Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!nntp.coast.net!howland.erols.net!news.nacamar.de!news-kar1.dfn.de!news-stu1.dfn.de!news.belwue.de!news.uni-ulm.de!borchert From: borchert@turing.mathematik.uni-ulm.de (Andreas Borchert) Newsgroups: comp.unix.solaris,comp.unix.bsd.misc,comp.unix.internals,comp.unix.osf.osf1 Subject: Re: Solaris 2.6 Date: 19 Dec 1996 16:49:44 GMT Organization: University of Ulm, SAI, Germany Lines: 72 Distribution: inet Message-ID: <59brn8$nk9@arktur.rz.uni-ulm.de> References: <32986299.AC7@mail.esrin.esa.it> <58n5q1$9ci@arktur.rz.uni-ulm.de> <58pdbt$mjv@web.nmti.com> <58r48o$t7q@arktur.rz.uni-ulm.de> <5948aj$nci@web.nmti.com> NNTP-Posting-Host: turing.mathematik.uni-ulm.de X-Newsreader: slrn (0.9.0.0 (BETA) UNIX) Xref: euryale.cc.adfa.oz.au comp.unix.solaris:93000 comp.unix.bsd.misc:1850 comp.unix.internals:11669 comp.unix.osf.osf1:17128 On 16 Dec 1996 19:35:47 GMT, Peter da Silva <peter@nmti.com> wrote: > In article <58r48o$t7q@arktur.rz.uni-ulm.de>, > Andreas Borchert <borchert@turing.mathematik.uni-ulm.de> wrote: > > Agreed. But this seems to be a common property of all commercial > > UNIX directions now. Once upon a time, there was a simple UNIX system > > named UNIX Edition VII. > > I don't agree. Digital UNIX (apart from the licensing stuff, but I hate license > managers generally) is still pretty simple under the hood. System V release > 4.0 is a lot simpler than Solaris, and even SVR4.2 (the last real System V > release) let you deal with a relatively simple low level environment if you > wanted. System V is far away from being a simple system. Take, for example, the System V interprocess facilities, namely shared memory, semaphores, and messages. These facilities have really ugly interfaces, come with myriads of limited system tables which needs to be increased, and are - in my opinion - not needed. Pipes, sockets, and mmap(2) are significantly better. I assume that the only reason for keeping them is upward compatibility. And that is the curse of commercial software systems: Upward compatibility leads sooner or later to unnecessary complex systems. > (the Solaris device file system is awfully confusing) The nice feature is that device entries are generated automatically during the boot phase (with the reconfiguration flag). You'll find the traditional names under /dev/ and the hardware device names under /devices. I do not know what you consider awfully confusing here. > > > (3) It doesn't provide good BSD administrative semantics. > > > (but it does use some BSD stuff as well) > > > What are you missing here? > > How about printcap? That'd be *really* nice. This went into terminfo which offers essentially the same features -- just with a different interface. There is no question, however, that the System V line printer software belongs to those parts which never should have found their way into Solaris. Better printing systems depend on NIS-tables (or other administrative tables which are distributed on the network) and do not spawn individual processes for each other printer on the network on each host. > > > (4) It's harder to port System V software to Solaris than to other > > > System V boxes. > > > (that's UHC, Intel UNIX, Unixware, and Xenix software) > > > I do *not* believe this -- I've even ported huge amounts of software > > (coming from System-V and from BSD environments) to early Solaris > > releases as Solaris 2.1 with nearly no problems. > > I've had problems. The most fun was one that truncated the first two characters > of every file name when you did filename expansion because there were two > different sets of directory reading routines with different structures with > the same name, and which one you got depended on the path to the C compiler > you picked. Thou shalt never use /usr/ucb/cc in favour of gcc or /opt/SUNWspro/bin/cc. Andreas. -- Andreas Borchert, Universitaet Ulm, SAI, Helmholtzstr. 18, 89069 Ulm, Germany E-Mail: borchert@mathematik.uni-ulm.de WWW: http://www.mathematik.uni-ulm.de/sai/borchert/ PGP key available via ``finger borchert@laborix.mathematik.uni-ulm.de''