Re: MD5

Just sent out the announcement for v7.0 ... the tree is now tag'd for v7.0
as REL7_0 ...

I would like to do a version v7.0.1 followup release by June 1st, and
create a branch at that time so that work can begin on v7.1 ...

Marc G. Fournier ICQ#7615664 IRC Nick: Scrappy
Systems Administrator @ hub.org
primary: scrappy(at)hub(dot)org secondary: scrappy(at){freebsd|postgresql}.org

Vince Vielhaber wrote:
>
> The md5 sources I got from Sverre (think I got that name right) were ready
> out of the box. I redid the test program to take just a string and
> return the md5 hash. I tested it on HP 9 & 10 (the HP8 machine with the
> ansi compiler on it disappeared so I'll test it when it resurfaces), IRIX,
> FreeBSD 3.2 and DOS/95/98/2k/NT - compiles cleanly and runs on each. I'd
> like to see if it still compiles and runs on some other platforms. I have
> the tarball sitting at:
>
> http://www.pop4.net/~vev/mdstuff.tar.gz
>
> To build and test it, unpack it, run make then try it:
>
> $ test asdf

or better try

$ ./test asf

if you are on any unix platform

----------------
Hannu

The md5 sources I got from Sverre (think I got that name right) were ready
out of the box. I redid the test program to take just a string and
return the md5 hash. I tested it on HP 9 & 10 (the HP8 machine with the
ansi compiler on it disappeared so I'll test it when it resurfaces), IRIX,
FreeBSD 3.2 and DOS/95/98/2k/NT - compiles cleanly and runs on each. I'd
like to see if it still compiles and runs on some other platforms. I have
the tarball sitting at:

http://www.pop4.net/~vev/mdstuff.tar.gz

To build and test it, unpack it, run make then try it:

$ test asdf
912ec803b2ce49e4a541068d495ab570 asdf

An ansi compiler is required - do we support any non-ansi compilers? If
so I'll make some changes to the sources.

Anyone who tests, please send me the platform you tried it on and the
results. I'd like to find a long word list to make some comparison
tests with, but it doesn't look like I ever got around to installing
a dictionary or spellchecker on any of my machines!!

Vince.
--
==========================================================================
Vince Vielhaber -- KA8CSH email: vev(at)michvhf(dot)com http://www.pop4.net
128K ISDN from $22.00/mo - 56K Dialup from $16.00/mo at Pop4 Networking
Online Campground Directory http://www.camping-usa.com
Online Giftshop Superstore http://www.cloudninegifts.com
==========================================================================

Vince Vielhaber <vev(at)michvhf(dot)com> writes:
> An ansi compiler is required - do we support any non-ansi compilers? If
> so I'll make some changes to the sources.

We've required ANSI-style prototypes all along. There are some other
ANSI features we're willing to work around the lack of, like token
pasting in macros --- so the real question is *what* ANSI features
does the code require?

> I'd like to find a long word list to make some comparison
> tests with, but it doesn't look like I ever got around to installing
> a dictionary or spellchecker on any of my machines!!

Just feed it a bunch of files. MD5 is supposed to work on any length
input ...

regards, tom lane

Tom Lane writes:

> We've required ANSI-style prototypes all along. There are some other
> ANSI features we're willing to work around the lack of, like token
> pasting in macros ---

You must mean this stuff: :-)

[src/include/nodes/nodes.h]
| #define nodeTag(nodeptr) (((Node*)(nodeptr))->type)
|
| #define makeNode(_type_) ((_type_*) newNode(sizeof(_type_),T_##_type_))
| #define NodeSetTag(nodeptr,t) (((Node*)(nodeptr))->type = (t))
|
| #define IsA(nodeptr,_type_) (nodeTag(nodeptr) == T_##_type_)

There are some provisions for working around this, but evidently they're
not used uniformly (which leads me to believe that there's little need to
do so).

> so the real question is *what* ANSI features does the code require?

I think it's fair to assume an ANSI C89 compiler in the year 2000, based
on the fact that we don't actually seem to consciously avoid any
constructs I know of, modulo the Autoconf safety net.

--
Peter Eisentraut Sernanders väg 10:115
peter_e(at)gmx(dot)net 75262 Uppsala
http://yi.org/peter-e/ Sweden

Peter Eisentraut <peter_e(at)gmx(dot)net> writes:
> Tom Lane writes:
>> We've required ANSI-style prototypes all along. There are some other
>> ANSI features we're willing to work around the lack of, like token
>> pasting in macros ---

> You must mean this stuff: :-)

Hmm, good point. I guess that HAVE_STRINGIZE stuff in c.h is dead code
after all, at least on the platforms people have tried Postgres on.

> I think it's fair to assume an ANSI C89 compiler in the year 2000, based
> on the fact that we don't actually seem to consciously avoid any
> constructs I know of, modulo the Autoconf safety net.

This is really a consequence of our development process: since we accept
patches from people who haven't studied the source very carefully, the
level of adherence to coding standards is pretty variable. Portability
problems get flushed out when someone reports "it doesn't work here",
but there's no process for removing no-longer-needed portability hacks.
Looks like HAVE_STRINGIZE is in that category now...

regards, tom lane