LARL vs. Literal Alignment

2010-08-21 Thread John P. Baker
When using the LARL instruction to reference a literal (i.e., =X'..'), I receive an ASMA058E error message due to the literal not being property aligned (on a halfword boundary). It would seem to me that when a relative instruction references a literal, the assembler should force an otherwise

Re: LARL vs. Literal Alignment

2010-08-21 Thread Paul Gilmartin
On Aug 21, 2010, at 12:04, John P. Baker wrote: When using the LARL instruction to reference a literal (i.e., =X'..'), I receive an ASMA058E error message due to the literal not being property aligned (on a halfword boundary). It would seem to me that when a relative instruction references a

Re: LARL vs. Literal Alignment

2010-08-21 Thread Paul Gilmartin
On Aug 21, 2010, at 12:04, John P. Baker wrote: When using the LARL instruction to reference a literal (i.e., =X'..'), I receive an ASMA058E error message due to the literal not being property aligned (on a halfword boundary). ... Is there any known method by which to force an otherwise

Re: LARL vs. Literal Alignment

2010-08-21 Thread John P. Baker
Gil, I found the same thing. However, I believe that it is reasonable to expect that HLASM, when referencing a literal by way of an instruction using relative addressing, should force the literal to a halfword boundary. Since this is not presently the case, I plan to submit a requirement to IBM

Re: LARL vs. Literal Alignment

2010-08-21 Thread Steve Comstock
John P. Baker wrote: When using the LARL instruction to reference a literal (i.e., =X'..'), I receive an ASMA058E error message due to the literal not being property aligned (on a halfword boundary). It would seem to me that when a relative instruction references a literal, the assembler

Re: LARL vs. Literal Alignment

2010-08-21 Thread John P. Baker
Steve, There is no problem when the literal is specified as H'nn', F'nn', or of any other form which requires halfword or better alignment. Likewise, there is no problem with a binary, character, or hexadecimal literal whose representation consists of an even number of bytes. However, a

[no subject]

2010-08-21 Thread john gilmore
John Gilmore Ashland, MA 01721-1817 USA

Re: LARL vs. Literal Alignment‏

2010-08-21 Thread john gilmore
The notion that customers cannot reasonably be deprived of literals suggests that what is in question is some fill-in-the-blanks situation. Such problems can be dealt with under the hood (bonnet), whether they be screen-input or macro keyword-parameter ones: query and note assembled lengths

Re: LARL vs. Literal Alignment

2010-08-21 Thread Tony Harminc
On 21 August 2010 17:27, John P. Baker jbaker...@comporium.net wrote: There is no problem when the literal is specified as H'nn', F'nn', or of any other form which requires halfword or better alignment. Likewise, there is no problem with a binary, character, or hexadecimal literal whose

Re: LARL vs. Literal Alignment

2010-08-21 Thread John P. Baker
Tony, A non-literal constant, if improperly aligned by the programmer, should be flagged in error. On the other hand, the placement of a literal constant is determined by the assembler. That being the case, I believe that it is reasonable for the assembler to provide correct alignment. John P.

Re: LARL vs. Literal Alignment

2010-08-21 Thread Steve Smith
John P. Baker wrote: When using the LARL instruction to reference a literal (i.e., =X'..'), I receive an ASMA058E error message due to the literal not being property aligned (on a halfword boundary). It would seem to me that when a relative instruction references a literal, the assembler

Re: LARL vs. Literal Alignment

2010-08-21 Thread Paul Gilmartin
On Aug 21, 2010, at 14:59, Steve Comstock wrote: John P. Baker wrote: When using the LARL instruction to reference a literal (i.e., =X'..'), I receive an ASMA058E error message due to the literal not being property aligned (on a halfword boundary). 1. LARL stores the address of the target

Re: LARL vs. Literal Alignment‏

2010-08-21 Thread Paul Gilmartin
On Aug 21, 2010, at 16:44, john gilmore wrote: The notion that customers cannot reasonably be deprived of literals suggests that what is in question is some fill-in-the-blanks situation. Such problems can be dealt with under the hood (bonnet), whether they be screen-input or macro

Re: LARL vs. Literal Alignment

2010-08-21 Thread Steve Comstock
Paul Gilmartin wrote: On Aug 21, 2010, at 14:59, Steve Comstock wrote: John P. Baker wrote: When using the LARL instruction to reference a literal (i.e., =X'..'), I receive an ASMA058E error message due to the literal not being property aligned (on a halfword boundary). 1. LARL stores the

Re: OF what use are one-bit counts for a bit string?

2010-08-21 Thread Tony Harminc
On 3 August 2010 07:55, Don Higgins d...@higgins.net wrote: I though I'd start a new thread with this subject so those searching index can find it. I stumbled across a recent paper in computational linguistics that refers to the usefulness of such a feature. The paper is far too specialized

Re: OF what use are one-bit counts for a bit string?

2010-08-21 Thread John P. Baker
Don, A useful application of a Count 1-Bits instruction could be page allocation. For example, let us consider the following case. My system is configured with 1M segments and 4K pages. A segment map of 32 bytes can be used to map the status of the 256 pages comprising the segment. I issue a