We had a meeting of the Mercury group on Fri April 27, 2001, from 3.15pm to 3.45pm Attendees: Zoltan Somogyi (zs) Simon Taylor (stayl) Mark Brown (dougl) David Overton (dmo) 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. Simon Taylor (stayl): - Still working on smart recompilation. Mark Brown (dougl): - Working on the search strategy for the declarative debugger. - Has been using the declarative debugger to debug itself. Zoltan Somogyi (zs): - Has been working on making the deep profiler more usable. - Started merging the deep profiling CVS branch back onto the main branch. - Produced a script (tools/cvdd) to perform a diff between two workspaces. - Reworked the optimization level handling to avoid invoking the lazy code generator and value numbering. With some recently added optimization passes, the eager code generator produces equally good code with less compilation time. David Overton (dmo): - Has been fixing performance problems with ROBBDs. Fergus Henderson (fjh): Didn't attend the meeting, but sent a progress report by e-mail: - Installed autoconf 2.13 on murlibobo (to fix a problem for Cygwin in the `configure' that we ship with our source distributions). - Tested the Mercury compiler on Tru64 5.1 using an Oct 2000 snapshot of gcc, since our sysadmins are planning to upgrade murlibobo's CPUs and OS. Happily, it worked fine out-of-the-box. - Sent off page for MS TechEd DVD. - Sent off long letter to MS regarding VSIP license. - Chased up GCC copyright assignment (still not yet resolved,). - Fixed a bug where the compiler was silently accepting some incorrect code involving typeclasses. - A little more work towards eliminating dependencies on Cygwin. - Some code reviewing. Minutes taken by Simon Taylor (stayl@cs.mu.oz.au).