Re: removing some cruft
| From: | Philip Olson | Date: | Tue, 31 May 2011 13:57:36 +0000 |
| Subject: | Re: removing some cruft | ||
| References: | 1 2 3 | Groups: | php.internals |
| Request: | Send a blank email to internals+get-52563@lists.php.net to get a copy of this message | ||
On May 28, 2011, at 10:42 AM, Daniel Brown wrote:
> On Sat, May 28, 2011 at 06:05, Kalle Sommer Nielsen <kalle@php.net> wrote:
>>
>> expose_php = Off?
>
> I think what he and others mean is that they want the option to
> leave the logo, credits, et cetera, completely out of the build at
> compile time.
We also have other possibilities here (this is not a proposal, but rather, a list of options/ideas
to ponder):
0. Do nothing
1. Disable expose_php within the distributed php.ini-* files
Trivial, but changing [at least] php.ini-production is a real possibility.
2. Disable expose_php by default within PHP
We could do this too. Default === when no php.ini file is used.
3. Remove the specific PHP version from the exposed PHP header
There are security concerns although I'm not sure what implications (BC?) or real help this
would offer. For example, version information can be deciphered by other means (e.g., logos and
credits change over time) but maybe it's a help because, for example, displaying phpinfo()
requires direct user action. See also #4.
4. Change its behavior and usage
For example, phpinfo() relies upon expose_php for both the logo and credits information to be
displayed. However, the phpinfo() output clearly exposes PHP (along with the version) regardless of
the expose_php setting so we could instead embed the logo into phpinfo(), and display credits as per
phpcredits(), instead of relying upon expose_php for this as done currently. In other words,
expose_php could be off yet phpinfo() behavior would remain close to the same with either a little
extra markup, or appended credits near the license.
But this raises the question: Why do we have magical logo and credit guids?
5. And a performance review
People are concerned with performance issues (e.g., images loaded into memory) although I'm not
sure what real performance is lost (some have made guesses). Is there a real issue with performance
here? How does it compare with, let's say, loading an additional extension? If so, how can it
be helped?
Regards,
Philip