[Info-vax] Long uptime cut short by Hurricane Sandy

[VAXman- at SendSpamHere.ORG]

In article <amvuioF8tilU2 at mid.individual.net>, billg999 at cs.uofs.edu (Bill Gunshannon) writes:
>In article <keeb7i$e15$1 at dont-email.me>,
>	Stephen Hoffman <seaohveh at hoffmanlabs.invalid> writes:
>> On 2013-01-31 17:38:45 +0000,   VAXman-  @SendSpamHere.ORG said:
>> 
>>> In article <kee7oh$m45$1 at dont-email.me>, Stephen Hoffman 
>>> <seaohveh at hoffmanlabs.invalid> writes:
>>>> On 2013-01-31 16:32:44 +0000,   VAXman-  @SendSpamHere.ORG said:
>>>> 
>>>>> Or, skip the myriad excursions through LIB$CVT_TIME and LIB$CVT_DTIME, and
>>>>> let DCL do one of the things it does best...
>>>>> 
>>>>> $ filename = filenameprefix+"_"+F$cvtime(,"COMPARISON")-"-"-"-"-"
>>>>> "-":"-":"-"."+".dmp"
>>>> 
>>>> Yes.  That's definitely better than what I'd posted[1].
>>>> 
>>>> But you're not going to convince me that DCL is superior to bash here;
>>>> that printf-style notation is just plain handy.
>>> 
>>> I never said that.  I was trying to get the focus on the scripting lingo's
>>> features and not all of the utilities it can invoke.
>> 
>> 
>> AFAIK, no command line interpreter and no shell is ever intended to be 
>> used in isolation.
>> 
>> The whole point of a CLI or of a shell is allow scripting and to allow 
>> gluing together various executables.
>> 
>> IMO, Unix just happens to be better at that gluing-together than is 
>> VMS, albeit with the Unix shells typically being more cryptic.  
>
>Forgetting everyone here's dislike for the old Unix naming convention
>as utility programs (like sed, awk or ls) are not part of the shell,
>just what is cryptic about any Unix shell?
>
>>                                                                 That's 
>> as much due to the general philospophy as anything else.  Unix 
>> classically does many small bits stitched together, where VMS does big 
>> and all-encompassing monolithic solutions.  Now is Unix perfect here?  
>> Nope.  There's no metadata around what data is shoved through the pipes 
>> (such as character encodings, for instance), so tossing certain sorts 
>> of data through a pipe can get... "interesting".  But then this is also 
>> the sort of solution that's seldom implemented on VMS.
>> 
>> 
>>> 
>>>> [1] Pedantic: what you've posted does not produce exactly same output.
>>>> It's very close, but not the same.
>>> 
>>> $ filename = 
>>> filenameprefix+"_"+F$fao("!19AS",F$cvtime(,"COMPARISON"))-"-"-"-"-" 
>>> "-":"-":"+".dmp"
>> 
>> Yeah.  Or add the prefix and the underscore into the f$fao, as that 
>> avoids clobbering caracters in the prefix.
>
>And tell me how the above is not cryptic?  Just what is the difference
>between the first dash and the second?  Or the third?  And would anyone
>normal determine that?  :-)

In DCL, one can delete a character from a string with - (subtract) and one
can add strings with a "+" (just like in Hoff's bash example!).  I can, for
example, remove the "$" in this string: "This is a $ sign." with -"$".  In
DCL, the subtract character will delete the very first occurrance of said 
character that's specified when encountered in the string.  If no character
is found, nothing is done.  HB pointed out that using the ()s would define
the scope better but I assumed no "funny" characters would be in the file
prefix.  The multiple -"-" -"-" remove the dashes in the time and the -":"
-":" remove the colons in the time.  Surely, you could see that if you'd
have tried to understand it instead of choosing to denigrate DCL.

A lack of understanding of the semantics doesn't make it cryptic.  And FWIW,
I can hack a bash script too!  I'm responding to this thread from one of my
Linux laptops!  I've spent the better part of my week trying to restore all
of my environment, customizations and scripts after updating another Linux
laptop.  I don't understand why the Linux folks insist on doing that but it
has been their modus operandi.

-- 
VAXman- A Bored Certified VMS Kernel Mode Hacker    VAXman(at)TMESIS(dot)ORG

Well I speak to machines with the voice of humanity.