From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on dcvr.yhbt.net X-Spam-Level: X-Spam-ASN: AS31976 209.132.180.0/23 X-Spam-Status: No, score=-2.3 required=3.0 tests=AWL,BAYES_00,DKIM_SIGNED, HEADER_FROM_DIFFERENT_DOMAINS,RCVD_IN_DNSWL_HI,T_DKIM_INVALID, T_RP_MATCHES_RCVD shortcircuit=no autolearn=no autolearn_force=no version=3.4.0 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by dcvr.yhbt.net (Postfix) with ESMTP id 40E911FAEB for ; Wed, 7 Jun 2017 21:06:54 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751638AbdFGVGw (ORCPT ); Wed, 7 Jun 2017 17:06:52 -0400 Received: from mail-io0-f172.google.com ([209.85.223.172]:35412 "EHLO mail-io0-f172.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751603AbdFGVGu (ORCPT ); Wed, 7 Jun 2017 17:06:50 -0400 Received: by mail-io0-f172.google.com with SMTP id k93so12425295ioi.2 for ; Wed, 07 Jun 2017 14:06:50 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=mime-version:sender:in-reply-to:references:from:date:message-id :subject:to:cc:content-transfer-encoding; bh=Cr3fu8RF+c9KjteBLiGyoaTNsu4ctIrQ1qLPfAqalKI=; b=TIVMOeIhV1KCL0SL3iwo1EyX2SymoBiYqDT1txpXSgPDPZOv59lc6fhPqpGptm7FDR nsDYkpCiWTeVETreJbftsKmlICNn891tv9YBMVt63ZmfegvFGvt8POOsqqE04otg3O8c J9ZkA76qXyw2pjlnc24WVJsryiSdBV6ztTnnf2tUnGaaczwIEMLT3wVMMqNce71nnUUL yI7r6bQPXldJOzgNBnz6P092l4AS0pe5x8G+z80Y4NfKVeN32LU2KBuTB1t+6UzYdqSi AI5jvNuguHDazirs12JsPa7fHBEkpyh51z6CZPT5XOqL8EB2+wMLCkk1COBqqHlM4tm9 VF9A== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:sender:in-reply-to:references:from :date:message-id:subject:to:cc:content-transfer-encoding; bh=Cr3fu8RF+c9KjteBLiGyoaTNsu4ctIrQ1qLPfAqalKI=; b=R3rIhEkCZgq0A6E9VqipXL9i7dpOLER9yzwukUUBZ3tybyHvCBgLWvZ7pms5XItcPv D98lCG2uFKLTSpIK+sz4bL0nm1sswjfknFACOXJwAtC5TvEwhV6rgURNP0zqy5JWPlD+ /EXtU9N3GFEPxxKc1bmeuuoNqtKCY6EtcsNEAjYjKb4Ikz5Gn8TgvFzWhGxeqPj5wh1L EE8uwfIaIeVIbbvQjVvuTfyUaH1rLLb4S5toXH33A+i/XOAknwbmGHZY+N6M+fhub2Wk Vg8gUiHddS6fW2mhSUD6nCaO2195NpMDBS+ak53wgS7HADc0TEq6NG1eLpyhrstLrb52 B1WQ== X-Gm-Message-State: AODbwcBosJs3gQxTBnpkvp7+ql47qrbFvEX4kvFRhkNouY2znxMJD+Lk tIATUEujVUlFN5VaIQDgb+kmA3BIJw== X-Received: by 10.107.160.143 with SMTP id j137mr39103372ioe.20.1496869609821; Wed, 07 Jun 2017 14:06:49 -0700 (PDT) MIME-Version: 1.0 Received: by 10.79.135.7 with HTTP; Wed, 7 Jun 2017 14:06:49 -0700 (PDT) In-Reply-To: References: From: Liam Breck Date: Wed, 7 Jun 2017 14:06:49 -0700 X-Google-Sender-Auth: Y4mWE6DWnK3x171iJ7PR9biAuLA Message-ID: Subject: Re: send-email: Net::SMTP::SSL failure To: =?UTF-8?B?w4Z2YXIgQXJuZmrDtnLDsCBCamFybWFzb24=?= Cc: Git Mailing List Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Sender: git-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: git@vger.kernel.org On Wed, Jun 7, 2017 at 1:43 PM, =C3=86var Arnfj=C3=B6r=C3=B0 Bjarmason wrote: > On Wed, Jun 7, 2017 at 9:39 PM, Liam Breck wrote= : >> On Wed, Jun 7, 2017 at 12:30 PM, =C3=86var Arnfj=C3=B6r=C3=B0 Bjarmason >> wrote: >>> On Wed, Jun 7, 2017 at 8:04 PM, Liam Breck wro= te: >>>> Thanks for your help! >>>> >>>> On Wed, Jun 7, 2017 at 10:16 AM, =C3=86var Arnfj=C3=B6r=C3=B0 Bjarmaso= n >>>> wrote: >>>>> On Wed, Jun 7, 2017 at 7:00 PM, Liam Breck w= rote: >>>>>> On Tue, Jun 6, 2017 at 10:47 PM, Liam Breck = wrote: >>>>>>> >>>>>>> This is configured to send via a gmail account >>>>>>> git send-email --to-cover --cc-cover >>>>>>> >>>>>>> I See >>>>>>> Attempt to reload IO/Socket/SSL.pm aborted. >>>>>>> Compilation failed in require at >>>>>>> /usr/share/perl5/vendor_perl/Net/SMTP/SSL.pm line 6. >>>>>>> BEGIN failed--compilation aborted at >>>>>>> /usr/share/perl5/vendor_perl/Net/SMTP/SSL.pm line 6. >>>>>>> Compilation failed in require at /usr/lib/git-core/git-send-email l= ine 1386. >>>>>>> fatal: 'send-email' appears to be a git command, but we were not >>>>>>> able to execute it. Maybe git-send-email is broken? >>>>>>> >>>>>>> Net/SMTP/SSL.pm v1.04 >>>>>>> >>>>>>> perl v5.26.0 >>>>>>> >>>>>>> Seen in git 2.11.1, 2.12.2, 2.13.0, 2.13.1 on Arch Linux >>>>>> >>>>>> Also fails with perl 5.24.1 & 5.24.0 >>>>>> >>>>>> Last working config was git 2.9.3 on perl 5.24.1 >>>>>> >>>>>> The relevant code from git-send-email is: >>>>>> >>>>>> require Net::SMTP; >>>>>> $smtp_domain ||=3D maildomain(); >>>>>> $smtp_server_port ||=3D 25; >>>>>> $smtp ||=3D Net::SMTP->new($smtp_server, >>>>>> Hello =3D> $smtp_domain, >>>>>> Debug =3D> $debug_net_smtp, >>>>>> Port =3D> $smtp_server_port); >>>>>> if ($smtp_encryption eq 'tls' && $smtp) { >>>>>> require Net::SMTP::SSL; >>>>>> $smtp->command('STARTTLS'); >>>>>> >>>>>> I really wish git bundled its non-core perl libs... >>>>> >>>>> What's the output from just: >>>>> >>>>> perl -MNet::SMTP -we1 >>>> >>>> No output, exit code 0, however... >>>> >>>> $ perl -MIO::Socket::SSL -we1 >>>> Can't load '/usr/lib/perl5/site_perl/auto/Net/SSLeay/SSLeay.so' for >>>> module Net::SSLeay: libssl.so.1.0.0: cannot open shared object file: >>>> No such file or directory at /usr/lib/perl5/core_perl/DynaLoader.pm >>>> line 193. >>>> at /usr/share/perl5/vendor_perl/IO/Socket/SSL.pm line 19. >>>> Compilation failed in require at >>>> /usr/share/perl5/vendor_perl/IO/Socket/SSL.pm line 19. >>>> BEGIN failed--compilation aborted at >>>> /usr/share/perl5/vendor_perl/IO/Socket/SSL.pm line 19. >>>> Compilation failed in require. >>>> BEGIN failed--compilation aborted. >>>> >>>> I don't have {vendor,site}_perl/auto/ tho I have the package for >>>> ssleay installed. >>>> >>>> Since which git release was that required? >>> >>> The actual issue here is that your Net::SSLeay package is broken >>> because it's linked to libssl.so.1.0.0 which has since gone away. You >>> should see that it's missing if you run whatever the Arch equivalent >>> is of these Debian commands: >>> >>> $ dpkg -L libnet-ssleay-perl|grep \.so$ >>> /usr/lib/x86_64-linux-gnu/perl5/5.24/auto/Net/SSLeay/SSLeay.so >>> $ /usr/bin/perldoc -l Net::SSLeay >>> /usr/lib/x86_64-linux-gnu/perl5/5.24/Net/SSLeay.pod >>> $ dpkg -S /usr/lib/x86_64-linux-gnu/perl5/5.24/Net/SSLeay.pod >>> libnet-ssleay-perl: /usr/lib/x86_64-linux-gnu/perl5/5.24/Net/SSLeay= .pod >>> $ dpkg -L libnet-ssleay-perl|grep \.so$ >>> /usr/lib/x86_64-linux-gnu/perl5/5.24/auto/Net/SSLeay/SSLeay.so >>> $ ldd -r /usr/lib/x86_64-linux-gnu/perl5/5.24/auto/Net/SSLeay/SSLea= y.so >>> 2>&1|grep libssl >>> libssl.so.1.1 =3D> /usr/lib/x86_64-linux-gnu/libssl.so.1.1 >>> (0x00007f2523bb5000) >>> $ dpkg -S /usr/lib/x86_64-linux-gnu/libssl.so.1.1 >>> libssl1.1:amd64: /usr/lib/x86_64-linux-gnu/libssl.so.1.1 >>> >>> But that this isn't reported is a bug in git-send-email. This >>> (untested) patch is probably the least invasive and easiest way to >>> deal with this: >>> >>> diff --git a/git-send-email.perl b/git-send-email.perl >>> index 7fd5874436..3f0fcf9040 100755 >>> --- a/git-send-email.perl >>> +++ b/git-send-email.perl >>> @@ -1354,6 +1354,8 @@ EOF >>> die __("The required SMTP server is not >>> properly