On Fri, May 4, 2012 at 11:46 PM, Glenn Fowler <[email protected]> wrote:
>
> the AT&T Software Technology ast beta 2012-05-04 source release
> has been posted to the download site
>        http://www.research.att.com/sw/download/beta/
>
> the git source repository has also been updated
>        http://www.research.att.com/sw/gitweb/
>
> the recent git web problems should be fixed

Thanks for beta... :-)
... quick test on SuSE Linux 12.1/AMD64/gcc4.6.2 with _and_ without
-O2 (which means optimiser settings are not relevant) triggers the
following testsuite failures:
-- snip --
test attributes begins at 2012-05-05+01:02:40
./src/cmd/ksh93/tests/attributes.sh: line 132: 11704: Memory fault
        attributes.sh[133]: typeset -L should not be inherited
test attributes failed at 2012-05-05+01:02:42 with exit code 1 [ 106
tests 1 error ]
test attributes(shcomp) begins at 2012-05-05+01:02:42
/tmp/tmp_3v.d89/shcomp-attributes.ksh: line 132: 11756: Memory fault
        shcomp-attributes.ksh[133]: typeset -L should not be inherited
test attributes(shcomp) failed at 2012-05-05+01:02:43 with exit code 1
[ 106 tests 1 error ]
-- snip --

The core dump has the stacktrace:
-- snip --
(gdb) where
#0  0x00007f4064674426 in __strcmp_sse42 () from /lib64/libc.so.6
#1  0x00007f406598e3f4 in nv_compare (dict=0x7f4065f495c0, sp=0x0,
dp=0x7f4065f24568, disc=0x7f4065d50c80)
    at 
/home/test001/work/ast_ksh_20120504/build_linux/src/cmd/ksh93/sh/nvdisc.c:37
#2  0x00007f4065a949f5 in dttree (dt=0x7f4065f495c0, obj=0x7f4065f24490, type=8)
    at 
/home/test001/work/ast_ksh_20120504/build_linux/src/lib/libast/cdt/dttree.c:447
#3  0x00007f4065a95b10 in dtwalk (dt=0x7f4065f495c0,
userf=0x7f40659d8534 <scanfilter>, data=0x7fff5cf9a300)
    at 
/home/test001/work/ast_ksh_20120504/build_linux/src/lib/libast/cdt/dtwalk.c:46
#4  0x00007f40659d8807 in nv_scan (root=0x7f4065f495c0,
fn=0x7f40659d8941 <sh_envnolocal>, data=0x7fff5cf9a350, mask=8192,
flags=0)
    at 
/home/test001/work/ast_ksh_20120504/build_linux/src/cmd/ksh93/sh/name.c:2312
#5  0x00007f40659b2105 in sh_reinit (argv=0x0) at
/home/test001/work/ast_ksh_20120504/build_linux/src/cmd/ksh93/sh/init.c:1630
#6  0x00007f406598c24d in sh_main (ac=2, av=0x7fff5cf9a5e8,
userinit=0) at 
/home/test001/work/ast_ksh_20120504/build_linux/src/cmd/ksh93/sh/main.c:168
#7  0x0000000000400699 in main (argc=2, argv=0x7fff5cf9a5e8) at
/home/test001/work/ast_ksh_20120504/build_linux/src/cmd/ksh93/sh/pmain.c:45
-- snip --
Problem seems to be that |sp| is |NULL|... which triggers the code
dump (Linux has no "coreadm" ... ;-( )...


Next crash is this one:
-- snip --
test functions begins at 2012-05-05+01:04:30
/home/test001/work/ast_ksh_20120504/build_linux/arch/linux.i386/src/cmd/ksh93/ksh[10]:
fun: line 8: 15212: Memory fault
./src/cmd/ksh93/tests/functions.sh: line 839: 15209: Memory fault
        functions.sh[849]: compound variable cleanup before script exec failed
test functions failed at 2012-05-05+01:04:33 with exit code 1 [ 105
tests 1 error ]
test functions(shcomp) begins at 2012-05-05+01:04:33
/home/test001/work/ast_ksh_20120504/build_linux/arch/linux.i386/src/cmd/ksh93/ksh[10]:
fun: line 8: 15304: Memory fault
/tmp/tmp_3t.aq5/shcomp-functions.ksh: line 834: 15301: Memory fault
        shcomp-functions.ksh[844]: compound variable cleanup before
script exec failed
test functions(shcomp) failed at 2012-05-05+01:04:35 with exit code 1
[ 105 tests 1 error ]
-- snip --
No coredump is produced... I have to dig later why this happens...


Next test failure is this one:
-- snip --
test io begins at 2012-05-05+01:04:38
        io.sh[128]: multiple exec 4< /dev/null can fail
test io failed at 2012-05-05+01:04:50 with exit code 1 [ 90 tests 1 error ]
test io(shcomp) begins at 2012-05-05+01:04:50
        shcomp-io.ksh[128]: multiple exec 4< /dev/null can fail
test io(shcomp) failed at 2012-05-05+01:05:02 with exit code 1 [ 90
tests 1 error ]
-- snip --

The _really_ frustrating part is that if I run the tests manually with
"shtests" I get no failures. Same happens if I use a "ksh" build with
libshell&&libcmd&&libdll&&libsum&&libast linked statically.


Finally... there are two new warnings:
-- snip --
/home/test001/work/ast_ksh_20120504/build_linux/src/cmd/ksh93/bltins/misc.c:
In function 'b_dot_cmd':
/home/test001/work/ast_ksh_20120504/build_linux/src/cmd/ksh93/bltins/misc.c:305:3:
warning: passing argument 2 of 'sh_argreset' discards 'volatile'
qualifier from pointer target type [enabled by default]
/home/test001/work/ast_ksh_20120504/build_linux/src/cmd/ksh93/include/defs.h:395:15:
note: expected 'struct dolnod *' but argument is of type 'volatile
struct dolnod *'
-- snip --
Someone please correct me but this can be solved by a simple cast in
the style of...
-- snip --
--- ./bltins/misc.c.original    2012-05-05 02:28:17.743880283 +0200
+++ ./bltins/misc.c     2012-05-05 02:28:36.661049953 +0200
@@ -302,7 +302,7 @@
                free((void*)shp->st.filename);
        shp->dot_depth--;
        if((np || argv[1]) && jmpval!=SH_JMPSCRIPT)
-               sh_argreset(shp,argsave,saveargfor);
+               sh_argreset(shp,(struct dolnod*)argsave,saveargfor);
        else
        {
                prevscope->dolc = shp->st.dolc;
-- snip --
... because no |longjmp()| is done and we don't do something like
|siglongjmp()|, right (dumb question... I know... but it's 2:30h AM
here) ?

----

Bye,
Roland

-- 
  __ .  . __
 (o.\ \/ /.o) [email protected]
  \__\/\/__/  MPEG specialist, C&&JAVA&&Sun&&Unix programmer
  /O /==\ O\  TEL +49 641 3992797
 (;O/ \/ \O;)

_______________________________________________
ast-users mailing list
[email protected]
https://mailman.research.att.com/mailman/listinfo/ast-users

Reply via email to