OPTABLE option of Disassembler
The High level Assembler accepts and uses an OPTABLE parm which lets you limit the valid op codes to an architecture level such as XA or 370 (and optionally list the valid OP codes at that level). The Disassembler (ASMDASM) has a comparable ÓPTABLE option which "Specifies the operation code table to be used in disassembling CSECTs." Except the ASMDASM we are using - no version shown, but linked 06.125 WITH Identify data of UK09726 - does not limit the op codes generated, or even check for a valid parameter. Is there a later version which has this function, or am I missing something? Gene Lynd -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Optable option of disassembler
Tony Harminc wrote: >It's useful to limit the opcodes understood, because the disassembler (any >disassembler for this architecture - not just IBM's) is less than perfect at >understanding what is code and what is data. If you know something about the >module you are working on (typically it is some legacy lost-source thing >written some time ago), then it is better to have things that could not be >code in that particular module shown as data rather than bogus instructions. >ASMDASM does allow you to tell it that an area is code-only or data-only, >but often enough you don't know that in detail early in the disassembling >process, and it helps not to have your work cluttered with instructions that >could not have been intended in, say, 1987. Responding to my question: "The High level Assembler accepts and uses an OPTABLE parm which lets you limit the valid op codes to an architecture level such as XA or 370 (and optionally list the valid OP codes at that level). The Disassembler (ASMDASM) has a comparable ÓPTABLE option which "Specifies the operation code table to be used in disassembling CSECTs." Except the ASMDASM we are using - no version shown, but linked 06.125 WITH Identify data of UK09726 - does not limit the op codes generated, or even check for a valid parameter. Is there a later version which has this function, or am I missing something?" Tony expressed my view of the problem very well - sorry I didn't. Typically when we use the Disassembler it is for very old code. But since there is in fact an OPTABLE parameter documented, that suggests IBM also sees the need for the feature. My question remains: is (or will there be) a version of the Disassembler that implements it? Gene Lynd -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Copy replacing without pseudo code
Clark Morris wrote: >>You might even make a stink about the lack of >>improvements in the language that is supposedly the >> one of the future. >>Personally, given IBM and other vendor actions, >>I believe it is a cash cow on a long term death bed. At the Tampa Share, I think, Tom Ross announced that he and his team are back on the job in COBOL development. We can therefore hope for better things from IBM. Gene Lynd -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
calling ISPLINK from COBOL
>>I'm trying to convert the following >> pseudo code to run under COBOL 01 CHAR PIC X(4) VALUE 'CHAR'. is incorrect. When you define a character value with a length less than 8 you need a trailing blank: 01 CHAR PIC X(5) VALUE 'CHAR '. Gene Lynd -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: IF/THEN/ELSE/ENDIF
"An unenforced prohibition is meaningless." Not really, with all due respect to Justice Holmes. The possibility of future enforcement hangs over my guilty head. Specifically, I am testing symbolic parameters in procs, although they are not among the valid operands. For example: PROC ABENDAID=1,0,2 0=NO A-A FILE,2=SHARED ... // IF (&ABENDAID GT 0 & RC < 8) THEN If IBM discovers (or knows) how well this works, I can only hope they decide to document it rather than flush it. Gene Lynd -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Mark load module non-executable
>>Paul Gilmartin asked "Is there any way to force >>the non-executable attribute on a load module?" File-AID 3.1 can do this. Use option I and disposition OLD - you can then overtype the attributes, including EXEC (executable). Gene Lynd -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Summarize length of all variable length records in a file?
John Reda wrote in reply to Sam Knudson: >The following sort control cards should provide what you are looking >for. It simply prints one line per record length and provides the >number of records with that length. You can add headers if you want >a more formal report. >SORT FIELDS=(5,2,CH,A) >INREC FIELDS=(1,4,1,2,X'0001') >SUM FIELDS=(7,4,BI) >OUTREC FIELDS=(1,4,5,2,BI,C' ',7,4,BI) Brilliant. We have had HISTOGRM for years, but it shows record lengths only within a range, which must be at least 4. Sample Histogram: COUNTMIN MAX 3 16 19 3 24 27 5 44 47 Sample with new control statements: 193 253 441 461 473 Totals are the same, but the exact record lengths are totaled, not an arbitrary range. Gene Lynd -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html