'mailread' considered dangerous?


#1

Hoi rubyists,

just wanted to state my opinion about mailreader. Imo it’s freaking
dangerous
to include something in the core distribution which doesn’t get emails
right :slight_smile:

It might be amusing to parse and look some at your emails, but if you
want
to seriously work with them, you can’t just use it. Neither is there any
order
of headers preserved (as is not required but asked for by rfc2822), nor
is
the storage of headers in a hash appropriate (you can have multiple
headers
with the same key and you shouldn’t be losing them). There’s probably
more
serious problems in there.

I’m aware this package is a ‘basic’ mail parsing capability, yet even
though
it’s ‘basic’ it should be correct! Next thing included is an Integer
class
whose ‘+’ method works ok if the number is even but does a ‘-’ when the
number
is odd ? (grmbl) goes off junking half of his code to test rubymail
now

Thanks for listening.

-Martin


#2

On Feb 24, 2006, at 7:56 AM, Martin S. Weber wrote:

Next thing included is an Integer class
whose ‘+’ method works ok if the number is even but does a ‘-’ when
the number
is odd ? (grmb

I don’t know anything about MailRead, but what do you mean in the above?


#3

On Tue, Feb 28, 2006 at 01:10:09AM +0900, Logan C. wrote:

On Feb 24, 2006, at 7:56 AM, Martin S. Weber wrote:

Next thing included is an Integer class
whose ‘+’ method works ok if the number is even but does a ‘-’ when
the number
is odd ? (grmb

I don’t know anything about MailRead, but what do you mean in the above?

It’s a sarcastic analogy for mailread working fine for some kinds of
inputs and just performing plain wrong for other kinds of inputs and the
fact that this fact is just ignored with its inclusion in the standard
distribution.

Would you expect a standard arithmetic package to get exponentation
wrong?

Would you expect a standard email parser package to get email headers
wrong?

-Martin


#4

Looks like this has recently been discussed on ruby-core…

start of thread
http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-core/6828

mention of dropping headers which may appear multiple times
http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-core/6858

I don’t see that anything has been committed to CVS since… though I
could be wrong.

-A