Fri Mar 24 15:40:10 GMT 2000 da Tested Proof General on win32 (NT4sp6) with Isabelle, Coq and XEmacs 21.1.9. Problems: Isabelle bombs on paths containing chars like \ : $ XEmacs barfs on reading some files in PG, why?? e.g. coq/coq.el Can fix this by reloading coq stuff again. A few probs remain, e.g. toolbar enablers for undo are flaky. Wed Mar 22 13:45:34 GMT 2000 da Tested file name quoting problem with Coq, Isabelle, LEGO. \ quoting triggers bug in Isabelle (complaint about pathname) " quoting not allowed in LEGO, \ quoting not needed. Coq works well with either. -------- Wed Nov 17 13:43:11 GMT 1999 Tested compiled version. Seems to work well for XEmacs! Also for FSF Emacs! So long as using their own elc's. Tue Nov 16 15:28:51 GMT 1999 Tested automatic multiple file handling: see etc/demoisa Tested FSF Emacs, after fixing several things. Tested proof by pointing in LEGO. Fixes for bugs, empty pbp response, and added a useful click (C-button 3) for undoing and deleting the last pbp command. Can now prove example.l by PBP. Proof-by-pointing lives! Thu Nov 11 19:05:39 GMT 1999 Tested response buffer display: see isa/message-test.ML Made output more regular by removing spurious space/newline after every prompt, and padding response buffer with newlines when messages aren't newline terminated. Testing window management for multiple frames. Could find no evidence of old bug message in code about with special-display-regexps, script buffer gets made into dedicated buffer. Removed the comment. Mon Aug 23 19:00:26 BST 1999 da Summary of tests today: Proof General 2.0: sanity check. Okay with XEmacs 20.4, lego 1.3.1 and Isabelle 98p1. Strange overlay disappearing problem with FSF Emacs 20.2, so must be X Server or architecture anomaly that causes different display order. Today's Proof General. 1. With Isabelle 98p1, no go. 2. Same show-stopper as above with Emacs 20.2 and 20.3. Argh! I'm really fed up of FSF Emacs, it goes wrong even when "nothing" has changed. 3. With current Isabelle (or, at least, 99pre180899). 4. Using x-symbol. No success, and a big mess (rebinds M-x !!?) Thu Jan 21 14:27:36 GMT 1999 da Quick test for pipe communication with emacs 20.3. Used C-c C-n and C-c C-u on example.ML file in Isabelle successfully. (Tested 990115 prerelease with piped communication patch in XEmacs already, for LEGO and Isabelle). Wed Dec 16 15:45:53 GMT 1998 da Quick test of Coq mode. xemacs -q -l ProofGeneral/generic/proof-site.el (setq proof-rsh-command "ssh hope") Assertion and retraction commands work as far as I can tell. Using toolbar on file coq/example.v Wed Dec 16 12:25:00 GMT 1998 tms Clarification of entry "Mon Dec 14 15:02:52 GMT 1998 da" The problem with LEGOVERSION "alpha" can also be reproduced with lego 1.3.1 (and XEmacs 20.4 or FSF Emacs 20.2) and the file lego/example2.l which accesses a module in a non-writable directory. You need to set chmod u-w readonly yourself; CVS doesn't like non-writable directories) It is a LEGO specific problem. LEGO forgets about annotations sometimes. This has been reported to lego@dcs.ed.ac.uk . Wed Dec 16 12:25:00 GMT 1998 tms On scar, tested Emacs 20.2.1 with lego 1.3.1 via "ssh craro", LEGOVERSION "std" emacs-20.2 -eval '(progn (load "/home/tms/emacs/ProofGeneral/generic/proof-site.el")(setq proof-rsh-command "ssh craro"))' lego/example.l Pressing C-c C-n crashes Emacs: Fatal error (11).Segmentation fault This must be a problem with my .emacs file. Including -q in the options, everyting seems to work just fine. Still, this is somewhat concerning. Mon Dec 14 15:02:52 GMT 1998 da Tested Emacs 20.2.1 with lego 1.3 via "ssh hope", with lego 1.3.1 via "ssh craro", LEGOVERSION "std" Both successfully process example.l With lego 1.3.1 via "ssh craro", LEGOVERSION "alpha", processing gets stuck, never reports "imports done". Is this a bug or problem with LEGO installation? Bugs: Killing off process shell via proof-shell-exit. Killing proof script buffer gives error.