code review 8325045: net/url: Add ParseUserinfo() function

net/url: Add ParseUserinfo() function

ParseUserinfo parses a user info string (e.g. foo:bar) into a UserInfo object.

[Hierro, 2013-04-08 08:54:49]

Hello golang-dev@googlegroups.com,

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

[adg, 2013-04-08 12:38:04]

Please hold onto this until after Go 1.1.

[Hierro, 2013-05-14 13:05:28]

On 2013/04/08 12:38:04, adg wrote:
> Please hold onto this until after Go 1.1.

Now that 1.1 is out, PTAL.

Regards,
Alberto

[gobot, 2013-05-14 22:19:15]

R=r (assigned by bradfitz)

[bradfitz, 2013-05-14 22:21:56]

I would rather not see a proliferation of GobEncode and GobDecode functions all
over the standard library.

It gives me Java Serializable flashbacks.

Can't these be registered with gob somehow instead?

https://codereview.appspot.com/8325045/diff/5001/src/pkg/net/url/url.go
File src/pkg/net/url/url.go (right):

https://codereview.appspot.com/8325045/diff/5001/src/pkg/net/url/url.go#newco...
src/pkg/net/url/url.go:298: if _, err := buf.Write([]byte(u.username)); err !=
nil {
WriteString, not Write.  here and elsewhere.

[Hierro, 2013-05-15 12:02:11]

On 2013/05/14 22:21:56, bradfitz wrote:
> I would rather not see a proliferation of GobEncode and GobDecode functions
all
> over the standard library.

Unfortunately, UserInfo objects have a private field, in order to store if they
have a password set (to differentiate no password from an empty password), so
there's no way to for encoding/<whatever> to handle them directly. Maybe we
should remove that private field rather than explicitly implementing
encoding/decoding?

> 
> It gives me Java Serializable flashbacks.
> 
> Can't these be registered with gob somehow instead?
> 
> https://codereview.appspot.com/8325045/diff/5001/src/pkg/net/url/url.go
> File src/pkg/net/url/url.go (right):
> 
>
https://codereview.appspot.com/8325045/diff/5001/src/pkg/net/url/url.go#newco...
> src/pkg/net/url/url.go:298: if _, err := buf.Write([]byte(u.username)); err !=
> nil {
> WriteString, not Write.  here and elsewhere.

Noted, will update those.

Regards,
Alberto

[Hierro, 2013-05-17 12:02:55]

I've restructured the code to move most of it to a new function called
ParseUserinfo(), which parses a user:password string into a Userinfo. This way
GobEncode(), GobDecode() become really small and ParseUserinfo() can be used for
implementing other encoders/decoders too (like e.g. JSON).

PTAL

[Hierro, 2013-05-18 10:12:49]

I have several CLs related to net/http and net/url, which add some functionality
and fix some bugs (including a security issue), depending on this code. Could
you please take a look at it?

[r, 2013-06-12 04:06:13]

remove the gob stuff for now. let's consider the refactoring and API addition
without that.

[Hierro, 2013-06-20 11:30:07]

On 2013/06/12 04:06:13, r wrote:
> remove the gob stuff for now. let's consider the refactoring and API addition
> without that.

Good. Should I update this CL or is it better if I create a new one for the API
addition?

Regards,
Alberto

[Hierro, 2013-06-20 11:36:07]

*** Abandoned ***

[Hierro, 2013-06-20 12:35:03]

On 2013/06/20 11:36:07, Hierro wrote:
> *** Abandoned ***

Closed this by accident, when submitting a CL from another checkout. Reopening
in the meantime.

[bradfitz, 2013-06-20 20:18:53]

On Thu, Jun 20, 2013 at 4:30 AM, <alberto.garcia.hierro@gmail.com> wrote:

> On 2013/06/12 04:06:13, r wrote:
>
>> remove the gob stuff for now. let's consider the refactoring and API
>>
> addition
>
>> without that.
>>
>
> Good. Should I update this CL or is it better if I create a new one for
> the API addition?
>

I'd just update this one.

But whatever.

Close this one if it's dead, though.

[Hierro, 2013-07-15 14:23:42]

PTAL