NetNews Usenet Archive 1992 #31/NN_1992_31.iso/spool/comp/sys/amiga/programm/17665

Internet Message Format | 1992-12-22 | 3.7 KB

Xref: sparky comp.sys.amiga.programmer:17665 comp.sys.amiga.misc:18919
Newsgroups: comp.sys.amiga.programmer,comp.sys.amiga.misc
Path: sparky!uunet!zaphod.mps.ohio-state.edu!sdd.hp.com!spool.mu.edu!agate!stanford.edu!rock!concert!sas!mozart.unx.sas.com!jamie
From: [e-mail] (James Cooper)
Subject: Re: SAS C6.1 datecmp bug [Was: Re: Advice to SAS/C 5.10b users]
Originator: [e-mail]
Sender: [e-mail] (Noter of Newsworthy Events)
Message-ID: <[e-mail]>
Date: Mon, 21 Dec 1992 19:30:46 GMT
References: <[e-mail]> <[e-mail]> <[e-mail]>
Nntp-Posting-Host: cdevil.unx.sas.com
Organization: SAS Institute Inc.
Lines: 59


In article <[e-mail]>, [e-mail] (Craig Fisher) writes:
>In article <@unx.sas.com> [e-mail] (Doug Walker) writes:
>> In article <>, [e-mail] (Sam Hulick) writes:
>> |> It would be rather safe to stick with 5.10b.  6.0 and 6.1 are too
>> |> buggy.. But you can probably see that from all the posts(complaints) on
>> |> c.s.a.p. :)
>>
>> I don't agree at all.  In anything the size of SAS/C, there are going
>> to be bugs.  I believe the majority of V6 customers are not experiencing
>> major problems - you just hear about the ones that are on the net
>> because they are reporting them.
>
>Seems to me thought that you could have used a few more beta testers.
>The number of bugs in this release doesn't really seem reasonable or
>excusable to me.

Let's see... counting everyone at Commodore (U.S.), plus the people here, plus
the regular poeple we recruited through other means, we had about 70 - 100
people testing the heck out of 6.0.  We all thought that was enough...

Doug was correct.  Out of all those people, things can still slip past... for
instance, one of the things fixed in 6.1 over 6.0 was not being able to compile
with the F4 key in SE.  This was finally traced down to having the system
WildStar bit turned on.  You might claim this is a common thing to do, but none
of the Beta testers uses that bit, so it wasn't found until after release.

It doesn't matter *HOW* many people you have testing something, there are going
to be bugs.  Period.  Oh, I suppose it is theoretically possible to come up with
a test suite which would test every single line of code for every possible case
it could run into, but we don't have man-centuries to wait for this test to run,
even if we had it.  And that's after the man-millenia to actually write that
test suite in the first place...

>> 6.1 fixes most of the problems reported in 6.0, except of course that
>> we broke the MATH=68881 option.  If you don't use that particular
>
>What about the problem with the datecmp() function? I reported this
>as broken via emits, but have not received any response (other than
>a receipt acknowledgement). Can you confirm that it is broken? Has it
>been fixed? Any idea when the next patch will be ready?

Yes, datecmp() is confirmed broken.  I can't get it to return anything other
than -1 myself.  I don't know why you haven't gotten a response via EMITS, but
I'm checking into that now.

As for why it was broken when you got it... well, I don't know any programmers
who use the ANSI functions unless they are porting code, and we've only had 3
people complain about the datecmp() function so far... and all of the complaints
were *after* 6.1 came out.  The function was there in 6.0, and hasn't been
touched since, but we never knew there was a problem, so we couldn't fix it.

--
---------------
Jim Cooper
([e-mail])                             bix: jcooper

Any opinions expressed herein are mine (Mine, all mine!  Ha, ha, ha!),
and not necessarily those of my employer.