Path: gmdzi!unido!mcsun!uunet!europa.asd.contel.com!darwin.sura.net!gatech! swrinde!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!cunixf.cc.columbia.edu!em21 From: e...@cunixf.columbia.edu (Eben Moglen) Newsgroups: comp.lang.perl Subject: Perl 4.x for MS-DOS Message-ID: <1991Sep15.151538.11499@cunixf.cc.columbia.edu> Date: 15 Sep 91 15:15:38 GMT Sender: e...@cunixf.cc.columbia.edu (Eben Moglen) Organization: Columbia University Lines: 4 Like another recent post, I seek Perl 4.x for MSDOS, and am using Borland C++ rather than Microsoft C. Last post requested response by e-mail, but obviously there are enough of us to justify a summary. Thanks very much.
Path: gmdzi!unido!mcsun!ukc!acorn!unipalm!ian From: i...@unipalm.uucp (Ian Phillipps) Newsgroups: comp.lang.perl Subject: Re: Perl 4.x for MS-DOS Message-ID: <1991Sep16.141441.12242@unipalm.uucp> Date: 16 Sep 91 14:14:41 GMT References: <1991Sep15.151538.11499@cunixf.cc.columbia.edu> Organization: Unipalm Ltd., Cambridge, England Lines: 30 e...@cunixf.columbia.edu (Eben Moglen) writes: >Like another recent post, I seek Perl 4.x for MSDOS, and am using >Borland C++ rather than Microsoft C. Last post requested response by >e-mail, but obviously there are enough of us to justify a summary. >Thanks very much. The author of the port (tth...@daimi.aau.dk) did, I belive post a news article on the subject, as well as kindly mailing me a copy. I can't remember exactly where he uploaded it (don't have FTP access yet, so I don't remember), but it's one of the well-known Finnish FTP sites. [I guess this a cue for Timo to followup...] The copy works (thanks very much, Tommy!) but is a lot larger than the 3.x version I've been using for a while - hence I find I can't do the "system" calls I want to. I was hoping to compile the lot using Borland, to win myself some space with the fabled "vroom". Sadly, neither eval.c nor toke.c came within a mile of being compilable (both ran out of some table space less than half way through). I did a little hackery which proved that it was neither the number of case/gotos nor all the unused #defines that killed the compiler, and gave up. [I removed all the gotos from eval.c, which just moved the collapse point down by the same amount. I ran the whole thing through a CPP, which worked, but the compiler stopped same place. I also tried optimisations off, and even small model. No dice.] Maybe I could use a MSC-compiled eval.c/toke.c, and BC++ for the rest. Or maybe not. Any comments? Anyone fancy taking this up with Borland?
Path: gmdzi!unido!mcsun!sunic!dkuug!daimi!tthorn From: tth...@daimi.aau.dk (Tommy Thorn) Newsgroups: comp.lang.perl Subject: Re: Perl 4.x for MS-DOS Message-ID: <1991Sep17.170249.607@daimi.aau.dk> Date: 17 Sep 91 17:02:49 GMT References: <1991Sep15.151538.11499@cunixf.cc.columbia.edu> <1991Sep16.141441.12242@unipalm.uucp> Sender: tth...@daimi.aau.dk (Tommy Thorn) Organization: DAIMI: Computer Science Department, Aarhus University, Denmark Lines: 61 i...@unipalm.uucp (Ian Phillipps) writes: >e...@cunixf.columbia.edu (Eben Moglen) writes: >The author of the port (tth...@daimi.aau.dk) did, I belive post a news Please, no! I was not the author, Diomidis Spinellis (d...@cc.ic.ac.uk) and Tom Dinger (tdin...@East.Sun.COM) was. I just compiled and linked it with gdbm-1.4 according to the instruction given by them in the README.msdos file, contained in the original perl-4.010 distribution. >don't remember), but it's one of the well-known Finnish FTP sites. [I >guess this a cue for Timo to followup...] Yes, it's available from nic.funet.fi:/pub/msdos/utilities/perl/perl4010.lzh. Or, more elaborate: ftp nic.funet.fi (you are asked for your name, answer with:) ftp (you are asked for a passwd, answer with:) your_own_email_address (give your email addresss) bin (to setup binary communication) cd pub/msdos/utilities/perl get perl4010.lzh bye Please, tell us if you make it available elsewhere (especially in the US). >The copy works (thanks very much, Tommy!) but is a lot larger than the >3.x version I've been using for a while - hence I find I can't do the >"system" calls I want to. Yes, the version I compiled leaves a lot to be desired: a better C compiler (I used MSC v5.1), overlays, extended/expanded memory, smart (exec)swap or whatever would help. I just haven't done anything about it. If I had time, I might try to build it with the gcc-for-msdos compiler. >I was hoping to compile the lot using Borland, to win myself some space >with the fabled "vroom". Sadly, neither eval.c nor toke.c came within a >mile of being compilable (both ran out of some table space less than >half way through). I did a little hackery which proved that it was >neither the number of case/gotos nor all the unused #defines that killed >the compiler, and gave up. Well, I guess that one has to cut eval and toke in severel pieces to make it compile, but it's far from impossible. >Maybe I could use a MSC-compiled eval.c/toke.c, and BC++ for the rest. No, I don't think that will not work. The low-level runtime support system is different, and don't easily coexits. Please keep comp.lang.perl informed of any better ports of perl, there seems to be sufficent interrest to justify making a good binary and including info on it in the perl FAQ. -- Tommy Thorn email: tth...@daimi.aau.dk Computer Science Department "People shouldn't work because they love it, Aarhus University they should work because it hurts." DENMARK -- Bob Sparacino, former Xerox executive
Path: gmdzi!unido!math.fu-berlin.de!ira.uka.de!yale.edu!qt.cs.utexas.edu! cs.utexas.edu!asuvax!gatech!emory!wa4mei!holos0!lbr From: l...@holos0.uucp (Len Reed) Newsgroups: comp.lang.perl Subject: Re: Perl 4.x for MS-DOS Message-ID: <1991Sep20.171428.27268@holos0.uucp> Date: 20 Sep 91 17:14:28 GMT References: <1991Sep15.151538.11499@cunixf.cc.columbia.edu> <1991Sep16.141441.12242@unipalm.uucp> Organization: Holos Software, Inc., Atlanta, GA Lines: 18 In article <1991Sep16.141441.12...@unipalm.uucp> i...@unipalm.uucp (Ian Phillipps) writes: >The copy works (thanks very much, Tommy!) but is a lot larger than the >3.x version I've been using for a while - hence I find I can't do the >"system" calls I want to. Well, it's not necessarily that it's bigger but rather that it doesn't swap itself out when doing system calls. My 3.041 port does exactly that; it needs only 2 K-bytes. Without swapping, it'll use over 300 K-bytes. See my postings on my 4.010 port. -- Len Reed Holos Software, Inc. Voice: (404) 496-1358 UUCP: ...!gatech!holos0!lbr