From: James Gray <james@...>
Date: 2008-11-18T05:05:32+09:00
Subject: Re: [solved, kinda] Re: Why does tmail stop my CGI script form working?

On Nov 17, 2008, at 1:54 PM, Chad Perrin wrote:

> On Tue, Nov 18, 2008 at 03:45:37AM +0900, James Gray wrote:
>> On Nov 17, 2008, at 11:38 AM, Chad Perrin wrote:
>>
>>> On Tue, Nov 18, 2008 at 01:20:16AM +0900, Jeremy McAnally wrote:
>>>> Why not try mailfactory if you're just sending e-mail?  If it's  
>>>> just
>>>> tmail screwing up.  I thought perhaps you were parsing e-mail or
>>>> something.
>>>
>>> The problem actually appears to be one of something on the server
>>> being
>>> misconfigured so that I cannot properly install and `require` gems.
>>>
>>> Also . . . as someone pointed out to me off-list, I forgot to  
>>> change a
>>> set of single quotes in the `mail` command into double quotes.  The
>>> line
>>> in question should actually read:
>>>
>>>  `echo "#{body_c}" | mail -s "#{sub_c}" #{to_c}`
>>
>> I would prefer to just avoid passing the body through the shell:
>>
>>  open(%Q{| mail -s "#{sub_c}" #{to_c}}, "w") { |mail| mail <<  
>> body_c }
>
> That's an excellent point, and I wish I had thought of it.  I took  
> your
> advice.  The relevant code now looks like this:
>
>    sent = open( %Q{| mail -s "#{sub_c}" #{to_c} }, 'w' ) do |msg|
>      msg << body_c
>    end

Is sent meant to contain the output?  If so, we should probably change  
that code to:

   sent = open(%Q{| mail -s"#{sub_c}" #{to_c} }, "r+") do |mail|
     mail << body_c
     mail.close_write
     mail.read
   end

>> You could use mail's escapes to avoid passing the subject as well.
>> See the man page for details.
>
> I pored over the man page for a little while, and didn't catch on to  
> what
> you meant by this.  The closest I found to it was the reference to  
> the ~s
> escape for the mail program, but I haven't figured out yet what you  
> meant
> should be done with it if that's your point.

I've never used that feature myself, but I suspect you would use it  
like this:

   mail << "~s #{sub_c}\n#{body_c}"

I'm guessing though and could be wrong.  Double check me there.

James Edward Gray II