am 02 11 2009 17:51 schrieb Justin J:

On Nov 2, 2009, at 12:47 AM, Peter Velan wrote:

am 02 11 2009 00:11 schrieb Justin J:

a) An eror still comes up:

"Wide character in print at /data/cgi-bin/ddm4/mail cgi line 1554 "

OK, this is an error that's only in the error log?

yes

OK - this has to do with writing UTF-8 encoded things to something
that doesn't support UTF-8 I don't know what that would be, still
yet We're going to get a lot of these in Dada Mail

If you can provide a simple message that causes this error, I'll look
at it closer, but I don't know if I can completely fix it

After some investigation I managed to reproducibly bring up this error with following constellation:

Send a Message: - Subject: € - Body: plain ASCII

[Send Test Message]

I sent the test message with "€" in subject three times varying Default Character Set in "Advanced Sending Preferences" as follows - Default Character Set: "UTF8 UTF8" - Default Character Set: "German (de) iso-8859-1" - Default Character Set: "German (de) windows-1252"

In all three cases I got

mail cgi: Wide character in print at /dadapath/mail cgi line 1554

in the error log

I could not trigger the error with other non-ASCII chars like umlauts (äöü), szligatur (ß), superscripts (²³) or µ All these characters have a numeric represantion < 128!

I had more luck with "š" (numeric representation: 161)

Could a linked stylesheet be included in the email itself like images (after setting the option appropriately)? I don't like to load additional parts from outside to the newsletter

In fact, it can! On the, "Send a Webpage" screen, if you select the
option,

Images in this Webpage Should: be embedded in the message itself

Thats the setup I'm using, so I could try to link in a css-file

The stylesheet will go along for the ride, as well There's no way to
say, "embed the stylesheet, not the images", I don't think

If I understand you correctly, then the "embedding option" will embed any linked in files like images, stylesheets, javascripts media-files, Nice to know!

May be the option

Images in this Webpage Should:
     be embedded in the message itself

should be worded a little bit differently, then

I'll make a note to change the automatic plain text creation to just simply use only what's in the body tag of the HTML message That'll save a lot of grief

Nice plan

I've now done this - the HTML to PlainText conversion now just looks
between the body tags and converts that

Ah fine

b2) Any german umlaut (in the webpage these kind of characters are coded as HTML entities is lost in the text version

Very strange, but I can recreate the problems here, as well I'll
look into this too

Thanks for looking after this

I found out what was happening here:

The built in HTML to PlainText converter basically just strips away
anything that's not an ascii character Brilliant

brilliant, indeed!

I'm not going to
change this behavior - I do say that the converter is pretty bad, so
it's at least advertised to be dumb, but! I've given Dada Mail support
for something called, "HTML::FormatText::WithLinks", which, if
available, can be used instead to do the conversion and it works a
whole lot better than whatever is I made one day

All that needs to be done to enable it is to install the CPAN perl
module:

HTML::FormatText::WithLinks

If you already know how to do that, you are in good shape! :)

I've posted the changes I've made since my last email to the git repo:

http://github com/justingit/dada-mail/commit/aabe397a8010d1711533acd687aadf557c6f820f

These changes add support for HTML::FormatText::WithLinks and also
fixes some of the problems with the built-in converter, basically with
parsing the, "head" of a HTML message

Need some more time to check this

Thanks a lot for the good job,

Peter