I wrote:
>3) I was unable to successfully insert ASCII 0's into MySQL. It
>would truncate the insert. To remedy this, I am escaping ASCII 0's
>*after* the solution has been scored, but *before* inserting into
>the database. Here's the regex I am using: s/\0/\\0/g; All other
>characters (ASCII 1-255) are being inserted successfully.
OK, I am a moron. Matthew Wickline provided some feedback to my
PGAS changes which spurred me to reconsider my ASCII 0 fix. As I
was preparing to email [EMAIL PROTECTED], I did a final ASCII 0
insert test when I realized that it had been working all along!
It was the mysql prompt that was breaking on ASCII 0, not the database
itself. I discovered this when I queried a solution with ASCII 0
directly to file and then viewed the file in vi. What looked like
"Dave " in the mysql prompt, looked like "Dave^@Hoover" in
vi. :-)
So, that little bit of nastiness has been disposed of. This is good,
because now you will be able to tell in the post-mortem whether
someone used a literal ASCII 0 or the two character \0.
>4) I used Eugene's post-mortem suggestions, tweaking them a little
>in order to not escape tabs, newlines or carriage returns. Here
>is the code being used to display the solutions in the post-mortem:
>
>$code =~ s{([\x01-\x08\x0B\x0C\x0E-\x1f])}
> {'<FONT COLOR="red">^'.($1^"\x40").'</FONT>'}ge;
This has been updated to:
$code =~ s{([\x00-\x08\x0B\x0C\x0E-\x1f])}
{'<FONT COLOR="red">^'.($1^"\x40").'</FONT>'}ge;
which prints ASCII 0 as ^@. This brings up another issue, should
ASCII 0 be represented as \0 or ^@?
>To submit solutions to help with testing:
>http://perlgolf.sourceforge.net/cgi-bin/PGAS/leader.cgi?course=12
>
>To view your test solutions as they would appear in the post-mortem:
>http://perlgolf.sourceforge.net/cgi-bin/PGAS/ref_peek.cgi?id=9&secret=pg
>
>To test the referee interface, let me know and I'll set you up.
>
>To view the source code:
>http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/perlgolf/
