On Fri, Sep 27, 2013 at 12:12 AM, Glenn Fowler <[email protected]> wrote:
>
> On Thu, 26 Sep 2013 23:52:35 +0200 Roland Mainz wrote:
>> On Thu, Sep 26, 2013 at 11:43 PM, Roland Mainz <[email protected]> 
>> wrote:
>> > On Thu, Sep 26, 2013 at 11:30 PM, Glenn Fowler <[email protected]> 
>> > wrote:
>> >> On Thu, 26 Sep 2013 23:07:14 +0200 Roland Mainz wrote:
>> >>> On Thu, Sep 26, 2013 at 7:19 AM, Glenn Fowler <[email protected]> 
>> >>> wrote:
>> >>> >
>> >>> > the AT&T Software Technology ast alpha 2013-09-26 source release
>> >>> > has been posted to the download site
>> >>> >         http://www.research.att.com/sw/download/alpha/
>> >>> > the package names and md5 checksums are
>> >>> >             INIT  95da0c907a6ed25a601c373e16a1862e
>> >>> >         ast-base  0aed6c2cd6bccf0ec23120179a1dfc38
>> >>> >         ast-open  529b4c33049de4100bc6f5ecaafa7722
>> >>> >          ast-ksh  d0762a194fe508641fa0706294236d3f
>> >>> > the md5 sums should match the ones listed on the download page
>> >>> >
>> >>> > note that these are binary md5sums :)
>> >>> [snip]
>> >>
>> >>> Erm... there are a couple of problems:
>> >>> 1. ast-ksh doesn't build on Solaris 11/AMD64/64bit:
>> >>> -- snip --
>> >>> + cc -KPIC -D_TS_ERRNO -D_REENTRANT -KPIC -g -xs -xspace -Xa
>> >>> -xstrconst -z combreloc -xildoff -xcsi '-errtags=yes'
>> >>> '-DSHOPT_CMDLIB_BLTIN=0' '-DSH_CMDLIB_DIR="/usr/ast/bin"' '-DSHO
>> >>> PT_CMDLIB_HDR="/home/test001/ksh93/ast_ksh_20130926/build_i386_64bit_debug/tmp_solaris_builtin_header.h"'
>> >>> -DSHOPT_SYSRC '-D_map_libc=1' '-D_lib_socket=1' -lsocket -lnsl -I.
>> >>> -I/hom
>> >>> e/test001/ksh93/ast_ksh_20130926/build_i386_64bit_debug/src/lib/libsum
>> >>> -I/home/test001/ksh93/ast_ksh_20130926/build_i386_64bit_debug/arch/sol11.i386-64/include/ast
>> >>> -D_PACKAGE_ast
>> >>> -c 
>> >>> /home/test001/ksh93/ast_ksh_20130926/build_i386_64bit_debug/src/lib/libsum/sumlib.c
>> >>> "/home/test001/ksh93/ast_ksh_20130926/build_i386_64bit_debug/src/lib/libsum/sumlib.c",
>> >>> line 198: undefined symbol: sha384_scale
>> >>> "/home/test001/ksh93/ast_ksh_20130926/build_i386_64bit_debug/src/lib/libsum/sumlib.c",
>> >>> line 198: non-constant initializer: op "NAME"
>> >>> "/home/test001/ksh93/ast_ksh_20130926/build_i386_64bit_debug/src/lib/libsum/sumlib.c",
>> >>> line 267: warning: statement not reached (E_STATEMENT_NOT_REACHED)
>> >>> cc: acomp failed for
>> >>> /home/test001/ksh93/ast_ksh_20130926/build_i386_64bit_debug/src/lib/libsum/sumlib.c
>> >>> mamake [lib/libsum]: *** exit code 2 making sumlib.o
>> >>> -- snip --
>> > [snip]
>> >>
>> >> patches not forgotten
>> >> just trying to get to a stable state
>> >
>> > Oh please... I hereby swear to eat 2 inches of a broomstick if the
>> > getconf table changes cause any trouble... :-)
>> >
>> >> as you can see we're not there
>> >
>> > Grumpf... I'd prefer the "stuff all patches in"-approach as it gives
>> > me the opportunity to fix more stuf... :-)
>> >
>> >> sum-sha2.c has
>> >>         #define sha384_scale    0
>> >> sumlib.c has
>> >>         #include "sum-sha2.c"
>> >>         #undef  METHOD
>> >>         #define METHOD(x)       
>> >> x##_match,x##_description,x##_options,x##_open,x##_init,x##_block,x##_data,x##_print,x##_done,x##_scale,x##_flags
>> >>         METHOD(sha384),
>> >>
>> >> you'll have to track down why for that particular cc and options
>> >>         x##_scale => sha384_scale doesn't expand to 0
>> >
>> > Erm... because it uses the libmd codepath ?
>
>> This patch should fix the issue:
>> -- snip --
>> diff -r -u original/src/lib/libsum/sum-lmd.c
>> build_i386_64bit_debug/src/lib/libsum/sum-lmd.c
>> --- src/lib/libsum/sum-lmd.c   2013-09-25 16:48:46.000000000 +0200
>> +++ src/lib/libsum/sum-lmd.c     2013-09-26 23:48:14.224181544 +0200
>> @@ -266,6 +266,7 @@
>>  #define sha384_description "FIPS 180-2 SHA384 secure hash algorithm.
>> The block count is not printed."
>>  #define sha384_options "[+(version)?sha384 (solaris -lmd) 2005-07-26]"
>>  #define sha384_match   "sha384|sha-384|SHA384|SHA-384"
>> +#define sha384_scale 0
>>  #define sha384_flags   SUM_INDICATOR
>>  #define sha384_init    lmd_init
>>  #define sha384_block   lmd_block
>> -- snip --
>
> that was my edit error during the SUM_INDICATOR edit -- thx
>
>> Still missing are the single command wrappers to get the digest part
>> of the utilities in sync with busybox and GNU coreutils, e.g. we need:
>> -- snip --
>> sha1sum(1)
>> sha2sum(1)
>> sha128sum(1)
>> sha224sum(1)
>> sha256sum(1)
>> sha384sum(1)
>> sha512sum(1)
>> -- snip --
>
> src/lib/libcmd/shasum.c was added in the last alpha with e.g.
>
> int
> b_sha512sum(int argc, register char** argv, Shbltin_t* context)
> {
>         return b_cksum(argc, argv, context);
> }
>
> (except b_sha2sum() was omitted by error)
> are you asking for something else

Erm... I'm herely officially declared "dumb" (in this case) ... I
forgot that our list of path bindings on Solaris is static to match
exactly what is allowed by Solaris PSARC rules (e.g. PSARC/2006/550
and follow-ups) ... which includes all stuff in /usr/ast/bin/&&co. ...
I just have to update that list. /usr/include/ast/cmdext.h contains
all matching |b_*()| functions except |b_sha2()| ...

... the only missing bit is sha224sum(1) (AFAIK originally conceived
as method to have maximum security allowed by US export laws; sha256
was too strong so it seems they went right for the next lower possible
digest length... ;-/ ) for which libsum&&libcmd do not have any code
right now...

----

Bye,
Roland

-- 
  __ .  . __
 (o.\ \/ /.o) [email protected]
  \__\/\/__/  MPEG specialist, C&&JAVA&&Sun&&Unix programmer
  /O /==\ O\  TEL +49 641 3992797
 (;O/ \/ \O;)
_______________________________________________
ast-developers mailing list
[email protected]
http://lists.research.att.com/mailman/listinfo/ast-developers

Reply via email to