Hi,

On Tue, Jul 28, 2026 at 3:50 PM Pierre Joye <[email protected]> wrote:

> Hey Jakub,
>
> On Tue, Jul 28, 2026 at 5:16 PM Jakub Zelenka <[email protected]> wrote:
>
> >
> > This is exactly a half baked deprecation because we need to keep it for
> internal use anyway (it uses http stream wrapper for chunked encoding) so
> this does not give us any code removal and we still need to maintain it. I
> don’t understand why we need to rush it as there is no real reason for that.
>
> Yes and no. Having internal uses only down the road is significantly
> easier to deal with than having to support this from a userland
> perspective. There is some time to think about it and make it cleaner
> or safer to be used for the purpose of fopen/file_get_contents like
> usages. But the select issues may remain, not sure yet. I don't think
> this is something that can't be solved.
>

I actually resolved that select issue (if it's what I think it is -
limitation using select on filtered streams which I got rid of) in the
current master so it will be part of 8.6. See
https://github.com/php/php-src/pull/20540 .


>
> > Also we cannot just decide not to deprecate it in RC. We would need
> extra approval from RM and then new RFC. We had exactly this sort of
> situation last year with __sleep and it was huge pain to change. It also
> took lots of our time that we could use for much more important stuff.
>
> I hear you.
>
> Important has very different lists depending who we talk to. And
> everyone has time when it is possible, being paid to do it or in their
> free time. Cumulated time to deal with something we know is broken and
> we could fix in a decade or more surely used an order of magnitude
> time for too many people than reveriting a deprecation or adapting it.
> Even if it happens some RMs will need to deal with it, but you and
> others are not alone, dropping a mail here does not hurt and I am sure
> some, or I when I know the area, can jump in too. Communication is
> sometimes hard but amazingly helpful :).
>
>
The problem is that there wasn't enough time to properly investigate if the
things can be improved on the C side. There was just some bug and then
immediate proposal to deprecate it without first not even realising that
it's used by chunked encdoing which I needed to point out. Then the idea
was to deprecate it for user space only which I actually wasn't against as
I assumed there was a proper research into the the usage but that wasn't
the case because Nicolas then pointed out that it is used in Symfony but
the deprecation proposal didn't get removed.

What I think is that deprecation should be proposed only if there a deep
knowledge of the problem and the context which I'm not sure is the case
here (and in some other proposals as well).

Kind regards,

Jakub

Reply via email to