rcj@moss.ATT.COM (04/06/88)
In article <4077@whuts.UUCP> chl@whuts.UUCP (LANG) writes: }When I originally posted the message about the recent SYS5 sources for }UNIX ARC being incompatible with the MS-DOS ARC, I had noticed the }-i switch. This does make it compatible; however, why should this }switch be needed. Aren't most people just using ARC to go between the }PC and UNIX (and vice versa). Why force an extra command line argument }which to me really makes it incompatible (OK! if not incompatible, its That's why I took two minutes to change the sense of the -i switch; mine now is IBM-compatible by default and -i makes it incompatible. }at least different and confusing). The resulting error messages complain }about the number of BITS packed with. Not the filename lengths. Judging from some mail I received from someone else, this is a problem unique to Amdahl UTS -- isn't that what whuts is/is running? }names. Let's keep the normal arguments to ARC the same on the PC }and UNIX so it is "compatible" in its truest sense. } }Any other feelings on this subject? } }P.S. Don't get me wrong! I appreciate the poster's effort in modifying }and in posting the sources. Same here. What follows is the text of a letter I sent to the poster asking some questions and pointing out what I consider to be some foibles: First of all, let me thank you for this version of arc for Unix (I'm using the System V version); it has made my life much nicer! Now, unfortunately, to the bugs: 1) With the IBM compatible option on, arc apparently truncates filenames to 12 characters in an attempt to be compatible with DOS. In actuality, DOS is much more restrictive than just that; it allows only 8 characters and, if those up-to-8 characters are followed by a dot, it allows 3 more characters for an extension. So a filename like "cid2astis.m", which your version of arc stores as-is because it is under 12 characters, is unpacked on the PC as "CID2ASTI.M". 2) Again, with the IBM compatible option on, since the PC cannot create files with lower-case names, it would be nice if the specified files on the command-line (those after the archive name) were automagically treated as caseless, and the files in the archive also created caseless. This would truly make it IBM-PC compatible. 3) On a similar note, if I have an archive that I have unpacked from a uuencoded file, and say it comes out as SKY.ARC (note the capitals), then "arc v SKY" results in arc trying to de-archive a non-existent file called SKY.arc. "arc v SKY.ARC" results in arc trying to de-archive SKY.ARC.arc. You may not want to change this; don't know. Thanks for your time, Men's Rights? NO Women's Rights? NO Equal Rights? YES The MAD Programmer -- 201-386-6409 (Cornet 232) alias: Curtis Jackson ...![ ihnp4 ulysses ucbvax allegra ]!moss!rcj ...![ ihnp4 ucbvax akgua watmath ]!clyde!rcj
pjh@mccc.UUCP (Peter J. Holsberg) (04/08/88)
In article <24453@clyde.ATT.COM> rcj@moss.UUCP (Curtis Jackson) writes: == In article <4077@whuts.UUCP> chl@whuts.UUCP (LANG) writes: == }When I originally posted the message about the recent SYS5 sources for == }UNIX ARC being incompatible with the MS-DOS ARC, I had noticed the == }-i switch. This does make it compatible; however, why should this == }switch be needed. Aren't most people just using ARC to go between the == }PC and UNIX (and vice versa). Why force an extra command line argument == }which to me really makes it incompatible (OK! if not incompatible, its == == That's why I took two minutes to change the sense of the -i switch; == mine now is IBM-compatible by default and -i makes it incompatible. == Sorry for having missed this, but what is the -i switch? -- Peter Holsberg UUCP: {rutgers!}princeton!mccc!pjh Technology Division CompuServe: 70240,334 Mercer College GEnie: PJHOLSBERG Trenton, NJ 08690 Voice: 1-609-586-4800