On 28 July 2017 at 12:46, Andrey Andreev wrote:
>
> This is a long shot, but here's an idea: an OOP API throwing exceptions.
>
> Would require more work both to implement and use, but it would surely
> be less error-prone and there'd be no BC break.
>
Hi Andrey, Yes that would resolve all BC iss
Hi,
On Fri, Jul 28, 2017 at 2:39 PM, Jakub Zelenka wrote:
> On Fri, Jul 28, 2017 at 12:23 PM, Craig Duncan wrote:
>
>> Hi Nikita,
>>
>> Thanks for your input. Would you vote yes for throwing an exception?
>>
>>
> Just to clarify exceptions are no go for 7.x. It would have to be 8.x and
> it woul
On 28 July 2017 at 12:36, Jakub Zelenka wrote:
>
> Also it would break a big amount of code because in case of converting
> errors to exceptions, you start getting a different exception so it would
> have to be handled. I would call the whole proposal just a big BC break
> (I'm aware that we don'
On Fri, Jul 28, 2017 at 12:23 PM, Craig Duncan wrote:
> Hi Nikita,
>
> Thanks for your input. Would you vote yes for throwing an exception?
>
>
Just to clarify exceptions are no go for 7.x. It would have to be 8.x and
it would be a huge BC break so I'm quite confident that this will fail the
vote
On Fri, Jul 28, 2017 at 11:59 AM, Craig Duncan wrote:
> Hi internals.
>
> As my initial thread about introducing warnings to the JSON functions was
> not immediately flooded with disagreement I took the liberty of creating an
> RFC for official discussion.
>
> The proposal is to have `json_encode
Hi Nikita,
Thanks for your input. Would you vote yes for throwing an exception?
On 28 July 2017 at 12:07, Nikita Popov wrote:
>
> Operations that are expected to fail should never generate warnings. We
> should not design functions such that their correct use requires the use of
> the error su
Hi,
On Fri, Jul 28, 2017 at 1:59 PM, Craig Duncan wrote:
> Hi internals.
>
> As my initial thread about introducing warnings to the JSON functions was
> not immediately flooded with disagreement I took the liberty of creating an
> RFC for official discussion.
>
> The proposal is to have `json_enc
On Fri, Jul 28, 2017 at 12:59 PM, Craig Duncan wrote:
> Hi internals.
>
> As my initial thread about introducing warnings to the JSON functions was
> not immediately flooded with disagreement I took the liberty of creating an
> RFC for official discussion.
>
> The proposal is to have `json_encode
Hi internals.
As my initial thread about introducing warnings to the JSON functions was
not immediately flooded with disagreement I took the liberty of creating an
RFC for official discussion.
The proposal is to have `json_encode()` and `json_decode()` raise warnings
whenever a failure occurs, in
