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

Conference azur::mcc

Title:	DECmcc user notes file. Does not replace IPMT.
Notice:	Use IPMT for problems. Newsletter location in note 6187
Moderator:	TAEC::BEROUD

Created:	Mon Aug 21 1989
Last Modified:	Wed Jun 04 1997
Last Successful Update:	Fri Jun 06 1997
Number of topics:	6497
Total number of notes:	27359

1729.0. "FCL bug: default for ENUMERATION datatype" by SWORD1::ES (Eugene Shvartsman) Mon Oct 28 1991 16:39

First, let me say that I do know about the FCL problem with the passing the
default values depending on the order of the arguments (notes 1112, 1185 1461).
We have this problem too.

But here is the new one.

If one of the request argument declared as having a default with a datatype of
the enumeration:

    ...

    TYPE DisplayType = 3001 (
		Barebone = 1,
		Brief = 2,
		Medium = 3,
		Full = 4
	);

     ...
	    REQUEST
     ...
		ARGUMENT output type = 3 : DisplayType
		    ECHO = TRUE,
		    DISPLAY = TRUE,
		    REQUIRED = FALSE,
		    DEFAULT = Barebone,
		    SYMBOL = OUTPUT_TYPE
		END ARGUMENT output type;
     ...
	    REQUEST

then if it is specified the last on the input line:

	MCC> trace node4 diehrd to_node = foo, output type

it will be encoded into ILV by FCL PM as

ILV dump of IN_P:

[  0 ] ( 
    [  1 ]     66 6f 6f  -- foo
    [  3 ] ( 
        [  5 ] 
        )
    )

But if there is one or more argument after it:

	MCC> trace node4 diehrd output type , to_node = foo
	%MCC-E-ENUMERATION_ERR, no such Enumeration value

i.e. FCL PM not able to handle it.

Best regards,
Gene

T.R	Title	User	Personal Name	Date	Lines
1729.1	I: Argument to_node probably your problem	TOOK::CALLANDER	MCC = My Constant Companion	`Sat Nov 23 1991 00:03`	12
	The problem is with the argument name. "TO" is reserved for the qualifier. We really shouldn't have a problem with your argument but... there is a problem with underscore handling in the fcl. I believe that if you rename your argument "to_node" t0 "tonode" or something else that your problem goes away. Mind you that this is just a guess, but I believe it isthe problem you are experienceing. sorry.
1729.2	No, it is not the name	SWORD1::ES	Eugene Shvartsman	`Mon Nov 25 1991 16:37`	11
	Sorry, Jill, but your guess is not right. I have noticed this problem with "TO" qualifier, that the reason why I have called my argument "TO_NODE". To make it completely clear I have renamed my argument to "FOONODE". Result is the same: MCC> trace node4 diehrd output type , foonode = foo %MCC-E-ENUMERATION_ERR, no such Enumeration value Gene
1729.3	another idea on enumeration error	TOOK::CALLANDER	MCC = My Constant Companion	`Wed Jan 08 1992 13:07`	3
	could you try it (or did you find the error) with an equals after "output type", this forces the parser to acknowledge that there is a value required.
1729.4	???	SWORD1::ES	Eugene Shvartsman	`Thu Jan 09 1992 16:12`	29
	Jill, What error "(or did you find the error)" do I suppose to find? Sure, if I type MCC> trace node4 diehrd output type =, to_node = foo I will have the following message: %MCC-W-ARGVALUEMISS, the argument value is missing So what? What does it prove? Seems to me that we have clear FCL bug, not critical, but a bug. Say, if we compare FCL with DCL we have an optional position independent qualifier with the default value, for example /LIST for many compilers. And in DCL you can say cc /noopt /list x.c and cc /list /noopt x.c without any problem, but in FCL you cannot do that. Gene
1729.5		BLUMON::SYLOR	Architect = Buzzword Generator	`Mon Jan 13 1992 12:27`	6
	Strictly speaking, MCC> trace node4 diehrd output type =, to_node = foo is not a syntax error. As defined in the NCL spec, that means to use the default value for output type.
1729.6	one more bug?	SWORD1::ES	Eugene Shvartsman	`Tue Jan 14 1992 02:31`	14
	> Strictly speaking, > >MCC> trace node4 diehrd output type =, to_node = foo > >is not a syntax error. As defined in the NCL spec, that means to use the default >value for output type. Sounds as one more bug for me, because in this case instead of passing default flag in ILV to my module, PM issues the following error: %MCC-W-ARGVALUEMISS, the argument value is missing Regards, Gene
1729.7	enumeration defaults broken	TOOK::CALLANDER	MCC = My Constant Companion	`Wed Jan 15 1992 19:56`	18
	regarding .4, thanks. It might have seemed like a useless test to you, but the difference in error messages confirmed that the parse tables do NOT see the default value for the output type argument. This is usually due to a dictionary update but no parse table update. Before jumping too far to conclusions I will verify that default enumeration values are working correctly. .....<as time passes>.... yup you win the grand prize, the enumeration defaults are NOT working in the FCL. And your prize is a qar number, 2142. jill (I tested it using the create domain rule directive and the auto enable argument)
1729.8	thanks for the prize	SWORD1::ES	Eugene Shvartsman	`Fri Jan 17 1992 22:39`	1