[cells-devel] Re: Celtk/Cells3 declarative, dataflow paradigm contrasted with Ltk/imperative paradigm - cells-devel - mailman3.common-lisp.net

5 Apr 2006


      John,
Sorry, it was late, I reacted too hastily. That error message looks wild,
not sure what to make of it.
Unfortunately, I have no Linux system. But the original developers of LTk
use Linux and are looking at Celtk these days. I have not heard from them,
but a query on the cells-devel mailing list might turn up someone using both
Linux and Celtk.
Otherwise, it might helpt to track down tk-format-now and tweak it so that
the branch to display messages sent to Tk always fires, then see (a) if
those messages look okay (post them to cells-devel) and (b) how far you get
before Celtk dies.
kt
On 4/5/06, John Landahl john@landahl.org wrote:
...
...
Right at the top of ltktest-cells-inside.lisp you will find:
#+test-ltktest
(progn
  (cells-reset 'tk-user-queue-handler)
  ; ...huge comment elided....
  (tk-test-class 'ltktest-cells-inside))
Yep, saw that, and I suppose I expected that loading celtk would set
test-ltktest.  Removing the #+ and re-evaluating does startup the
test.
However, the result is a bunch of errors and a big traceback.  Here's
what looks like the most useful chunk:
0> numparse | :=> 42read-data:UNKNOWNcolor name "systembuttonface"
read-data:INVALIDcommand name ".cv-scroller.test-canvas"
INVALIDCOMMANDNAME.cv-scroller.test-canvasUNKNOWNCOLORNAMEsystembuttonfaceBADWINDOWPATHNAME.cv-scroller.test-canvasINVALIDCOMMANDNAME.cv-scroller.test-canvas.bkg-popINVALIDCOMMANDNAME.cv-scroller.test-canvas.bkg-popINVALIDCOMMANDNAME.cv-scroller.test-canvas.bkg-popINVALIDCOMMANDNAME.cv-scroller.test-canvas.bkg-popCANREAD.cv-scroller.test-canvas.bkg-pop.bkg
An error of type READER-ERROR has occured: READER-ERROR on
#<TWO-WAY-STREAM
                                                             :INPUT-STREAM
#<SB-SYS:FD-STREAM for "descriptor 14" {A755EC9}>
:OUTPUT-STREAM #<SB-SYS:FD-STREAM for "descriptor 13" {A755541}>>:
illegal terminating character after a colon: #\
0: (SB-DEBUG:BACKTRACE
    536870911
    #<SB-SYS:FD-STREAM for "a constant string" {AFCDB01}>)
1: (LTK::TRIVIAL-DEBUGGER #<READER-ERROR {AA34AF1}> #<unavailable
argument>)
2: (INVOKE-DEBUGGER #<READER-ERROR {AA34AF1}>)
3: (ERROR READER-ERROR)
4: (SB-IMPL::%READER-ERROR
    #<TWO-WAY-STREAM
      :INPUT-STREAM #<SB-SYS:FD-STREAM for "descriptor 14" {A755EC9}>
      :OUTPUT-STREAM #<SB-SYS:FD-STREAM for "descriptor 13" {A755541}>>
    "illegal terminating character after a colon: ~S")
5: (SB-IMPL::READ-TOKEN
    #<TWO-WAY-STREAM
      :INPUT-STREAM #<SB-SYS:FD-STREAM for "descriptor 14" {A755EC9}>
      :OUTPUT-STREAM #<SB-SYS:FD-STREAM for "descriptor 13" {A755541}>>
    #:)
6: (READ-PRESERVING-WHITESPACE
    #<TWO-WAY-STREAM
      :INPUT-STREAM #<SB-SYS:FD-STREAM for "descriptor 14" {A755EC9}>
      :OUTPUT-STREAM #<SB-SYS:FD-STREAM for "descriptor 13" {A755541}>>
    NIL
    NIL
    T)
7: (READ-PRESERVING-WHITESPACE
    #<TWO-WAY-STREAM
      :INPUT-STREAM #<SB-SYS:FD-STREAM for "descriptor 14" {A755EC9}>
      :OUTPUT-STREAM #<SB-SYS:FD-STREAM for "descriptor 13" {A755541}>>
    NIL
    NIL
    NIL)
8: (LTK::MAIN-ITERATION :BLOCKING T)
9: ((LAMBDA ()))
10: ((LABELS LTK::USE-DEBUGGER)
     #<FUNCTION LTK::TRIVIAL-DEBUGGER>
     #<CLOSURE (LAMBDA #) {AA3477D}>)
11: (LTK:MAINLOOP :SERVE-EVENT NIL)
I'm running SBCL 0.9.11 on Linux, if it matters.  My Tk version is
8.4.
...
No, but do compare it to 'ltktest-cells-inside to get a feel for how
Cells transforms development.
I definitely will.  I'd very much like to do more declarative
programming where it makes sense, and (as you and others have pointed
out) UIs are a natural fit.  So it will be good to compare your
approach to the more usual (less intelligent) approach.
--
John Landahl john@landahl.org
http://landahl.org/john