# RC2RC1

- Thread: https://externals.io/message/9104
- Original archive: https://news-web.php.net/php.internals/9104
- Started by Andi Gutmans on 2004-04-13T15:39:22+00:00
- Messages: 7
- Reply by email to internals@lists.php.net
- More for agents: search with https://externals.io/search.md?q={query} · JSON API at https://externals.io/api · documentation at https://externals.io/llms.txt

---

## #9104 — Andi Gutmans — 2004-04-13T15:39:22+00:00

URL: https://externals.io/message/9104#9104

Guys,

I want to roll RC2RC1. There are three major changes I think we should get 
out of the door ASAP:
- zend.ze1_compatibility_mode fix.
- MT-safe crash bug.
- studlyCaps changes.

The only important issue I see is John's change of tiny to throw 
exceptions. As there's no way we will finalize our policy on exceptions, it 
might be the best to roll it back and release RC2. I think the fixes are 
too important to let them linger for too long. John, what says you?

Andi

---

## #9109 — Derick Rethans — 2004-04-13T18:34:02+00:00

URL: https://externals.io/message/9104#9109 · In reply to: #9104

Hello Guys/Gals,



> I want to roll RC2RC1. There are three major changes I think we should get
> out of the door ASAP:

I've one other thing to propose, and that is enabling iconv by default.
Now you might ask why iconv and not mbstring? Well, that's easy as the
author of mbstring (Moriyoshi) thinks of moving it to PECL, out of the
core. Why I think iconv should (and can) be enabled by default is:

- libxml2 which we 'require', requires a working iconv implementation
  (for Linux this is embedded in glibc, for FreeBSD this is the iconv
  library, other Operating Systems of course have some form of
  implementation too).

- conversion between charsets is getting more and more of a problem, and
  to support some of the larger multi-language frameworks a bit better,
  having an implementation in C for this is a very good thing. (There
  are already implementations of conversions out in PHP, which is
  ofcourse awfully slow)

- I am not in favor of enabling things by default *at all*, but I do
  think it's worthwhile to do for iconv.

regards,
Derick

---

## #9110 — John Coggeshall — 2004-04-13T18:43:49+00:00

URL: https://externals.io/message/9104#9110 · In reply to: #9104

The change wasn't actually to make Tidy throw exceptions, but rather to
throw errors instead of exceptions (until recently, Tidy would throw
exceptions in either OO or procedural contexts).

Although I could change Tidy to not throw exceptions, I am inclinded to
leave it as is (maybe change the error code from `E_ERROR` to E_WARNING).
It's really bothersome to me to think of these new PHP 5 OO extensions
being released without throwing exceptions. There is a lot of debate as
we know about when exceptions should be thrown, but I haven't heard
anyone flat out say exceptions in the OO context is bad. Thus, my hope
is that we can roll RC2 and re-visit this issue to come up with a
long-term exceptions standard.

If for some reason we can't agree to throw exceptions internally, I've
got no problem reverting back to no exceptions in Tidy.


John


-----Original Message-----
From: Andi Gutmans [mailto:andi@zend.com] 
Sent: Tuesday, April 13, 2004 11:39 AM
To: internals@lists.php.net
Subject: [PHP-DEV] RC2RC1


Guys,

I want to roll RC2RC1. There are three major changes I think we should
get 
out of the door ASAP:
- zend.ze1_compatibility_mode fix.
- MT-safe crash bug.
- studlyCaps changes.

The only important issue I see is John's change of tiny to throw 
exceptions. As there's no way we will finalize our policy on exceptions,
it 
might be the best to roll it back and release RC2. I think the fixes are

too important to let them linger for too long. John, what says you?

Andi

---

## #9111 — Wez Furlong — 2004-04-13T18:54:14+00:00

URL: https://externals.io/message/9104#9111 · In reply to: #9109

+1 (and it's already on by default in the win32 build).

--Wez.

> I've one other thing to propose, and that is enabling iconv by default.
> Now you might ask why iconv and not mbstring? Well, that's easy as the
> author of mbstring (Moriyoshi) thinks of moving it to PECL, out of the
> core. Why I think iconv should (and can) be enabled by default is:
> 
> - libxml2 which we 'require', requires a working iconv implementation
>   (for Linux this is embedded in glibc, for FreeBSD this is the iconv
>   library, other Operating Systems of course have some form of
>   implementation too).
> 
> - conversion between charsets is getting more and more of a problem, and
>   to support some of the larger multi-language frameworks a bit better,
>   having an implementation in C for this is a very good thing. (There
>   are already implementations of conversions out in PHP, which is
>   ofcourse awfully slow)
> 
> - I am not in favor of enabling things by default *at all*, but I do
>   think it's worthwhile to do for iconv.

---

## #9112 — Ilia Alshanetsky — 2004-04-13T18:58:22+00:00

URL: https://externals.io/message/9104#9112 · In reply to: #9109

> I've one other thing to propose, and that is enabling iconv by default.

+1 given that we enable xml extensions that via libxml2 use iconv, we might as 
well enable the iconv extension as well.

Ilia

---

## #9130 — Andi Gutmans — 2004-04-13T21:52:47+00:00

URL: https://externals.io/message/9104#9130 · In reply to: #9104

Derick,

Can you commit a patch so that I can roll it tomorrow?

Thanks,

Andi

At 08:34 PM 4/13/2004 +0200, Derick Rethans wrote:
>Hello Guys/Gals,
>
>
>
> > […]
>
>I've one other thing to propose, and that is enabling iconv by default.
>Now you might ask why iconv and not mbstring? Well, that's easy as the
>author of mbstring (Moriyoshi) thinks of moving it to PECL, out of the
>core. Why I think iconv should (and can) be enabled by default is:
>
>- libxml2 which we 'require', requires a working iconv implementation
>   (for Linux this is embedded in glibc, for FreeBSD this is the iconv
>   library, other Operating Systems of course have some form of
>   implementation too).
>
>- conversion between charsets is getting more and more of a problem, and
>   to support some of the larger multi-language frameworks a bit better,
>   having an implementation in C for this is a very good thing. (There
>   are already implementations of conversions out in PHP, which is
>   ofcourse awfully slow)
>
>- I am not in favor of enabling things by default *at all*, but I do
>   think it's worthwhile to do for iconv.
>
>regards,
>Derick

---

## #9142 — Derick Rethans — 2004-04-14T07:43:03+00:00

URL: https://externals.io/message/9104#9142 · In reply to: #9130

> Can you commit a patch so that I can roll it tomorrow?

Done.

Derick
