From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <45219fb00801030254y72b5ac1fjd5100c4dd6888e4c@mail.gmail.com> Date: Thu, 3 Jan 2008 11:54:27 +0100 From: "=?ISO-8859-1?Q?Llu=EDs_Batlle?=" To: "Fans of the OS Plan 9 from Bell Labs" <9fans@cse.psu.edu> Subject: Re: [9fans] opera under linuxemu In-Reply-To: <20080103083708.GA826@shodan.homeunix.net> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <13426df10801021655r28a34e4ep98bdde53f18f447d@mail.gmail.com> <32d987d50801021929v1a3c478awdd22daebb96d8aa@mail.gmail.com> <20080103083708.GA826@shodan.homeunix.net> Topicbox-Message-UUID: 26111b0e-ead3-11e9-9d60-3106f5b1d025 I agree on this. On the other hand, I seem to remember that any program using the glibc's network name resolver can be totally statically linked in Linux. That has probably something to do with those dynamic /lib/libnss* modules for name resolution, specified in /etc/nsswitch.conf. Maybe a special compilation of glibc with those modules in static form can produce static binaries, but I don't think any distribution comes with a precompiled glibc like that. 2008/1/3, Martin Neubauer : > For one, I think opera-static doesn't mean it's a static binary but qt is > linked in statically. On the other hand, until just a couple years ago > static linking in linux was no problem at all. But around the time linux 2.6 > came out, the glibc guys apparently decided nobody used static linking > anyway and merrily bollocksed it up. Great, now I'm depressed. > > * Federico G. Benavento (benavento@gmail.com) wrote: > > hola, > > > > getting real static binaries in linux is a bit tricky and no one seems to > > be doing so, they always need ld-linux.so, libnss and others. > > cinap creates some kind of bundles that create a fake ns in /tmp/$lbun > > with this (http://9hal.ath.cx/usr/cinap_lenrek/lbun/mklbun) or something > > like, but I know he got opera running in Plan 9. > > http://9hal.ath.cx/usr/cinap_lenrek/plan9opera.png >