fringe
14 years agoMakefile: Gather statistics and report them.
Mark Wooding [Wed, 18 Nov 2009 13:32:22 +0000 (13:32 +0000)]
Makefile: Gather statistics and report them.

For now, just the lengths of the source files.  Maybe we can think of
more interesting things to measure later.

14 years agof#: Fix Emacs mode header.
Mark Wooding [Wed, 18 Nov 2009 13:31:45 +0000 (13:31 +0000)]
f#: Fix Emacs mode header.

14 years agocl: Tidy up parser slightly.
Mark Wooding [Wed, 18 Nov 2009 13:28:06 +0000 (13:28 +0000)]
cl: Tidy up parser slightly.

Most of the other implementations don't have separate cases for end-of-
input and not-open-paren; so don't penalize Lisp because I decided to
make the parser more extendable.

14 years agocl: Unpleasant hack to fetch program name.
Mark Wooding [Wed, 18 Nov 2009 13:27:02 +0000 (13:27 +0000)]
cl: Unpleasant hack to fetch program name.

It seems that `cl-launch' doesn't set CL_LAUNCH_FILE when doing `-X',
which is rather inconvenient.  Stuff using `mdw.sys-base' wins because
(massive cheat) ASDF systems are loaded before the program name is
stripped off the front of the argument list; we don't get that luxury
here.  So `*load-pathname*' will have to do; take `pathname-name' to
strip away the unpleasant realities of FASL-caching.

Of course, if we're dumping images then CL_LAUNCH_FILE is set properly.
<sigh>

14 years agoscheme: Use `resume' to switch coroutines.
Mark Wooding [Wed, 18 Nov 2009 11:42:17 +0000 (11:42 +0000)]
scheme: Use `resume' to switch coroutines.

The name `switch' was taken by Chicken as an undocumented `case'-like
control structure.  Grumble.

14 years agocl: Dump a core image to improve startup times.
Mark Wooding [Tue, 17 Nov 2009 23:33:52 +0000 (23:33 +0000)]
cl: Dump a core image to improve startup times.

This is cheating: the core image is massive because SBCL doesn't tree-
shake.

14 years agoMakefile: Print timing information in test output.
Mark Wooding [Tue, 17 Nov 2009 23:28:27 +0000 (23:28 +0000)]
Makefile: Print timing information in test output.

For some reason, C makes a very poor showing in the timings.  I'd love
to know why.  (Guess at `malloc' being a bottleneck.)

14 years agoInitial version.
Mark Wooding [Tue, 17 Nov 2009 21:51:44 +0000 (21:51 +0000)]
Initial version.