Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:46963
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 209.85.210.172 as permitted sender)
DomainKey-Status: bad
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:in-reply-to:references:date:message-id:subject:from:to
         :cc:content-type;
        b=q/S1VO4MImIDsSrHqYwqr5zcMXr0zCFI+ogdRXtvpGLww6oK91mQq/+PEonDzXFKtu
         6qSk7/FdM/ZN2ZaAsjlf82dFqGZkP8a6KwT01InOq0LtjZdopchnhH/DtahAyn3xAsDH
         9s83rUF8U+6VT/6epJ5I5h9+8riaaxcabquaw=
MIME-Version: 1.0
In-Reply-To: <679B196B-E276-44BE-87EB-6949A673A205@meteonews.ch>
References: <679B196B-E276-44BE-87EB-6949A673A205@meteonews.ch>
Date: Tue, 9 Feb 2010 01:18:49 +0100
Message-ID: <fe05d1541002081618s223aaa02le922b470fd4fd90@mail.gmail.com>
To: Beat Vontobel <b.vontobel@meteonews.ch>
Cc: internals@lists.php.net
Content-Type: text/plain; charset=ISO-8859-1
Subject: Re: [PHP-DEV] [PATCH] Fix for bug #48607 (ftp_fopen_wrapper)
From: pierre.php@gmail.com (Pierre Joye)

hi,

On Mon, Feb 8, 2010 at 12:03 PM, Beat Vontobel <b.vontobel@meteonews.ch> wrote:

> I'm new to the PHP internals mailing list (and also to PHP internals in
> general), so let's first say hello to everybody and thanks for your work!

Thanks and welcome to the list :)

> I needed to dive into PHP's internals myself because I hit a problem (which
> I later found to be bug #48607) with the ftp_fopen_wrapper: fwrite basically
> doesn't work at all with FTP, because PHP's implementation is in violation
> of the FTP specs. No EOF is signalled at the end of a transmission -- PHP's
> behaviour after writing to an FTP server has to be treated as an aborted
> connection by a clean server. If an fwrite to an FTP server ever succeeded,
> it was only due to lucky timing coincidence or buggy servers.
>
> The patch below fixes this issue for me. I run it successfully since more
> than a month on two high-load production webservers. I'd be happy if
> somebody more familiar with PHP's internals could review it, apply it to SVN
> and then hopefully close bug #48607.

Do you have tests cases please? Also specify which server you use so
we can run the tests in the similar context, or compare with
alternative ftp servers if necessary.

> Just one important note for the reviewer: I wasn't quite sure on how to
> correctly "hand up" errors to the higher levels of the API. That's where I
> just used php_error_docref(), which is probably not the intended way, but
> worked for me to at least generate a warning. I also now set the return
> value of the function on failure, but realized that it's again hardcoded to
> "success" on higher levels of the code. This would probably make for another
> bug report and more patches.

It would be cleaner to fix them as well then no? Can you give it a try?

Thanks for your work!

Cheers,
-- 
Pierre

@pierrejoye | http://blog.thepimp.net | http://www.libgd.org