Forum: Mongrel Patching Ruby 1.8.6 p11X To Avoid SEGFAULTs

8c43ed7f065406bf171c0f3eb32cf615?d=identicon&s=25 Zed A. Shaw (Guest)
on 2008-06-25 19:08
(Received via mailing list)
Hey everyone,

I know some of you have ran into the latest security fix causing
SEGFAULTs in Rails applications.  This is apparently due to changes in
the class duplication code in Ruby, but I don't have much more
information.

I do however have instructions for people who need these security fixes
now.  The very nice and smart Hongli created a patch for his Ruby2EE
project that also works for Ruby 1.8.6-p111 or Ruby 1.8.6-p114 with
some modification.

PATCHING P114

Here's how you can use it to patch p114.  Grab the Ruby 1.8.6 p114
source, untar it, then cd into the source directory.  You have to be in
the source directory when you start this process.  Not above it, not
below it, right in it.  I show you this command as the first thing.

---------
$ cd ruby-1.8.6-p114
$ wget http://blog.phusion.nl/assets/r8ee-security-patch-...
...
2008-06-25 12:46:39 (63.1 KB/s) - `r8ee-security-patch-20080623-2.txt'
saved [11939/11939]

$ patch -p1 < r8ee-security-patch-20080623-2.txt
patching file array.c
patching file bignum.c
patching file eval.c
patching file intern.h
patching file io.c
patching file lib/webrick/httpservlet/filehandler.rb
Reversed (or previously applied) patch detected!  Assume -R? [n] n
Apply anyway? [n] n
Skipping patch.
4 out of 4 hunks ignored -- saving rejects to file
lib/webrick/httpservlet/filehandler.rb.rej patching file sprintf.c
patching file string.c
---------

Notice how I had to tell it to skip changes to Webrick?  Nobody here
runs webrick so that's just fine.  After this you can do the
usual ./configure, make, make install and get your Ruby back.

PATCHING P111

The process should be exactly the same, just you won't have to tell it
skip the patch to webrick.

WHAT's IN THIS PATCH?

Hongli collected patches from the FreeBSD crew, and then pulled them
together with a security fix in eval.c he was given.  You can read the
thread here:

http://www.ruby-forum.com/topic/157034

The md5sum that I have for this patchfile is:

74405e3f4a0c1e0484c303a33c0a6f0d  r8ee-security-patch-20080623-2.txt

If your md5sum is different then I recommend contacting Hongli for
help.  Consider giving him money for a short consulting contract since
he obviously knows his shit.

THE CATCH: NOT TESTED BY ME

Alright, so don't go running out trying this shit without some
testing.  Not testing is what got everyone in this mess.  All the
bigger ruby players I know are doing this, and they say it works.
Hongli is using it and it works for him.  You are not a big ruby player
or Hongli.  So, test your stuff completely, then roll it out.

Please report back to me if you have problems with the patch and/or if
it works great for you so I can help some other folks out.

Thanks people.  Always looking out for ya.
F8634aca904bc63cb047cb1bd93bdc74?d=identicon&s=25 Evan Weaver (eweaver)
on 2008-06-26 01:51
(Received via mailing list)
Awesome! Thanks.

Evan
E7cff3cfd41c495e1012227d7dc24202?d=identicon&s=25 Luis Lavena (luislavena)
on 2008-06-26 02:02
(Received via mailing list)
On Wed, Jun 25, 2008 at 7:01 PM, Zed A. Shaw <zedshaw@zedshaw.com>
wrote:
> some modification.
> $ wget http://blog.phusion.nl/assets/r8ee-security-patch-...
> patching file lib/webrick/httpservlet/filehandler.rb
> usual ./configure, make, make install and get your Ruby back.
> thread here:
>
>
> Thanks people.  Always looking out for ya.
>

Thanks Zed, we are merging those with our own patches for Ruby
One-Click Installer (MinGW version).

