Re: [cc65] Optimizer bug found

From: Groepaz <groepaz1gmx.net>
Date: 2009-09-11 01:30:07
On Donnerstag 10 September 2009, Oliver Schmidt wrote:
> Hi Uz,
>
> After spending a "little" time I finally succeeded in isolating an
> (maybe THE) optimizer bug causing Contiki to fail.
>
> From my user perspective it is very interesting that the bug shows up
> with compiler option -O but does _not_ show up with -Oi.
>
> @Groepaz: Maybe your tests should be repeated with different optimizer
> settings...

yes true, i will do that :)

a quick test shows that your testcase works if a) no optimizations are enabled 
or b) any combination of optimize flags that contain "i" are enabled. in all 
other cases it fails.

and about the damned yacc.c program: WARNING! :) it contains hardcoded 10's 
for checking line-endings, so it wont work when eg compiled for c64 and then 
using a proper petscii file as input. (and it might have worked before, only 
because i didnt convert the input files to proper petscii =P) i am trying to 
change such things right now to make the program portable in that respect ... 
maybe someone can try to compile it for apple2 (that has ascii like char 
mapping right?) and check it there. (to make sure charset is actually the 
problem here, and not some f****d up bug in my simulator that for some evil 
reason only shows up in the yacc programm too =P)

whats interisting though, when i run it in my cpu simulator thingy (i am 
trying to get of vice completely so everything can really run without 
interaction) which doesnt use any special character encoding, then yacc.c 
appears to work (!) when no optimizations are enabled. it fails in all other 
cases. (this didnt show up in previous overview pages because i always run 
certain tests with vice too to make sure my simulator doesnt produce bogus 
results)

-- 

http://www.hitmen-console.org    http://magicdisk.untergrund.net
http://www.pokefinder.org        http://ftp.pokefinder.org

Every adult needs a child to teach; it's the way adults learn. 
<Frank A. Clark>

----------------------------------------------------------------------
To unsubscribe from the list send mail to majordomo@musoftware.de with
the string "unsubscribe cc65" in the body(!) of the mail.
Received on Fri Sep 11 01:30:14 2009

This archive was generated by hypermail 2.1.8 : 2009-09-11 01:30:16 CEST