jrb@petro.UUCP (Jon Boede) (10/13/87)
In article <1842@gryphon.CTS.COM> greg@gryphon.CTS.COM (Greg Laskin) writes: >>>I assume there's an option to cc(1) to increase the amount of heap space. >The -LARGE switch (not present in all implementations of Xenix) invokes a >large MODEL pass of the compiler which allows multiple heap segments during >compilation. Ok, ok... thinking this was going to solve my problems I summarily tried the -LARGE flag only to find that "...: /lib/p1L cannot execute". It was there -- "file /lib/p1L" returned a "separate executable, Large model". The file had permissions "rwx--x--x" and I was trying to run it as a person other than root or the owner. Upon further investigation I found that my execution request was denied because /lib/p1L was thought to be a shell script! In fact, NOTHING compiled in the Large Model runs on this box -- it tries to execute them as shell scripts (and MAN, do you get some wierd errors! :-) I'm now in this mostly for the academic challenge (I've since gotten access to a VAX 11/780 that only one other person uses -- nyuck, nyuck) ... I'm very interested in the concept of porting what I've wrote to all variety of hardware. So, if anybody has seen this before, please let me know... Danke! -- Jon Boede ...!{gatech,ihnp4,ssbn,swrinde,tness1,utanes}!petro!jrb 512/599-1847 -^^^^^^- 2555 N.E. Loop 410, #1403, 78217 "People who are incapable of making decisions are the ones who hit those barrels at freeway exits."
jim@applix.UUCP (Jim Morton) (10/15/87)
In article <462@petro.UUCP>, jrb@petro.UUCP (Jon Boede) writes: > Ok, ok... thinking this was going to solve my problems I summarily tried the > -LARGE flag only to find that "...: /lib/p1L cannot execute". > > It was there -- "file /lib/p1L" returned "separate executable, Large model". > > The file had permissions "rwx--x--x" and I was trying to run it as a person > other than root or the owner. Upon further investigation I found that my > execution was denied because /lib/p1L was thought to be a shell script! > In fact, NOTHING compiled in the Large Model runs on this box -- it tries to > execute them as shell scripts (and MAN, do you get some wierd errors! :-) The most probable reason /lib/p1L is not executable and is thought to be a shell script is due to the SCO serial number authorization. If you look at the first characters of the file using "od -c", you'll see something like "this file has not been serialized yet". For this reason, you must pull these files off using SCO's custom utility, NOT just tar'ing them off the distribution diskette... -- Jim Morton, APPLiX Inc., Westboro, MA UUCP: ...ames!harvard!m2c!applix!jim ...rutgers!harvard!m2c!applix!jim
jrb@petro.UUCP (Jon Boede) (10/16/87)
In article <610@applix.UUCP> jim@applix.UUCP (Jim Morton) writes: >Because in article <462@petro.UUCP>, I write: >> Ok, ok... thinking this was going to solve my problems I summarily tried the >> -LARGE flag only to find that "...: /lib/p1L cannot execute". >> >> It was there -- "file /lib/p1L" returned "separate executable, Large model". >> [and I go on to say that exec must think it's a shell script] > >The most probable reason /lib/p1L is not executable and is thought to be >a shell script is due to the SCO serial number authorization. If you look >at the first characters of the file using "od -c", you'll see something >like "this file has not been serialized yet". I looked at an "od -c" of it an all I got was binary gobbildy gook, starting with 006 002. So I don't think what you describe is the problem. It's not just /lib/p1L that won't execute... ANYthing compiled with -Ml gets "executed" as if it were a shell script. I should point out that what I'm using is a Tandy 1200 (4.77Mhz 8088(6?)), not a 3000 (PC AT) as I'd said. I don't see why that would make any difference; perhaps it doesn't. It might be pointed out that the person who has been most helpful in solving my problems (Hi Greg) also says that this is "pretty weird". Perhaps one of the Boys From Santa Cruz would care to comment. Thanks! Jon -- Jon Boede ...!{gatech,ihnp4,ssbn,swrinde,tness1,utanes}!petro!jrb 512/599-1847 -^^^^^^- 2555 N.E. Loop 410, #1403, 78217 "People who are incapable of making decisions are the ones who hit those barrels at freeway exits."
shap@bunker.UUCP (10/30/87)
In article <464@petro.UUCP> jrb@petro.UUCP (Jon Boede) writes: >In article <610@applix.UUCP> jim@applix.UUCP (Jim Morton) writes: >>Because in article <462@petro.UUCP>, I write: >>> Ok, ok... thinking this was going to solve my problems I summarily tried the >>> -LARGE flag only to find that "...: /lib/p1L cannot execute". >>> >>> It was there -- "file /lib/p1L" returned "separate executable, Large model". >... >I should point out that what I'm using is a Tandy 1200 (4.77Mhz 8088(6?)), not ^^^^ >a 3000 (PC AT) as I'd said. I don't see why that would make any difference; >perhaps it doesn't. I quote (without permission) from the SCO Development system release notes (v2.1.3) "This release includes large model passes of the cmerge compiler. These are invoked using the -LARGE flag with cc. Only 286 machines can ^^^^^^^^^^^^^^^^^ run the large model passes of the compiler..."