Right now:
1564 tests, 14742 assertions, 6 failures, 50 errors

I need to take a look at the tests patches too, since before I got
only 3 failures :-D

--
Luis Lavena
AREA 17
-
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams
8c43ed7f065406bf171c0f3eb32cf615?d=identicon&s=25 Zed A. Shaw (Guest)
on 2008-06-26 03:31
(Received via mailing list)
On Thu, 26 Jun 2008 02:02:30 +0200
"Luis Lavena" <luislavena@gmail.com> wrote:

> Thanks Zed, we are merging those with our own patches for Ruby
> One-Click Installer (MinGW version).
>
> Right now:
> 1564 tests, 14742 assertions, 6 failures, 50 errors
>
> I need to take a look at the tests patches too, since before I got
> only 3 failures :-D

If you're using the p114 branch, then those are related to changes in
REXML and soap4r stuff.  If you can find out how to make it all clean
then great, but most people seem to not care about that.

On the p111 it should be much better since p114 was when they added the
soap and xml changes.
E7cff3cfd41c495e1012227d7dc24202?d=identicon&s=25 Luis Lavena (luislavena)
on 2008-06-26 10:52
(Received via mailing list)
On Thu, Jun 26, 2008 at 3:25 AM, Zed A. Shaw <zedshaw@zedshaw.com>
wrote:
>> only 3 failures :-D
>
> If you're using the p114 branch, then those are related to changes in
> REXML and soap4r stuff.  If you can find out how to make it all clean
> then great, but most people seem to not care about that.
>
> On the p111 it should be much better since p114 was when they added the
> soap and xml changes.
>

Thankfully I was wrong, those were a mix of p230 results, sorry for the
noise.

Now we have 4 failures, +1 from the p114 which ruby-core told just to
ignore because are some silly japanesse ruby on windows tests (yeah, I
know, don't get me started).

Thank you man for the heads up, this add the 3rd patch to out
collection based on p114, now fully automated to build One-Click
Installer ;-)

Take care,
--
Luis Lavena
AREA 17
-
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams
67b49d6bc040dfcf5bb982039cde5d32?d=identicon&s=25 Tiago Macedo (Guest)
on 2008-06-26 15:51
(Received via mailing list)
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I turned this patch (
http://blog.phusion.nl/assets/r8ee-security-patch-... ) into
a Gentoo ebuild for 1.8.6p114. I removed the webrick diff as it was
already fixed in p114.

I tested it with Rails 2.1, 2.0 and 1.2 and I had no issues (just the
same minor errors as with plain p114).

http://takk.webreakstuff.com/~tmacedo/ruby-1.8.6_p...

I'm going to deploy it to some applications that are already running on
p114.

Thanks for the help,
Tiago Macedo

Luis Lavena wrote:
| On Thu, Jun 26, 2008 at 3:25 AM, Zed A. Shaw <zedshaw@zedshaw.com> wrote:
|> On Thu, 26 Jun 2008 02:02:30 +0200
|> "Luis Lavena" <luislavena@gmail.com> wrote:
|>
|>> Thanks Zed, we are merging those with our own patches for Ruby
|>> One-Click Installer (MinGW version).
|>>
|>> Right now:
|>> 1564 tests, 14742 assertions, 6 failures, 50 errors
|>>
|>> I need to take a look at the tests patches too, since before I got
|>> only 3 failures :-D
|> If you're using the p114 branch, then those are related to changes in
|> REXML and soap4r stuff.  If you can find out how to make it all clean
|> then great, but most people seem to not care about that.
|>
|> On the p111 it should be much better since p114 was when they added the
|> soap and xml changes.
|>
|
| Thankfully I was wrong, those were a mix of p230 results, sorry for
the noise.
|
| Now we have 4 failures, +1 from the p114 which ruby-core told just to
| ignore because are some silly japanesse ruby on windows tests (yeah, I
| know, don't get me started).
|
| Thank you man for the heads up, this add the 3rd patch to out
| collection based on p114, now fully automated to build One-Click
| Installer ;-)
|
| Take care,
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEUEARECAAYFAkhjlmEACgkQxFuRTtCTMvIhqQCVEJmfcMGxjRh+31iMjnqxFjZC
dACfZDnLfS0mEsByG7NAayMLp7YNs00=
=AExa
-----END PGP SIGNATURE-----
D6c3a72ead48f1276b26bb564a651fe2?d=identicon&s=25 Alvinson Dj (alvinson)
on 2008-06-28 06:22
so good,i will study


Create questionnaires and surveys quickly and easily
www.surveylover.com
This topic is locked and can not be replied to.