Bugs: String literals with escapes \\ in them don't clean up correctly. - fixed Cleanup kills whitespace in strings. i.e: "hello\n world" will get collapsed into "hello\nworld" - fixed, but trying to fix case with weird selection. - fixed ;;;;;;;;;;;;;;; (x (cond (y blah))) ;;;;;;;;;;;;;;; Highlight blah, then press enter. - fixed. Silly index deletion problem. Copy-and-paste does not use the cleanup functions at the moment. Also cut and paste are not dedotting anymore. - fixed - However, copy-and-paste between different applications isn't right anymore. Hmm... - Workaround done. Here strings don't work. - fixed Nested comments #| |# don't work. - fixed, sorta. They're parsed by semi-parse-syntax finally. But Insert mode doesn't quite work right with nested comments. in-something doesn't count how deeply we've nested. - fixed for real Closing a "---" that doesn't have an $expr$ holder kills an sexpr that probably shouldn't die. Copy and paste from the interaction window, before the last unread point, runs into our manger. Check syntax in insert mode does not defocus as it should. All of the button presses should get us out of insert mode. - I couldn't get this to work, so as a workaround I disable Check Syntax while in insert mode. See diva-link.ss's to-command-mode for details. Major slowness on large files. - identified as parser. Inserting a snip in Insert mode breaks. -- reason: the internals of insert mode and interpreter.ss are still very much about symbols. Need to fix this in the next release. Avoid using pregexp for simple string munging. I traced a computation that took about a second using pregexp-replace* --- the equivalent hand-rolled code is about a millisecond. Cursor motion: map: expects type as 2nd argument, given: #f; other arguments were: # === context === /home/dyoo/work/divascheme/traversal.ss:164:5 /home/dyoo/work/divascheme/interpreter.ss:655:2: eval-column-motion /home/dyoo/work/divascheme/diva-link.ss:315:6: interpreter/imperative /home/dyoo/work/divascheme/diva-link.ss:306:6: get&set-mred/handlers call-with-break-parameterization - not sure how this is happening, or if it still is a problem: I've changed the syntax list implemenation to compute lazily now, so it should never be visible as #f from traversal.ss.