Martin Gasbichler <gasbichl@informatik.uni-tuebingen.de> writes:
> I put a beta version of 0.6.3 at
>
> ftp://ftp.scsh.net/pub/scsh/0.6/scsh-0.6.3.tgz
Thanks! I tested a few script of mine, including my embryonic IMAP
client library, which contains its share of dynamic regexps. It
appears to work. I have another problem, though, with the AWK macro,
as described below. I also have some general comments:
- The following script used to work without errors with 0.5.3. With
0.6.3, it invariably outputs an error when reaching the end of file:
(define (main cmd-line)
(awk (read-line) (line) ()
(#t (format #t "line: ~a\n" line))))
Here is a session showing the problem:
----------------------------------------------------------------------
lamppc3% ~linuxsoft/apps/scsh-0.5.3/bin/scsh -e main -s test.scm < test.scm
line: (define (main cmd-line)
line: (awk (read-line) (line) ()
line: (#t (format #t "line: ~a\n" line))))
lamppc3% ~linuxsoft/apps/scsh-0.6.3/bin/scsh -e main -s test.scm < test.scm
line: (define (main cmd-line)
line: (awk (read-line) (line) ()
line: (#t (format #t "line: ~a\n" line))))
Error: returning zero values when one is expected
(values)
lamppc3%
----------------------------------------------------------------------
- I cannot find the documentation about the new MD5 functions in the
scsh manual, but maybe that's just because it's not written yet.
- I found out that the structures containing srfi-1, srfi-13 and
srfi-14 (the list, character and string libraries) are not
automatically opened anymore. While I can understand this decision
from an aesthetic point of view, it breaks pretty much all of my
scripts, and I guess this alone can give a bad impression of 0.6.3.
Wouldn't it be possible to open them automatically, and make a big
note in the documentation saying that they will be closed by default
for the next release, or something like that?
- In the PDF version of the manual, the table of contents and the
index do not appear in the bookmarks on the left, which makes it
hard to jump to either of them.
- Wouldn't it be possible to fix let-match (and if-match) even without
Olin's input (see bug #441953)? With the current situation, I do not
want to use either of them since their behaviour doesn't correspond
to the documentation, and I do not know which one will change in the
future. Couldn't we say that the code is wrong, and change it to
match the documentation? (The same is true of bug #489901 BTW).
- I found out that for some structures, s48 asks whether one really
wants to open it, while for others it opens them immediately.
Example:
----------------------------------------------------------------------
> ,open srfi-1 srfi-2 srfi-13
Load structure srfi-2 (y/n)? y
[srfi-2 /home/linuxsoft/apps/scsh-0.6.3/lib/scsh/srfi/srfi-2.scm]
----------------------------------------------------------------------
Why does it ask for srfi-2 and not for the others (this has always
been the case, IIRC, but I'm just asking the question now)?
Apart from that, thanks a lot for this future relase!
Michel.
|