Hello everybody.
karajorma told me to fix this bug
http://scp.indiegames.us/mantis/view.php?id=1992. I still don't have a working mantis account

(BTW: can someone assign a random password to pedro mantis account ?), that's why I'm posting here a note in order to share my correction proposal.
The color bug has two causes:
1 - brief_color_text_init terminates every lines by a {letter=0 color=0 } colored_char. It is useless to add a strign termination character because the SCP_vector stores the string length. In addition it makes SCP_vector.size() = string length +1 which is not intuitive.
2 - brief_render_line draws briefing lines made of character sequences separated by spaces with the color of the last character of the sequence. FUBAR-BDHR notes shows every line are terminated by a {letter=0 color=0 } colored_char. This char is part of the character sequence because it is not a whitespace (whitespace is char 32, not 0), as a consequence the color 0 is used to draw the last character sequence of a line. This is surprising that this method tries to detect words because the coloring information have already been built by brief_color_text_init function. It would be more appropriate to focus on character colors.
I propose a double fix where the two fix are sufficient to solve mantis issue (
http://scp.indiegames.us/mantis/view.php?id=1992):
1 - Modify brief_color_text_init to remove the extra terminal character
2 - Rewrite brief_render_line to work with character sequences delimited by color change.
Note that the current implementation of brief_color_text_init does not recognize ',.:;()' as word separators which causes the dot of "$g DELTA." string to be green whereas It could be considered to be white. I can fix it too, but it's a behavior modification, and I'm not able for now to be sure it's not going to add bugs.
Corrections are tested. I have a svn patch waiting for approval. I'm still not aware of the code commit procedure. Can someone show me the way ?
Thanks.