SSL testing on localhost - segfaults aplenty!

Rails
1

Hello,

I’ve followed the tutorial @
http://weeatbricks.com/2007/10/31/how-to-use-a-rails-app-with-https/ to
setup SSL on localhost, wanting to use https for logging in and such.

Using the method described above I get sensibly higher load times and of
course very very frequent and random segfaults (be it ActiveRecord,
Mongrel or any other .rb causing it) !!

Platform is…well Win32. I’ve tried the method with Ruby 1.8.6 latest,
1.8.7 latest (today), 1.9.1 latest (today) and Ruby on Rails 2.2.2,
2.3.2 . I might add that 1.9.1 latest doesn’t even start the server,
sheesh !

I’m quite baffled, I’ve developed before with 1.8.6, 2.2.2 and no SSL
thoughts, it went splending. Having to went the extra 1-2 minutes to
restart the server every time a segfault occurs is really dampening my
dev time :expressionless:

On secound thought, is there another way of enabling SSL, other than
that from the abovementioned tutorial for Win32 ? There are a lot for
the UNIX world yet for the less inclined in the field only this one.

Any help (Apache config, Mongrel config, other tutorial, anything
whatsoever :slight_smile: ) is greatly appreciated !

Thank you in advance !

2

Where is the segfault?

Does Apache work like normal if you hit http and https ? You should be
able hit the site with both before you even install the rail ssl
require plugin. Mongrel does not care about if the client is talking
http or https with Apache.

Does the site work like normal if you hit Mongrel direct?

3

Hello Freddy,

Thank you for the reply ! Here is one sample segfault after 2 refreshes
on one of the pages :

"Processing MainController#signup (for 127.0.0.1 at 2009-05-20 09:24:34)
[GET]
D:/nu
codes/R/ruby-1.8.6/lib/ruby/gems/1.8/gems/mongrel-1.1.5-x86-mswin32-60/lib/mongrel.rb:114:
[BUG] Segmentation fault
ruby 1.8.6 (2009-03-31) [i386-mswin32]

This application has requested the Runtime to terminate it in an unusual
way.
Please contact the application’s support team for more information."

I tried hitting Mongrel direct but forgot to edit the line in my
/etc/hosts Win32 file - contents thereof are :

127.0.0.1 localhost
127.0.0.1 www.niche.eu
::1 localhost

So it did not work with Mongrel direct until after introducing the first
line. As in the server not responding on http://localhost:3333,
responding on the fake domain name http://www.niche.eu[:3333,
eventually] → on the fake domain name I’m able to do https.

Should I dig into the Apache logs ? Maybe there is a timeout value in
the httpd config, or it is not properly setup ? I must say I’ve never
looked into Apache before, just wanted to have a solution of working
https. Another striking issue is that I cannot get the web interface to
Apache (as in http://localhost) to work, perhaps it has something to do
with the hosts file having those redirects.

Nevertheless, Apache works - going to http://www.niche.eu redirects it
to Mongrel, the site shows up.

Mongrel with https obviously not reachable (https://localhost:3333/).
Apache https with port included not reachable
(https://www.niche.eu:3333/)

Or maybe I didn’t get the hit Mongrel / hit Apache correctly ? :slight_smile:

Thank you for your input ! I’m gonna try and investigate the matter
further.

Freddy A. wrote:

Where is the segfault?

Does Apache work like normal if you hit http and https ? You should be
able hit the site with both before you even install the rail ssl
require plugin. Mongrel does not care about if the client is talking
http or https with Apache.

Does the site work like normal if you hit Mongrel direct?

4

Note to all those bewildered by random segfaults – REPLACE MYSQL 5.1
WITH A PREVIOUS VERSION and then copy a older libMySQL.dll to your ruby
bin directory.

Already seen here :

http://groups.google.com/group/rubyonrails-talk/browse_thread/thread/66c916fabd91406a?pli=1

http://railsforum.com/viewtopic.php?id=28080

Enjoy, it took me more than 1 month to figure out what the bleep it
could be…