CR check in base64_decode

[David Relson]

On Tue, 04 Jan 2005 06:59:37 +0100
Matthias Andree wrote:

> David Relson <relson at osagesoftware.com> writes:
> 
> > lexer_v3.l calls text_decode() which calls base64_decode().  The EOL
> > sequence is included in the word_t that's passed along.  An interesting
> > experiment would be to have lexer_v3.l exclude the EOL.....
> 
> Ja. Particularly might this fix some of the performance issues with
> unclosed tags that some users may have observed, although it might still
> be susceptible to overlong lines. These can be chopped into pieces
> though. Longer than c. 1,000 characters including CRLF doesn't pass
> through SMTP anyhow.

I wish it were so.  The RFC doesn't _prohibit_ over 1000 characters and
I've seen messages with significantly longer lines.