Re: Problems with the fix for the BC break introduced in 5.4.29 and 5.5.13
| From: | Remi Collet | Date: | Wed, 18 Jun 2014 17:06:42 +0000 |
| Subject: | Re: Problems with the fix for the BC break introduced in 5.4.29 and 5.5.13 | ||
| References: | 1 | Groups: | php.internals |
| Request: | Send a blank email to internals+get-74976@lists.php.net to get a copy of this message | ||
Le 17/06/2014 16:05, Ferenc Kovacs a écrit :
> Hi,
>
> Could you check out my last mail about the unserialize stuff?:
> http://news.php.net/php.internals/74947
The initial commit try to fix a segfault for bug #67072.
https://bugs.php.net/67072
This case is really awful... something which should not exists.
Ok, a segfault is bad, but we have live with it since years.
My proposal, for a quick solution, trying to be pragmatic and
trying to make the most of php users happy.
Revert in 5.4.30 to the 5.4.28 code (only var_serialize.* + tests)
Revert in 5.5.14 to the 5.5.12 code
This should allow to release without new delay.
In 5.6+ (or master), keep the fix
(don't allow to unserialize things which don't have to be)
And allow newInstanceWithoutConstructor for internal classes.
(make this unserialize hack used by phpunit and doctrine unneeded)
And fix classes which need to be fixed.
Remi.