git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCH] send-email: don't call methods on undefined values

From
Junio C Hamano <gitster@pobox.com>
Date
Sep 13, 2013, 06:09 UTC
Message-ID
<xmqqd2odl0ba.fsf@gitster.dls.corp.google.com>
In-Reply-To
<20130909224950.GA101065@vauxhall.crustytoothpaste.net>
"brian m. carlson" <sandals@crustytoothpaste.net> writes:
Show 23 quoted lines
> On Mon, Sep 09, 2013 at 09:45:10AM -0700, Junio C Hamano wrote:
>> "brian m. carlson" <sandals@crustytoothpaste.net> writes:
>> > --- a/git-send-email.perl
>> > +++ b/git-send-email.perl
>> > @@ -1234,7 +1234,7 @@ X-Mailer: git-send-email $gitversion
>> >  				if ($smtp->code == 220) {
>> >  					$smtp = Net::SMTP::SSL->start_SSL($smtp,
>> >  									  ssl_verify_params())
>> > -						or die "STARTTLS failed! ".$smtp->message;
>> > +						or die "STARTTLS failed! ".IO::Socket::SSL::errstr();
>> 
>> I agree that $smtp->message may be bogus at this point, but could
>> "require IO::Socket::SSL" have failed on us in ssl_verify_params?
>> In that degraded mode, we do not do SSL peer verification, but
>> otherwise we would still attempt to talk with the peer, and in such
>> a case, IO::Socket::SSL would not be available to us at this point,
>> no?
>
> Since Net::SMTP::SSL uses IO::Socket::SSL (in fact, it is an
> IO::Socket::SSL), we can be guaranteed that it is, in fact, available at
> this point.  I guess strictly we don't need that require in
> IO::Socket::SSL since we'll already be guaranteed that it exists by the
> require of Net::SMTP::SSL.

That "require" came from around $gmane/230533, which was about making sure the update to pacify newer Net::SMTP::SSL does not break folks with older packages, I think.

I just didn't/don't know the history of Net::SMTP::SSL, and that was where my comments came from.

As long as "Net::SMTP::SSL uses IO::Socket::SSL" has been true since the ancient past, I agree that that 'require' of the latter does not need to be guarded by an 'eval'; at that point, we are already in the Net::SMTP::SSL codepath. And your patch I replied to should be the right thing to do.

Thanks for clarifying.
Previous: brian m. carlson
Message 4 of 4 in “send-email: don't call methods on undefined values”
  1. send-email: don't call methods on undefined valuesbrian m. carlson, Sep 8, 2013
  2. Junio C HamanoSep 9, 2013
  3. brian m. carlsonSep 9, 2013
  4. Junio C HamanoSep 13, 2013

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.