Search
17:13:36
jackdaniel
so this patch makes it work on arm?
17:13:44
jackdaniel
maybe make a PR for iolib?
17:57:21
Posterdati
jackdaniel: I tested it on mipsel
17:57:55
Posterdati
but the error was the same on arm
18:13:47
pjb
AFAIK the coerce should work, (it does in the REPL, doesn't it?) So it looks like a bug in ECL, not in iolib!
18:39:15
jackdaniel
in that case please fill the report! :)
18:39:27
jackdaniel
preferably with a fix, but it's not necessary
18:40:13
Posterdati
pjb: the coerce works in the ecl repl as well as the functions
18:41:33
Posterdati
pjb: I think it is confused by the macro reader #+ inside it
19:03:00
Posterdati
I do not see mips in *features* mmmh
19:08:34
Posterdati
jackdaniel: did you see my bug report?
19:25:48
jackdaniel
Posterdati: yes I did
19:25:52
jackdaniel
mipsel is mips feature
19:26:58
Posterdati
I saw in swank-loader.lisp
19:27:16
jackdaniel
Posterdati: could you add also a snippet, which reproduces the error without loading iolib?
19:27:21
jackdaniel
I think it was something with coerce
19:27:56
Posterdati
there is an error outside loading iolib
19:28:22
Posterdati
only if quickload iolib
19:29:03
jackdaniel
could you try to construct a minimal test case (code) which doesn't requier loading iolib? it may requier calling compile
19:29:11
Posterdati
if you take the function definition and copy in the repl it works!
19:29:13
jackdaniel
for instance (defun xxx () (some-bad-operation)
19:29:27
jackdaniel
Posterdati: repl bytecompiles things, try to call (compile 'fun)
19:31:23
jackdaniel
bytecompiler and native compiler are not entirely consistent (what is a problem by itself)
19:37:16
Posterdati
is the reader macro inside the coerce!
19:47:16
Posterdati
I compiled a very similar function, no error
19:52:18
Posterdati
no problem whatsoever
20:01:05
Posterdati
could be a memory related problem?
20:05:04
Posterdati
or something done to the reader?
20:38:20
jackdaniel
also, I don't own mips machine