I have been using Montezuma to compile document indexes of all of the
in-image documentation available in quicklisp loadable systems. This
results in a ~90meg index.
I have been finding and fixing a couple of bugs in Montezuma along the
way. There are patches available at my github acct [1]. If you would
prefer these in some other format, please let me know.
The first patch tries to make sure that error messages remain
printable even in exceptional circumstances (eg slots being unbound).
I did this by wrapping many calls in print object with ignore-slot-
unbound and using with-slots-ignoring-unbound instead of with-slots
(also only in print-object methods).[2]
The second patch fixes a bug in the read-internal method of fs-store.
Previously when I executed queries on the index I built, I always got
EOF errors. I noticed that read-internal was checking against the
return value of read-sequence against length rather than position.[3]
Thanks much for the great software, I hope I can help it improve
slightly,
Russ Tyndall
Acceleration.net
[1]
https://github.com/AccelerationNet/montezuma/tree/accel-error-fixing
[2]
https://github.com/AccelerationNet/montezuma/commit/7ef06e490bb19894df5a0c98de62ea3fa3d81021
[3]
https://github.com/AccelerationNet/montezuma/commit/0b22ec053234a79868f1486c4f35482e10d3f37c