[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Rails 3.2 sessions broken with Rack 1.5



On 06/08/2013 10:25 PM, Antonio Terceiro wrote:
On Thu, Jun 06, 2013 at 04:28:23PM -0400, Sam Ruby wrote:
On 06/05/2013 11:50 AM, Antonio Terceiro wrote:

I will be happy to replace my very simple Rails test which yours. Is
there a way to run those tests without installing anything, so that I
can just do `apt-get install rails` and then run the tests? I couldn't
figure this out from a quick read at the README.

I'm now at the point where you should be able to reproduce my
results. See the attached testrails.sh for instructions.

That's awesome - I have just run the whole thing on a clean chroot here!

The only missing part in the script was installing git, but that should
be an issue in the CI environment since git would be a prerequisite for
pulling the test repository in the first place.

Oopsie! :-)

I have no plans to support rvm at all. Can we drop rvm-capistrano there?

Sure. For background, this script is intended to regression test the scenario used in the Agile Web Development with Rails book. That book currently recommends installing via RVM and gems. That's due to the currently lacking support for recent versions of Ruby and Rails on various platforms (its even worse for Mac OS X), and with RVM typically being the path of least resistance to getting a newbie up and running.

I'd love to be able to tell people "apt-get install awdwr" (most likely with a PPA) and you are up and running!

Realistically, though, my book tends to be ahead of the curve -- the current version of the book covers Rails 4. It was made available within 48 hours of the first beta, and I push out an update each time a release candidate ships. It looks like there will be another rc this week, so this will be a busy week.

The TODO list for getting your tests running on a pure-Debian system
look quite encouraging:

- get jquery-ui-rails packaged
- get gorp packaged
- fix the rack issue

I saw you are already working on the rack issue, thanks a lot. Do you
want to tackle the packaging too? :-)

Sure... but I'm time slicing this effort, and I plan to focus first on the rack issue as that has the most unknowns and would be best if I were to get it pushed upstream into Rails itself and let it flow from there back to here.

I'll stop in on #debian-ruby when I get to the point where I may be ready to start packaging.

I've also attached the consolidated output from the run as
checkdepot.html.  A quick spot check of the output indicates that
many (if not all) of the errors are covered by bug 709431[1].

Cool.

Along the way I encountered two other problems that I merely
captured and moved on.  Now that I have completed the first-pass at
this, please let me know the best way to report these problems.

The first problem is that an install of ruby-bcrypt doesn't make the
bcrypt-ruby gem available.  See bcrypt.log.

Cédric already pushed a fix for this (thanks Cédric!).

Yup, thanks!

The second problem is that mysql-server-5.5 isn't installable using
the default configuration.  See mysql.log.

What you mean with default configuration? I tried installing
mysql-server-5.5 (5.5.31-1) on a clean chroot twice (one time setting
the DB root password, and the other not doing it), and both times it
worked for me.

You may want to review the existing bugs at
http://bugs.debian.org/cgi-bin/pkgreport.cgi?pkg=mysql-server-5.5;dist=unstable
and see if any of those matches your problem, if not you might want to
report a new one.

Will do. Again, I didn't look in this with any depth, I just noted it and moved one. Will circle back to this at some point.

So all those tests were run against sqlite3 only, right?

These tests, sure. If you look into awdwr you will see that there is another set of tests: deploydepot and checkdeploy. These bring in Apache, Passenger, Capistrano, MySQL, and RVM. (And, before you ask -- rvm isn't central to these tests either).

- Sam Ruby


Reply to: