>> Unfortunately query-pr doesn't seem to find closed PRs even if you
>> ask for them by number - or at least I haven't managed to persuade
>> it to - so [...]

> It is really easy to browse our gnats-database with the www-frontend.

Unfortunately there's a small bug somewhere in it.  Look at PR 1251, go
to the end, and scroll back a bit.  Notice there's a point at which
everything suddenly turns into italics.  That point is the following
line (taken from the real PR):

q+A\l(I<8eG_Je<i>P3(rVg6C.B`^?46q]te;SV+AMF3^nJ!A5YX["A_Brq=[OmU]G"W8RPVf't`A%

Unfortunately when this got HTMLified, it turned into

q+A\l(I&lt;8eG_Je<i>P3(rVg6C.B`^?46q]te;SV+AMF3^nJ!A5YX[&quot;A_Brq=[OmU]G&quot;W8RPVf't`A%

Notice the <i> should have been turned into &lt;i&gt; but wasn't.  As a
result, a spurious begin-italics code appears and three characters of
the PR disappear.

I have a script to fetch a PR from the web database, if anyone is
interested.  It needs a networking program akin to Hobbit's netcat;
I'll be glad to send the program I use to anyone, or you can probably
use Hobbit's for the purpose - the script may need tweaking if you do.
The script seems to work fine, except for this bug which of course is
beyond its power to do anything about.

					der Mouse

			    mouse@collatz.mcrcim.mcgill.edu