Re: [Axiom-developer] [build-improvements] more character stuff

[Gabriel Dos Reis]

On Sun, 13 May 2007, Waldek Hebisch wrote:

| > 
| > Hi,
| > 
| >   The previous patch that used the ANSI Lisp character type, in lieu
| > of the non-standard but cltl conformant string-char, misses a few
| > places.  Fixed thusly.
| > 
| 
| Gaby, it looks like you are re-doing ANSI stuff from scratch.

Well, as you know, I've been working on and off on support for SBCL and 
CLISP (ECL is secondary).  So, it is more in that line...

| Did you notice that string-char issues were already handled
| in Jurgen Weiss version?

No, I did not.

|  Greg Vanuxem sbcl port picked them.

OK.

| In private tree I have merged sbcl changes into wh-sandbox.
| There are some questionable parts in the code (notably using
| eq on functions) and some tests still fail, so I did not
| commit the code, but I can make it available.

Well, if you did part of the work, we can always debate the
questionable part.  Here is my work flow:

  (1) identify troublesome parts
  (2) identify the root causes
  (3) implement minimal changes to solve the issue
  (4) discuss the fundamental problem and the bigger changes
      later.

| BTW. Jurgen Weiss made choce between character and string-char
| conditional.  It is not clear for me if GCL supports ANSI
| enough to work well using character -- we probably should ask
| Camm if there are any issues making string-char preferable
| for GCL.

That sounds reasonable.

-- Gaby