from CPAN which is the latest stable version
> >> available.
> >
> > Ping? This patch should be harmless and it fixes a real error, can it be
> > applied please?
>
> I cannot tell if it is harmless. The original used
>
> decode("utf8", $str, Encode::FB_DEFAULT);
>
> and you made them to:
>
> decode_utf8($str);
>
> According to the documentation, decode_utf8($octets [,CHECK])
> should be equivalent to decode("utf8", $octets [,CHECK]), and
> the documentation further says that without CHECK, these
> functions assume Encode::FB_DEFAULT; in other words, these two
> should be equivalent.
>
> Which means that there is something else going on. Your change
> may fix what you observed (I do not doubt that it fixed what you
> observed for you), but without understanding what really is
> going on (iow, why it is a fix, when the documentation clearly
> indicates they should be equivalent and it should not fix
> anything), we cannot tell what *ELSE* we are breaking with this
> change.That might be a bug in Encode itself indeed, I will dig a bit more. Thanks.
Regards, ismail