[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference vaxaxp::vmsnotes

Title:	VAX and Alpha VMS
Notice:	This is a new VMSnotes, please read note 2.1
Moderator:	VAXAXP::BERNARDO

Created:	Wed Jan 22 1997
Last Modified:	Fri Jun 06 1997
Last Successful Update:	Fri Jun 06 1997
Number of topics:	703
Total number of notes:	3722

226.0. "Subtle USGOTO error VAX only V6.2 & V7.0" by GIDDAY::GILLINGS (a crucible of informative mistakes) Wed Feb 19 1997 20:53

    A customer has submitted some DCL code which behaves differently on
    OpenVMS/VAX V6.2 from OpenVMS/Alpha. The distilled version is:
    
label_test.com
$ ThisSymbol="The symbol value"
$ GOTO TestLabel
$ ThisSymbol "test"
$ TestLabel:
$ EXIT
    
    
    On OpenVMS/Alpha and OpenVMS/VAX V6.1 and below this "works" as one
    might expect at first sight. On OpenVMS/VAX V6.2 and V7.0, an error
    is generated (verify on):
    
$ @label_test
$ ThisSymbol="The symbol value"
$ GOTO TestLabel
%DCL-W-USGOTO, target of GOTO not found - check spelling and presence of label
 \TESTLABEL\
    
    Now, if you change the RHS of the symbol assignment to anything NOT
    starting with "THE", the code "works" again. So, DCL must be seeing the
    symbol usage of "ThisSymbol" and interpreting it as a synonym for
    "THEN", hence the label is inside a structured IF block and is
    therefore not visible to the GOTO command. 
    
    The customer discovered this curiosity by accidently leaving out the
    "=" symbol in an assignment which was continued across 2 lines.
    
    Now, I find it difficult to call this as bug or feature. On the one
    hand you'd have to be completely crazy to define synonym symbols for
    "IF", "THEN", "ELSE" and/or "ENDIF" (although some non-native english
    speakers may disagree). On the other hand, the nature of an interpreted
    language would suggest that it is legal and should work.
    
    There must be a problem here though as VAX and Alpha behave
    differently. Any opinions on which one should be QARed?
    
    						John Gillings, Sydney CSC

T.R	Title	User	Personal Name	Date	Lines
226.1		AUSS::GARSON	DECcharity Program Office	`Wed Feb 19 1997 22:35`	28
	re .0 Er, cute. In fact symbols are not relevant. The following command procedure minimally demonstrates the difference between Alpha VMS V6.2 and VAX VMS V6.2. $ if 1 $ th $ endi Output from Alpha: %DCL-E-INVIFNEST, invalid IF-THEN-ELSE nesting structure or data inconsistency Output from VAX: none although if one has spell checking on then one gets %DCL-I-MISSPELL, Verb TH is misspelled, using THEN in file A.COM i.e. Alpha will not accept the THEN verb abbreviated unambiguously to TH or THE. Which is correct? Well, IMHO this is a pretty sordid area of DCL but the VAX is probably correct.
226.2	Caught by a bugfix	EVMS::DAVIDB::DMILLER	This bug fix broke what???????	`Thu Feb 20 1997 11:22`	10
	Yep. The bugfix was to accept abbreviations for the THEN and ELSE verbs just like they're allowed everywhere else. As noticed, you had a busted command procedure all along, and now its obvious. The fix should have been on both VAX and Alpha, though. Will have to check into that... -Dave
226.3	How do you spell check (again) please?	IOSG::PYE	Graham - ALL-IN-1 Sorcerer's Apprentice	`Fri Feb 21 1997 03:56`	6
	RE .1 Can you remind me how to turn on spell checking? Last time I saw how to do this, I was left a little confused! Graham (maintainer of 100k lines of DCL!!)
226.4	DCL spell checking	MARVIN::CARLINI		`Fri Feb 21 1997 15:13`	10
	Well it was 17 months ago when last you asked :-) $ SET COMMAND SYS$INPUT: DEFINE VERB EVIL CLIROUTINE SETDCL QUALIFIER SPELL_CHECK, NEGATABLE $! $ EVIL/SPELL Antonio
226.5	Warning!	COMEUP::SIMMONDS	lock (M); while (not *SOMETHING) { Wait(C,M); } unlock(M)	`Sun Feb 23 1997 00:05`	9
	Re: .4 .. there's some problem when SPELL_CHECK is turned ON on OpenVMS Alpha V6.2 and you start a command procedure.. 100% CPU Busy in a fairly tight loop involving DCL and emulated MOVC3 .. interrupts are ignored. Does anyone consider this SPELL_CHECK feet-chewer QAR-able? :) John.
226.6		AUSS::GARSON	DECcharity Program Office	`Sun Feb 23 1997 16:35`	9
	re .5 (Er, yes, my LOGIN.COM only turns on spell checking if VAX.) > Does anyone consider this SPELL_CHECK feet-chewer QAR-able? :) A quick look through the list of QARs entered by me suggests that even I drew the line at QARing an undocumented feature that you have to define a verb manually for in order to use. But don't let me stop you.
226.7	will QAR it soon..	COMEUP::SIMMONDS	lock (M); while (not *SOMETHING) { Wait(C,M); } unlock(M)	`Sun Feb 23 1997 18:38`	6
	Re: .6 Given that this (CPU consuming) behaviour could be exploited as a mild denial-of-service attack on OpenVMS Alpha, I'll write a QAR on it. John.