code review 8926046: encoding/json: document that marshaling invalid utf-8 s...

encoding/json: document that marshaling invalid utf-8 sequence will return error
Also added docs for InvalidUTF8Error.
Fixes issue 5360.

[minux1, 2013-04-29 15:19:48]

Hello golang-dev@googlegroups.com (cc: golang-dev@googlegroups.com),

I'd like you to review this change to
https://code.google.com/p/go/

[adg, 2013-04-29 16:05:40]

https://codereview.appspot.com/8926046/diff/4001/src/pkg/encoding/json/encode.go
File src/pkg/encoding/json/encode.go (right):

https://codereview.appspot.com/8926046/diff/4001/src/pkg/encoding/json/encode...
src/pkg/encoding/json/encode.go:43: // if invalid UTF-8 sequence is encountered.
s/if invalid/if an invalid/

[minux1, 2013-04-29 16:22:11]

Hello golang-dev@googlegroups.com, adg@golang.org (cc:
golang-dev@googlegroups.com),

Please take another look.

[minux1, 2013-04-29 16:23:17]

Please wait a moment, i will take the chance to document InvalidUTF8Error.

[minux1, 2013-04-29 16:29:46]

PTAL.

Now also added docs for InvalidUTF8Error.

[r, 2013-04-29 22:17:00]

LGTM

https://codereview.appspot.com/8926046/diff/16001/src/pkg/encoding/json/encod...
File src/pkg/encoding/json/encode.go (right):

https://codereview.appspot.com/8926046/diff/16001/src/pkg/encoding/json/encod...
src/pkg/encoding/json/encode.go:42: // String values encode as JSON strings,
InvalidUTF8Error will be returned
s/,/./

[minux1, 2013-04-30 03:22:02]

*** Submitted as https://code.google.com/p/go/source/detail?r=6aab8f011070 ***

encoding/json: document that marshaling invalid utf-8 sequence will return error
Also added docs for InvalidUTF8Error.
Fixes issue 5360.

R=golang-dev, adg, r
CC=golang-dev
https://codereview.appspot.com/8926046

[tylere_gmail.com, 2013-05-01 14:38:12]

Is there anything in the spec that actually says which way this should 
behave? 

As an application developer, the (originally) documented behavior is much 
less annoying - an invalid UTF rune isn't really worth barfing over. 
Especially annoying since there doesn't appear to be any handy stdlib to 
clean a UTF-8 string.

[minux1, 2013-05-01 18:31:47]

On Wed, May 1, 2013 at 10:38 PM, <tylere@gmail.com> wrote:

> Is there anything in the spec that actually says which way this should
> behave?
>
AFAIK, no. the RFC only states that utf-8 is the default encoding, but any
unicode encoding
is acceptable if i read correctly. however, it do say that strings should
be comprised of valid
unicode characters.

>
> As an application developer, the (originally) documented behavior is much
> less annoying - an invalid UTF rune isn't really worth barfing over.
> Especially annoying since there doesn't appear to be any handy stdlib to
> clean a UTF-8 string.
>
i think for marshaling this behavior is more useful that silently
corrupting the string (potentially
binary data) with the unicode replacement character.

[rsc, 2013-05-01 19:18:21]

It's too late to change this behavior. We're just clarifying the docs.