Help please strugglng to get multiple options out of ReadArgs

[chocsplease]

Hi a/b,

Many thanks for yet another extremely helpful and useful post.

Regarding the *ptr=((char)((STRPTR*)clparams[0])[i][0]); line.

I was trying to get the pointer to point *at* the 1st character. I thought if I could I could then run through the string simply incrementing the pointer till I hit a /0. However as you have pointed out I stuffed up badly when I tried to find the length of it using strlen since I'd just got the 1st char and not a pointer to it.

Think I understand what I've just typed...

Thank you for the info regarding casting and the code snippet I've copied it into my library as I'm sure it'll be useful later on.

Luckily I don't have to do any of this as my version of gcc includes a non ANSI strcasecmp (as well as strncasecmp) so thanks for pointing out their existence - my C book makes no mention of them. Could you suggest a good current one?

Oh I'm using messages in the code so the user can localise the template if they wish - messages is defined as -

static char *messages[]={MSG_READWRITE,MSG_NODISK,MSG_BADDISK,MSG_KICKSTART,MSG_NOT_DOS,MSG_READONLY,MSG_VALIDATING,\
MSG_MOUNTED,MSG_VOL_HEAD,"Unit","DOS","Size","Used","Free","Full","Errors","Status","Name","Amiga Disk","PC Disk",MSG_ERROR_MEM,MSG_BREAK,\
	MSG_ERROR_LIB,TEMPLATE};

TEMPLATE is #defined as I mentioned in the earlier post. I guess I could #define the whole lot, as it looks messy as it is, not sure if there is a recommended way of doing this. Also not sure if I need the static. The pointers get moved to point to localised text in a malloc'ed array if the user has defined them in a preferences file.

Once again massive thanks for all the help.

[idrougge]

I can't remember ever seeing any localised command line arguments.

[chocsplease]

Originally Posted by idrougge

I can't remember ever seeing any localised command line arguments.

Oh. I didn't know that! Having never used an Amiga in any language other than English I had assumed that commands etc would be in the native language. When I got a set of blank disks from the folks at Vesalia they were labelled leer (German for empty).

I guess that Commodore must have hard coded everything then? Even after they introduced the localisation system in whatever os version it was.

So much of the old OS info has been replaced by OS 4 specific stuff on the net its very hard to find the info you need.

Anyway my app does offer this and provided it works and I can nail a stupid crash I'm fairly near finishing thanks to all the help I have had from folks on this forum.

[idrougge]

Indeed Workbench uses localised default names for formatted disks.

[nogginthenog]

Originally Posted by chocsplease

Oh. I didn't know that! Having never used an Amiga in any language other than English I had assumed that commands etc would be in the native language.

Wouldn't make any sense. Scripts etc would have been language dependant!

[chocsplease]

Originally Posted by nogginthenog

Wouldn't make any sense. Scripts etc would have been language dependant!

Hi,

To be honest I hadn't even thought about scripts! However I've fixed it by simply adding the switch; script. If this is included in the command line then the app uses the the default English arguments no-matter what the user has set the local command line to be.

Obviously if they change the name of the app itself then there is little I can do about it.

Thanks for pointing the possible issue out.

[chocsplease]

Oh well so much for *that* idea.

Having added SCRIPT/S to my template I am now getting a software failure when the app exits of #80000004. Its not a Guru, but one of the suspend/Reboot type.

The bizarre thing is that all the other switches work fine and the app exits correctly whether you enter anything on the command line or not. It only crashes if you enter SCRIPT.

Here's the template.

#define TEMPLATE "DEVICE/M,HELP/S,WINDOW/S,NOCOLOURS/S,COLOURS/S,LEADINGSPACE/S,TRAILINGSPACE/S,NORMALRAM/S,DECIMALSIZE/S,SCRIPT/S,DUD/S"

I added the DUD switch just in case it was something odd with SCRIPT being the last switch. Adding DUD to the command line also causes the same error!

Here's the code that reads the options. Initially I wasn't doing anything with the switches but then added the if checks with the zz=0; in them in pure desperation.

