# super-globals corruption

- Thread: https://externals.io/message/1461
- Original archive: https://news-web.php.net/php.internals/1461
- Started by Sascha Schumann on 2003-05-13T01:16:02+00:00
- Messages: 5
- 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

---

## #1461 — Sascha Schumann — 2003-05-13T01:16:02+00:00

URL: https://externals.io/message/1461#1461

Are there any known corruption issues with the super globals
    in the current PHP 4.3 CVS?

    While deploying an application on Solaris/Sparc today,
    $_SERVER["HTTP_HOST"] sometimes contained

    - the right value
    - nothing
    - $_SERVER["PATH"]

    After that, I just hardwired the right value.  I'm in the
    process of replicating this on Solaris/x86, but the bug might
    be arch-dependent; so any additional info is appreciated.

    - Sascha

---

## #1466 — Rasmus Lerdorf — 2003-05-13T06:48:00+00:00

URL: https://externals.io/message/1461#1466 · In reply to: #1461

>     Are there any known corruption issues with the super globals
>     in the current PHP 4.3 CVS?
>
>     While deploying an application on Solaris/Sparc today,
>     $_SERVER["HTTP_HOST"] sometimes contained
>
>     - the right value
>     - nothing
>     - $_SERVER["PATH"]
>
>     After that, I just hardwired the right value.  I'm in the
>     process of replicating this on Solaris/x86, but the bug might
>     be arch-dependent; so any additional info is appreciated.

I haven't seen it on Linux/fbsd.

-Rasmus

---

## #1474 — Jay Smith — 2003-05-13T14:17:18+00:00

URL: https://externals.io/message/1461#1474 · In reply to: #1461

Sascha Schumann wrote:
>     Are there any known corruption issues with the super globals
>     in the current PHP 4.3 CVS?
> 
>     While deploying an application on Solaris/Sparc today,
>     $_SERVER["HTTP_HOST"] sometimes contained
> 
>     - the right value
>     - nothing
>     - $_SERVER["PATH"]
> 
>     After that, I just hardwired the right value.  I'm in the
>     process of replicating this on Solaris/x86, but the bug might
>     be arch-dependent; so any additional info is appreciated.
> 
>     - Sascha

I haven't seen that on Solaris/SPARC yet. I'm assuming this is on Apache? 

J

---

## #1476 — Sascha Schumann — 2003-05-13T14:41:30+00:00

URL: https://externals.io/message/1461#1476 · In reply to: #1474

> I haven't seen that on Solaris/SPARC yet. I'm assuming this is on Apache?

    Nope; are you using the current PHP_4_3 CVS in production?

    I have another Solaris 2.6 site running with thttpd since
    mid-Feb on the PHP_4 branch and never saw these problems
    there.  Thus, if it isn't a fault in Solaris 2.8 or the build
    toolchain, it must be a bug in the PHP_4_3 CVS.

    - Sascha

---

## #1478 — Jay Smith — 2003-05-13T14:48:16+00:00

URL: https://externals.io/message/1461#1478 · In reply to: #1476

Sascha Schumann wrote:

> > […]
> 
>     Nope; are you using the current PHP_4_3 CVS in production?
> 
>     I have another Solaris 2.6 site running with thttpd since
>     mid-Feb on the PHP_4 branch and never saw these problems
>     there.  Thus, if it isn't a fault in Solaris 2.8 or the build
>     toolchain, it must be a bug in the PHP_4_3 CVS.
> 
>     - Sascha

We don't run Solaris at all in production, as our Solaris box is somewhat
dated and insanely slow. It's strictly for testing, mostly so we can say
our stuff runs on more than just linux, as we'd like to try and remain as
cross-platform as we can across unices. 

If you need testing, I can test on iplanet (ick), apache and cli/cgi, and if
need be, I could probably set up thttpd. 

J