#38 incorrect character handling

[Bartosz Zapalowski]

CVSMonitor incorrectly assumes character coding for logs.

It should either use Unicode or allow to specify
character set.

The best way of doing this is to use Unicode for every
page cvsmonitor produces.
Log entries should be converted from a given character
set to Unicode based on the choice in admin panel for
every module (separetely) is handles.

[Adam Kennedy]

Logged In: YES
user_id=153576

If I were to make a move towards fixing this, would you be
willing to provide advice and testing for any fixes?

I do have some experience with Unicode, but assistance would
be essential to fixing this.

[Bartosz Zapalowski]

Logged In: YES
user_id=734772

I will provide testing and advice as much as I can.

As far as I got to know you can just use the following code
to convert the strings to unicode:
use Encoding;
from_to($string, $string_encoding, "utf-8");

where $string_encoding is the enconding specified for the
CVS module.

The strings coming from cvsmonitor and not from cvs logs
etc. just you just have to recode to unicode.

There is, however, one problem which was too hard for me to
trace. I put at the beggining of the HTML outputing code the
sentence:
print "Content-type: text/html; charset=iso-8859-2\n\n";

Then when I pointed my webbrowser to cvsmonitor the first line
in the outputed HTML was
Content-type: text/html; charset=ISO-8859-1

and cvsmonitor stopped working (it's database got broken).

However, in cvsmonitor file there is no such string as
'ISO-8859-1' no matter the case. I couldn't find the module
which forced printing this line.

[Adam Kennedy]

Logged In: YES
user_id=153576

The charset thing is an option to CGI::header. The CGI
module defaults to iso-8859-1, but you can set it manually
if needed.

The database is probably going to be the hardest part of
this. I'll need to check up on some things before we attempt it.

[Bartosz Feński]

Logged In: YES
user_id=770596

I'm sure that he will help you with testing these features.
I will try to help you too. It's important to me to have Polish
characters properly encoded on webpages generated by
CVS Monitor.
Take a look at
http://debian.linux.org.pl/cgi-bin/cvsmonitor/cvsmonitor.pl
It doesn't look good. Especially log entries describing changes.
If you create some patches/fixes to handle it properly I'll
willingly test them.