Howdy,
After Ruby 2.7 entered Debian I re-ran the test, as far as I was aware
ruby-openssl didn't make a release in time for Ruby 2.7, and that seems
to have held true. I have much more recently tried pulling buffering.rb
(the specific file with the bug) from upstream's git, and it would
I suppose this is still an issue with ruby 2.7? The bug hasn't been
reported to upstream, and I don't see any relevant change there. Maybe
we can run the reproducer script again, to decide whether the patch is
needed by https://launchpad.net/ubuntu/+source/ruby2.7/ as well?
--
You received this
** Tags added: bionic-openssl-1.1
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1835968
Title:
Regression in backported patch for openssl 1.1
To manage notifications about this bug go to:
** Changed in: ruby2.5 (Ubuntu)
Status: Confirmed => Fix Released
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1835968
Title:
Regression in backported patch for openssl 1.1
To manage
** Changed in: ruby2.5 (Ubuntu)
Status: Confirmed => Fix Released
--
You received this bug notification because you are a member of Ubuntu
Server, which is subscribed to ruby2.5 in Ubuntu.
https://bugs.launchpad.net/bugs/1835968
Title:
Regression in backported patch for openssl 1.1
To
This bug was fixed in the package ruby2.5 - 2.5.1-1ubuntu1.5
---
ruby2.5 (2.5.1-1ubuntu1.5) bionic; urgency=medium
* Add d/p/restore_buffer_newline_check.patch to fix failure sending
files with mixed newline encoding styles; this regression was
introduced by
** Tags removed: verification-needed
** Tags added: verification-done
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1835968
Title:
Regression in backported patch for openssl 1.1
To manage
Howdy,
Well I tested this before, but the current version in -proposed seems to
work just as expected after a few days.
** Tags removed: verification-needed-bionic
** Tags added: verification-done-bionic
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is
Hello Unit, or anyone else affected,
Accepted ruby2.5 into bionic-proposed. The package will build now and be
available at
https://launchpad.net/ubuntu/+source/ruby2.5/2.5.1-1ubuntu1.5 in a few
hours, and then in the -proposed repository.
Please help us by testing this new package. See
Marc and Andreas are assisting with bileto, will give it a shot.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1835968
Title:
Regression in backported patch for openssl 1.1
To manage notifications
Hm, I've never used "bileto" before, but thanks was wondering what the
rejection was for.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1835968
Title:
Regression in backported patch for openssl 1.1
Bryce,
The existing ruby2.5 2.5.1-1ubuntu1.4 was built in a security-only
bileto ppa and copied to the Ubuntu archive so that it would be suitable
for publication to bionic-security. (https://launchpad.net/~ci-train-
ppa-service/+archive/ubuntu/3662-deletedppa/+build/16681231) Can you
please
Great to hear, and thanks for the quick feedback. I've finalized and
uploaded the package for the SRU team. If they accept, they'll explain
next steps.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
Howdy,
While I haven't done anything specific to put it to the test, I've let
cron do its thing and so far haven't hit any issues on either line
ending type. So I'm going to say it looks good to me!
~Unit 193
--
You received this bug notification because you are a member of Ubuntu
Bugs, which
Thanks Unit 193. The changes have successfully passed review, and just
await your thumbs up on verification before upload.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1835968
Title:
Regression
Howdy,
You came up with the same conclusion I had above, including suspecting
patch 16 first as well. I've added the repo, upgraded, and we'll see
how this goes for a little bit.
~Unit 193
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to
** Description changed:
+ [Impact]
+ Transfer of certain files will fail with an error, which had been
+ transferring properly prior to ruby2.5 (2.5.1-1ubuntu1.4), which
+ introduced improved openssl 1.1 support.
