Fwd: [PATCH] added: openssl_csr_subject() View as Markdown

Wez Furlong wrote:

Two or three comments:

• You leak the dynamically allocated buffer returned by X509_NAME_oneline()
• X509_NAME_oneline() man page says that its use is discouraged in
  favour of smarter functions. Is there a possibility that someone will
  need that functionality?
• if you can solve/answer both of the above, please post a link to the patch

Now uses X509_NAME_print_ex(), using the XN_FLAGS_RFC2253 option. The
only other flags for X509_NAME_print_ex() control the output string. I
was going to add options for the php function to select them, but I
don't see the use as the RFC2253 output is easily parsable and the other
options don't do much (reverse order, or spaces instead of commas as
delimiters).

I also changed the prototype to be more like with the other openssl
functions (returning bool):

bool openssl_csr_subject(mixed csr, string &out)

Now, I don't do very much C coding, and I'm pretty new to the zend API..
One thing I couldn't figure out was why (in php) openssl_csr_subject()
won't put anything into $out if it's passed an undefined variable (where
as openssl_csr_export, which is very similar, works).

For example (let's pretend $csr is a string containing a PEM-encoded CSR):

function test($csr) {
    openssl_csr_subject($csr, $out);
    var_dump($out);
}

will show "NULL", whereas:

function test($csr) {
    openssl_csr_subject($csr, &$out);
    var_dump($out);
}

and

function test($csr) {
    $out = NULL;
    openssl_csr_subject($csr, $out);
    var_dump($out);
}

both work fine, showing string (90) {
"emailAddress=blah@php.net,OU=php...." }

It looks to me like it has something to do with $out not being set
(though, with $out = NULL isset($out) still returns false), but I'm
still not sure why passing $out by reference at call-time would work in
that case.