cvs code from this morning (about 9am PST) breaks on slackware 7.1 w/ gcc
2.95.2.1 with an undefined reference to session_setup_sia in session.o.
this seems to be the culprit here:
#ifdef HAVE_OSF_SIA
#else /* HAVE_OSF_SIA */
session_setup_sia(pw->pw_name, ttyname);
since i have no idea what that's trying to accomplish (and seems to be a bit
backwards to me from looking at the rest of the code dealing with sia), i
figure i'll just let whoever wrote it sort it out.
devon
before i do my shopping i'll just answer a few of the other questions: re: the fox machine. jordan sent an email to fox telling them about the larger apache process, but not recommending a memory upgrade, just saying "we'll keep a close watch on it to make sure that it doesn't become an issue." the memory IS going to become an issue. maybe not today, maybe not next week, but if fox does their job and promotes the movies and the websites with the movies, it's going to become an issue very quickly and pretty much without notice. which makes 'keeping a close watch on it' pretty pointless, especially considering the turnaround time it'd take to get a memory upgrade for the machine. IMO we need to hit 'em for it NOW, before it's even a baby issue. re: using the e450 as a db machine for miramax. the e450 is best known for its disk space and _amazing_ disk and network i/o. the network i/o isn't so much of a feature we'd want from a db machine as the disk space and i/o performance. large db's have to read large files every time someone wants some info from 'em. i think that the e450 will probably be better suited for fox as a dedicated db machine, but if miramax wants to pay for it as a dedicated db machine then i'm not going to complain. miramax could probably get away with a x86 machine with dual 800's, 2GB of ram, and some software raid5'd disk space (we'd need at least 5 disks in the machine: 2 mirrored OS drives, and 3+ 18GB raid5'd content drives, spread across at least 2 scsi ultra2 channels). the e450 would be total overkill for miramax to use as a dedicated db machine. as a shared machine, it just doesn't sound right, especially if we'd have to tell fox.com they're going on a shared db machine. if jordan hasn't told miramax they're going on a shared db machine, then there's really no damage done, as far as i can see. we just don't put anybody else on the e450 with 'em! we might also consider putting paramount's mailgun stuff on the e450. that nt machine just doesn't cut the mustard (even marcus agrees with me that we need to get that thing off of NT and onto something more stable and robust). i'll have some part no.s and specs on that ram for you shortly. devon ----- Original Message ----- From: "Devon Bleak" <devon at admin2.gisnetworks.com> To: <openssh-unix-dev at mindrot.org> Sent: Tuesday, February 13, 2001 9:43 AM Subject: cvs bulid breaks on slackware> cvs code from this morning (about 9am PST) breaks on slackware 7.1 w/ gcc > 2.95.2.1 with an undefined reference to session_setup_sia in session.o. > > this seems to be the culprit here: > > #ifdef HAVE_OSF_SIA > #else /* HAVE_OSF_SIA */ > session_setup_sia(pw->pw_name, ttyname); > > since i have no idea what that's trying to accomplish (and seems to be abit> backwards to me from looking at the rest of the code dealing with sia), i > figure i'll just let whoever wrote it sort it out. > > devon > > >
before i do my shopping i'll just answer a few of the other questions: re: the fox machine. jordan sent an email to fox telling them about the larger apache process, but not recommending a memory upgrade, just saying "we'll keep a close watch on it to make sure that it doesn't become an issue." the memory IS going to become an issue. maybe not today, maybe not next week, but if fox does their job and promotes the movies and the websites with the movies, it's going to become an issue very quickly and pretty much without notice. which makes 'keeping a close watch on it' pretty pointless, especially considering the turnaround time it'd take to get a memory upgrade for the machine. IMO we need to hit 'em for it NOW, before it's even a baby issue. re: using the e450 as a db machine for miramax. the e450 is best known for its disk space and _amazing_ disk and network i/o. the network i/o isn't so much of a feature we'd want from a db machine as the disk space and i/o performance. large db's have to read large files every time someone wants some info from 'em. i think that the e450 will probably be better suited for fox as a dedicated db machine, but if miramax wants to pay for it as a dedicated db machine then i'm not going to complain. miramax could probably get away with a x86 machine with dual 800's, 2GB of ram, and some software raid5'd disk space (we'd need at least 5 disks in the machine: 2 mirrored OS drives, and 3+ 18GB raid5'd content drives, spread across at least 2 scsi ultra2 channels). the e450 would be total overkill for miramax to use as a dedicated db machine. as a shared machine, it just doesn't sound right, especially if we'd have to tell fox.com they're going on a shared db machine. if jordan hasn't told miramax they're going on a shared db machine, then there's really no damage done, as far as i can see. we just don't put anybody else on the e450 with 'em! we might also consider putting paramount's mailgun stuff on the e450. that nt machine just doesn't cut the mustard (even marcus agrees with me that we need to get that thing off of NT and onto something more stable and robust). i'll have some part no.s and specs on that ram for you shortly. devon ----- Original Message ----- From: "Devon Bleak" <devon at admin2.gisnetworks.com> To: <openssh-unix-dev at mindrot.org> Sent: Tuesday, February 13, 2001 9:43 AM Subject: cvs bulid breaks on slackware> cvs code from this morning (about 9am PST) breaks on slackware 7.1 w/ gcc > 2.95.2.1 with an undefined reference to session_setup_sia in session.o. > > this seems to be the culprit here: > > #ifdef HAVE_OSF_SIA > #else /* HAVE_OSF_SIA */ > session_setup_sia(pw->pw_name, ttyname); > > since i have no idea what that's trying to accomplish (and seems to be abit> backwards to me from looking at the rest of the code dealing with sia), i > figure i'll just let whoever wrote it sort it out. > > devon > > >
It a problem when on irix systems too. At 09:43 AM 2/13/01 -0800, Devon Bleak wrote:>cvs code from this morning (about 9am PST) breaks on slackware 7.1 w/ gcc >2.95.2.1 with an undefined reference to session_setup_sia in session.o. > >this seems to be the culprit here: > >#ifdef HAVE_OSF_SIA >#else /* HAVE_OSF_SIA */ > session_setup_sia(pw->pw_name, ttyname); > >since i have no idea what that's trying to accomplish (and seems to be a bit >backwards to me from looking at the rest of the code dealing with sia), i >figure i'll just let whoever wrote it sort it out. > >devon
On Tue, 13 Feb 2001, Devon Bleak wrote:> cvs code from this morning (about 9am PST) breaks on slackware 7.1 w/ gcc > 2.95.2.1 with an undefined reference to session_setup_sia in session.o. > > this seems to be the culprit here: > > #ifdef HAVE_OSF_SIA > #else /* HAVE_OSF_SIA */ > session_setup_sia(pw->pw_name, ttyname); > > since i have no idea what that's trying to accomplish (and seems to be a bit > backwards to me from looking at the rest of the code dealing with sia), i > figure i'll just let whoever wrote it sort it out. >This was reported a few moment ago by Kevin. [..snip of patch..] #ifdef HAVE_OSF_SIA -#else /* HAVE_OSF_SIA */ session_setup_sia(pw->pw_name, ttyname); +#else /* HAVE_OSF_SIA */ #ifdef HAVE_CYGWIN