Minutes of the Mercury meeting
Fri Oct 6th, 3:15pm to 4:40pm.

Attendees:
	David Jeffery 
	David Overton 
	Mark Brown 
	Zoltan Somogyi 
	Tyson Dowd 
	Fergus Henderson 
	Kevin Glynn 
	Peter Stuckey   (for most of the meeting)

1.  Progress reports.

	We went around the table, with each person in turn discussing what
	they had been working on recently and/or what they were planning
	to work on.

	David Jeffery :
		- Has been working on HAL stuff, and on merging the changes
		  on the HAL branch (`no_alias') of the cvs repository back
		  into the main branch.
		- Has been working on a paper on existential types (with fjh).

	Zoltan Somogyi :
		Has been working with Tom Conway on a paper on the
		more accurate version of the profiler.

		Has been working on handling retry from the middle of
		procedures in mdb.

		Asked whether we should include goal paths in stack traces.
		The concensus was that since we already have the line number,
		that is enough in most cases, but the goals paths should be
		included if you specify the `--detailed' option to the
		`stack' command.

		Has been working on implementing retry in procedures
		which do I/O (by tabling the results).

	Fergus Henderson :
		Has been working on solving a GC-related bug with the
		asm_fast.gc.par grade, and on code reviewing.

	David Overton :
		Has finished coding higher-order support in new mode checker,
		but hasn't tested it yet.

	Mark Brown :
		Has been working on writing up article on declarative debugging
		of ICFP programming contest solution.

		Has been fixing a problem that someone reported where
		if you try to use the declarative debugger on code
		that wasn't compile with the right options, the
		declarative debugger throws an exception rather than
		reporting a decent error message.

		Zoltan suggested that it might be good to add
		initialization code to link all the module_layouts
		into a single program_layout table, so that the
		declarative debugger can tell when the program
		includes some modules which haven't been compiled with
		declarative debugging.  Mark responded that he though
		the declarative debugger could use the program
		representation instead.  Fergus suggested that having
		a unified program_layout might be useful for other
		purposes too.

		There was some discussion about whether the
		declarative debugger could debug shallowly-traced
		code.

	Tyson Dowd :
		Has been working on responding to fjh's review
		comments on IL back-end.

		Has been working on the C++ interface.

		Had been working on a .par GC-related bug with MCORBA,
		now hand-balled that to fjh.

	Kevin Glynn :
		Mentioned `hi' interactive Hugs-like Haskell compiler
		front-end, and asked about Mercury equivalent.  We
		discussed some potential improvements to mdb.

2. Summer Studentship Projects

	Zoltan had another suggestion for a summer studentship
	project: improve/extend the pretty printer (library/pprint.m),
	e.g. to handle depth limits, and possible also get the
	debugger's term browser to make use of these extensions.

3.  Miscellaneous.

	We discussed RTTI support for field names.