I'm currently going thru the print.c file with the goal of offering a cleaner output for the HTML documents.
I found that we have some parser code for detecting signatures and the beginning of an HTML document while using the showhtml option.
I think it's safer to remove this automatic detection, which works on some cases, but not on all, and instead use the MIME attachments to send HTML documents rather than try to autodetect them.
Well, for signatures, unfortunately, we don't have any specific text/signature MIME type... and text/plain will continue being parsed into HTML. However, there's no standard either that says that signatures will always begin with a "--" line either. By using this rule in hypermail, we're assuming a de-facto standard which isn't really true :)
So, that's why I propose to remove this part of the code.
What do you think?
-Jose
If we're using the showhtml option and we have a line that
has a "--" sequence, we assume this is a signature and switch to using
<PRE>.
If a line begins with <HTML> line begins with a -- sequence, we assume this is the beginning or end of an HTML document and output as is.
While this conversions may work in many cases, they make it impossible to use the -- sequence or <HTML> one in other cases.
For example, if I wrote something like the following:
Here's an HTML example:
--Received on Tue 02 Nov 1999 06:41:16 PM GMT
<HTML>
<HEAD><TITLE>test</TITLE></HEAD>
<BODY>
<P>Hello, web.</P>
</BODY>
</HTML>
-- In this example, you can see blah blah ================ That wouldn't come out correctly thru hypermail with showhtml.
This archive was generated by hypermail 2.3.0 : Sat 13 Mar 2010 03:46:11 AM GMT GMT