Search
Saturday, 21st of July 2018, 2:39:17 UTC
5:48:19
beach
Good morning everyone!
8:43:34
frgo
A wonderful good Saturday morning all.
14:15:23
drmeister
MPS build works again
14:24:36
kpoeck
Drmeister why did you change in load.cc the outputs for :verbose and :print to BFORMAT_T (initially eval::funcall(cl::_sym_format ....)
14:26:02
kpoeck
BFORMAT_T does not seem to be redirected in (with-output-to-string ....)
14:26:50
drmeister
kpoeck: Right - BFORMAT_T is hard coded to print to *standard-output*.
14:27:27
drmeister
Does that cause a problem?
14:27:45
kpoeck
Can I change that back to eval::funcall, so that some ansi-tests that capture output from load work again?
14:28:03
kpoeck
Same problem in compile-file
14:29:06
drmeister
How would you normally pass the stream to load?
14:30:02
drmeister
I'm just waking up. I'm not all here yet.
14:30:53
kpoeck
(with-input-from-string (s "(defun foo())") (load s :print t :verbose t))
14:31:12
drmeister
Ok, so I changed from my old, weird _lisp->print(...) to BFORMAT_T because BFORMAT_T prints properly to the standard output.
14:31:13
kpoeck
should return the output from load, but currently it doesn't
14:31:55
kpoeck
If I change e.g. BFORMAT_T(BF(";;; Loading %s\n") % _rep_(lsource));
14:32:21
kpoeck
to eval::funcall(cl::_sym_format, _lisp->_true(), SimpleBaseString_O::make("~&;;; Loading ~s~%"), lsource);
14:32:34
kpoeck
load works again as expected
14:33:26
drmeister
Right - but that may have trouble or at least look weird in the interpreter or aclasp - because format isn't fully installed at that point.
14:34:18
drmeister
The interpreter provides a seriously crippled version of format and only bclasp contains the proper version implemented in Common Lisp.
14:34:45
kpoeck
I think I have seen a if (format-allready-works)(format ...)
14:36:14
drmeister
What about BFORMAT(_lisp->_true(),BF(";;; Loading %s%N") % _rep_(lsource) );
14:36:26
drmeister
Or do you need the ~& to work?
14:36:56
kpoeck
In compile-file there is describe-form using only write-string and terpri, is that bootstrap safe?
14:37:17
drmeister
What I mean to say is "yes", we can change it to eval::funcall(cl::_sym_format,...) to make it work properly.
14:37:43
drmeister
I'm just messing more with the interpreter and aclasp and it would be nice to have load print stuff properly in there as well.
Saturday, 21st of July 2018, 14:39:17 UTC