+
+ [Test Case]
+ $ lxc create ubuntu:18.04/amd64 ruby25-sru-1835968-bionic
+ $
** Merge proposal linked:
https://code.launchpad.net/~bryce/ubuntu/+source/ruby2.5/+git/ruby2.5/+merge/370639
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1835968
Title:
Regression in
Unit 193,
I've packaged up a minimal fix into a PPA for testing, can you install
it and verify it does indeed work for the various use cases you had
tested earlier? It would also be extremely helpful if you could run it
against some cases that weren't failing before, just to verify this
doesn't
As an addendum, poking around some more I downloaded the test file, and
noticed it contains a mixture of unix and dos style line encodings:
$ file example.html
example.html: HTML document, UTF-8 Unicode text, with very long lines, with
CRLF, LF line terminators
Maybe irrelevant, but
I don't know Ruby very well, but I've narrowed it down to a specific
line within the patch. The original code had:
if @sync or @wbuffer.size > BLOCK_SIZE or idx =
@wbuffer.rindex("\n")
The patch in question drops the final conditional:
if @sync or @wbuffer.size > BLOCK_SIZE
The rest
With just this single patch reverted, the test case passes:
$ sudo patch -R -p2 <
/home/ubuntu/0017-Reduce-memory-allocation-when-writing-to-SSLSocket.patch
patching file openssl/buffering.rb
ubuntu@ruby25-sru-1835968-bionic:~$ ./testcase-ruby-ssl.rb
1
2
3
4
5
6
6.16.2Message sent via
Running the test case I ran into the same error Christian did, mentioned
in comment #8, but after hacking on the testcase a bit (I don't know
Ruby, but studying the code in /usr/lib/ruby/2.5.0/net/smtp.rb gave
enough clues) I was able to reproduce:
$ ./testcase-ruby-ssl.rb
1
2
3
4
5
6
Looking through the upstream repository for what landed after the
aforementioned patches, there aren't further changes against
buffering.rb. I'm not noticing any clear followup fixes to changes to
patches in this patchset. So, it might be worth testing if this issue
is also reproducible in the
Doing a quick scan through the patches, these look interesting:
* 0013-test-test_x509name-change-script-encoding-to-ASCII-8.patch
* 0014-x509name-refactor-OpenSSL-X509-Name-to_s.patch
* 0015-x509name-fix-handling-of-X509_NAME_-oneline-print_ex.patch
*
Here is a broken-out set of the individual patches included in 0009
-openssl-sync-with-upstream-repository. Perhaps just one of the patches
in this set causes the problem, that we could revert/fix.
** Attachment added: "openssl-sync-with-upstream-repository.tgz"
** Also affects: ruby2.5 (Ubuntu Bionic)
Importance: Undecided
Status: New
** Changed in: ruby2.5 (Ubuntu Bionic)
Status: New => Triaged
** Changed in: ruby2.5 (Ubuntu Bionic)
Importance: Undecided => High
--
You received this bug notification because you are a member of
I got the testcase (and forwarded to xnox/bryce as it is only to be
shared confidently)
I got a auth error at first and modified it to use a local server that one
could spawn in a container. But I still fail at:
/usr/lib/ruby/2.5.0/net/smtp.rb:539:in `initialize': no implicit conversion of
Hash
Howdy,
"to confirm - reverting just [2] (and no other components of this) fixes
the issue for you?" Yes, this is precisely what I am saying. Reverting
just that makes everything work for me.
Eh, the "testcase" isn't really a proper testcase, but it was narrowed
down enough for me to reproduce
Thanks for the extra info Unit193.
The first assumed change [1] is in file debian/patches/0001-openssl-
buffering.rb-no-RS-when-output.patch
But [2] is the offending commit according to your analysis so far. And I
can confirm that it is applied in Bionic. That comes in
** Changed in: ruby2.5 (Ubuntu)
Status: Incomplete => Confirmed
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1835968
Title:
Regression in backported patch for openssl 1.1
To manage
Howdy,
So I got a bit of a simpler testcase worked out (though due to reasons,
it's not something I'd like to make fully public, can PM it) and it
ended up being 251b5be20d5b58c27490f44cdeb6e655f9be6f19 that was giving
me the trouble.
I also made an Eoan chroot and tried the test script, it
Howdy,
My apologies I never mentioned what version I was on. And I went from
2.5.1-1ubuntu1.2 to 2.5.1-1ubuntu1.4 it would seem.
I don't exactly have a minimal testcase since the error happened when
interacting with a smtpd.
As part of the openssl 1.1 upgrade, a couple patches were added to
Also, the title you selected for the bug, "Regression in backported
patch for openssl 1.1" suggests you suspect this relates to the recent
openssl 1.1.1 changes, however I'm unclear on how that would relate to
encoding issues in ruby, so it would help if you could elaborate on your
suspicions
Thanks for flagging this.
Version 2.5.1-1ubuntu1.4 matches to what's currently carried in 18.04 so
am guessing you're on that release. What ruby2.5 version did you
upgrade from? (Check your /var/log/apt/*.log files.)
Also, please provide a minimal test case to reproduce the bug. I.e.
** Changed in: ruby2.5 (Ubuntu)
Importance: Undecided => High
** Tags added: regression-update
** Tags added: rls-bb-incoming
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1835968
Title:
36 matches
Mail list logo
