When sending HTML e-mails that contain a background,
BulkMailer parses the HTML code of the e-mail. The goal
of the parsing is to identify the tag that specifies
the background. The identified tag should then be
tweaked so that the background file can be embedded and
linked in the e-mail.
The tag identification algorithm is as follows:
1. locate the "<body" tag
2. check if there is a "background" attribute in it
3. get the background URL (absolute or relative)
4. tweak the background URL to an embedded reference (cid:)
5. locate (and eventually download) the background file
6. embed the
This algorithm works when the background file is
specified directly in the HTML file. If the background
file is specified indirectly, i.e. outside the HTML
file, the algorithm fails. For example, if the
background file is specified in a CSS stylesheet, which
is linked from the HTML file, the algorithm fails to
locate the background file.
To correct this failure, the HTML parsing algorithm
should be extended to indirect references of the
background file.
Logged In: YES
user_id=386220
I am not sure that altering the html is good. Instead of
altering the HTML, perhaps we could validate it and suggests
ways the user can fix it, or ignore the warnings. I fear
that the "fixes" will create other problems, and also users
will not realize or understand what changes are happening to
their html. If users cannot predict what will happen in the
application, they may choose not to use it.
We could also provide ways to attach files.