defined.") >>> } >>> >>> + delete $INC{"Net/SMTP.pm"} if exists $INC{"Net/SMTP.pm"= } >>> + and !defined $INC{"Net/SM= TP.pm"}; >>> require Net::SMTP; >>> my $use_net_smtp_ssl =3D >>> version->parse($Net::SMTP::VERSION) < version->parse("2.34"); >>> $smtp_domain ||=3D maildomain(); >>> >>> on closer inspection none of the other require() uses in that script >>> are run twice, so they don't have the same issue with hiding the >>> initial error. >> >> Arch didn't have packages for these perl modules until recently, >> forcing git users to install them with cpan. And then the new packages >> didn't emit a warning about checking for conflicts in the site_perl/ >> directories. Grrr. >> >> Fixed this by uninstalling /usr/{lib,share}/perl5/site_perl/* I had >> only cleaned out share/ previously. > > Ah, so you installed Net::SSLeay via CPAN, and then upgraded your Arch > openssl, breaking the CPAN-built *.so object? Trouble started after an Arch system upgrade which didn't uninstall site_perl/*, so yes probably. I actually fixed the problem first by installing SSLeay via cpan, then noticed where it installed and removed all the cpan stuff. >> Agreed that send-email should be report errors properly. It's a rather >> essential tool. > > Indeed, do you get a meaningful error if you apply my patch? I'm not a git hacker, so I'm not set up to test this quickly, and I lost a day's work trying to fix Arch :-( >>>>> I have not looked deeply at this, but the error you're getting means >>>>> "we tried to load it before and failed, and here you are trying >>>>> again". >>>>> >>>>> This is almost definitely due to this line in git-send-email: >>>>> >>>>> if (eval { require Net::SMTP; 1 }) { >>>>> >>>>> And more generally, this code is all buggy: >>>>> >>>>> 4 matches for "eval.*require" in buffer: git-send-email.perl >>>>> 153:my $have_email_valid =3D eval { require Email::Valid; 1 }; >>>>> 154:my $have_mail_address =3D eval { require Mail::Address; 1 }; >>>>> 1118: if (eval { require Net::Domain; 1 }) { >>>>> 1129: if (eval { require Net::SMTP; 1 }) { >>>>> >>>>> Well, "buggy" in the sense that we're just happy-go-lucky trying to >>>>> load these modules, and if they have an error we don't report it, the= n >>>>> when we try to load them again perl just emits a generic error saying >>>>> you're trying to require() something that already failed somewhere >>>>> before, a minimal test case for that is: >>>>> >>>>> $ cat /tmp/Fails.pm >>>>> package Fails; >>>>> die "oh noes"; >>>>> $ perl -I/tmp -we 'eval { require Fails }; require Fails' >>>>> Attempt to reload Fails.pm aborted. >>>>> Compilation failed in require at -e line 1. >>>>> >>>>> Whereas what we really want to do is some variant of: >>>>> >>>>> $ perl -MData::Dumper -I/tmp -we 'eval { require Fails } or warn >>>>> $@; require Fails' >>>>> oh noes at /tmp/Fails.pm line 2. >>>>> Compilation failed in require at -e line 1. >>>>> Attempt to reload Fails.pm aborted. >>>>> Compilation failed in require at -e line 1. >>>>> >>>>> Or even the more adventerous, this can have some bad side-effects wit= h >>>>> some libraries (you lie to perl saying you haven't seen it before), >>>>> but I doubt Net::SMTP cares much, particularly when we're just about >>>>> to report an error: >>>>> >>>>> $ perl -MData::Dumper -I/tmp -we 'eval { require Fails } or do { >>>>> delete $INC{"Fails.pm"} }; require Fails' >>>>> oh noes at /tmp/Fails.pm line 2. >>>>> Compilation failed in require at -e line 1.