[Mono-list] Question on MonoBURG's DAG-related mechanisms

[Nikolaos Kavvadias]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
 
Hi there

i would like to ask you (Mono users :) a few questions regarding MonoBURG.

1. When using the "-e" or "--dag" option for enabling DAG-based
pattern matching, the data structure that the matcher will work on is
an MBState and not an MBTree. It is reasonable that the MBTree would
not suffice since a) it does not support CSEs, b) there is no case for
a multi-output IR tree. Both a),b) occur in DAGs. My question is: how
is the matching algorithm altered when using MBState? Does it have any
relation to the more recent (when compared to BURG 1992 paper) method
of "Finite-State Code Generation"? (PLDI 1999, Fraser-Proebstring). If
we are talking for another technique, is it anywhere documented (e.g.
paper)?.

2. Is there any detailed documentation for MonoBURG? I could only find
the ChangeLog (that doesn't count for documentation :) and the man
page which is helpful but it's just a man page.

3. Is there any toy example (similar to "sample.brg") for a
DAG-enabled matcher?

I would like to thank you in advance.

Kind regards
Nikolaos Kavvadias

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (MingW32)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
 
iD8DBQFEYF7+MPiy0tCWlz4RAsvwAJ9CYkwMbcY4f3LAg7YQ69FTXQbqPQCfVY8i
QH6+kddEgHsoPi/GoKeSCVc=
=GSvk
-----END PGP SIGNATURE-----