/*get args if any from command line*/


			Args=ReadArgs(TEMPLATE,clparams,NULL); /*check first on the default template for script switch*/				

		if (Args)	/*entered something*/
		{
		printf("entered args- clparams[0]=%d \n",clparams[0]);
			if (clparams[0]!= NULL)
			{
			printf("entered device\n");
				i=0;
				while (((STRPTR*)clparams[0])[i]!=NULL)
				{
					printf("%s\n",((STRPTR*)clparams[0])[i]);
					i++;
				}

			}
			
				
		printf("OPTIONS STARTS AS %d\n",options);	
/*			if ((clparams[3]==-1) && (clparams[4]==-1)) DO SOMETHING! <<<<<< cant have colours and no colours speced!*/
		
			if ((clparams[3]==-1) && ((options & NO_COLOURS) ==0)) options=options | NO_COLOURS;	/*if nocolours on cmd line and not in prefs file set*/
			
			if ((clparams[4]==-1) && ((options & NO_COLOURS)> 0)) options=options - NO_COLOURS;		/*if colours on cmd line and nocolours in prefs file reset */

			if ((clparams[5]==-1) && ((options & LEADING_SPACES) == 0)) options=options | LEADING_SPACES;	/*if want leading spaces and not in prefs file set*/
		
			if ((clparams[6]==-1) && ((options & TRAILING_SPACES) == 0)) options=options | TRAILING_SPACES;	/*if want trail scs and not in prefs fileset*/
		
			if ((clparams[7]==-1) && ((options & NORMAL_RAM) == 0)) options=options | NORMAL_RAM; 		/*if want ram disk always 100% and not in prefs set*/
		
			if ((clparams[8]==-1) && ((options & DEC_VALUES) == 0)) options=options | DEC_VALUES;			/*if want fake values for sizes and not in prefs set */
	

			if ((clparams[9]==-1))
			{
				zz=0;
				/*test SCRIPT switch */
			}
			
			if ((clparams[10]==-1))
			{
				zz=0;
			/*test for odd crash */
			}

		
		}
	
	printf("OPTIONS ENDS AS %d\n",options);

The error occurs right at the very end of the main function at the line I have marked <<<<<<<<< :

				/*clear any stored volist*/
				printf("Clearing vollist_head\n");
				while(vollist_head)
				{
					vollist_temp=vollist_head->next;
   					
   					printf("vollist_temp mem is now %d\n",&vollist_temp);
   					printf("About to freevec vollist_head %d\n",&vollist_head);
   					
   					FreeVec(vollist_head);
   					vollist_head=vollist_temp;
					printf("OK\n");
   				}
		
		
							
				printf("FREE ARGS\n");
				FreeArgs(Args); /* clear mem used for command line*/
				printf("FREE INFODAT\n");
   				FreeVec(infodat);
   				return_code=RETURN_OK;
   			}

			printf("close lib UBase\n");
   			CloseLibrary(UBase);
			printf("close lib DBase\n");

   			CloseLibrary((struct Library *)DBase);
		printf("OK\n");
   		}
   		else
   		{
   			if (DBase)
   			{
   				CloseLibrary((struct Library *) DBase);
   				
   			}
   			Printf("%s\n",(long)messages[22]); 

   		}
   	  	if (PrefsFile) /*if we allocated a prefs file free it*/
	  	{
  			printf("close PREFS file\n");
	  		printf("free prefs =%x\n",PrefsFile);
  			free (PrefsFile);
			printf("OK\n");
  		}
  	}
   	 	
   	
   	printf("END\n");
return return_code;  /*<<<<<<<<<<<<CRASH!<<<<<<<<<<*/
}

I'm baffled. Why would it work with all the other switches?

Why does it work fine with multiple devices specified, no devices specified, all the other switches no other switches etc etc.

It just crashes if I specify SCRIPT or DUD!

Is there a limit to the number of switches you can use?

This is driving me nuts.

[a/b]

How is clparams defined, as [9] or [11]?
This looks like [9] and local (on stack) resulting in stack corruption.

[chocsplease]

Originally Posted by a/b

How is clparams defined, as [9] or [11]?
This looks like [9] and local (on stack) resulting in stack corruption.

I'll give you three guesses and the first 2 don't count....

It's 9 - why I did't even think of checking this is beyond me

Many thanks for saving my sanity once again a/b. (Wish there was a way of upvoting or something on this forum).