It's ENTIRELY possible that having the HP ANSI/C compiler will resolve this issue as well -- but the ONLY reason for that compiler that we've found so far is for 'samba' -- so I'm not able to justify the expense (I never even bothered to check, but I assume there is an expense associated with it)
we built using the gcc 3.3.2 compiler from the HP Porting site on the TCOE environment -- so my 'theory' would be that the libraries that are munged (and there are most certainly munged libraries in the plain OE environment) are 'fixed' in the two configurations we've discussed (ANSI C compiler from HP, or TCOE environment) so it would appear that these two solutions work.... On Thu, 29 Jul 2004 16:44:02 -0400 (EDT), Ryan Novosielski <[EMAIL PROTECTED]> wrote: > Even with HP-AnsiC? I don't have this kind of problem. I have been > building Samba for years with the HP compilers. There is the occasional > snafu (Ying Li's patch does solve the problem however), but it generally > works. I don't believe I have this TCOE environment. > > ---- _ _ _ _ ___ _ _ _ > |Y#| | | |\/| | \ |\ | | | Ryan Novosielski - Jr. UNIX Systems Admin > |$&| |__| | | |__/ | \| _| | [EMAIL PROTECTED] - 973/972.0922 (2-0922) > \__/ Univ. of Med. and Dent. | IST/ACS - NJMS Medical Science Bldg - C630 > > > > On Thu, 29 Jul 2004, topher wrote: > > > well -- after MUCH messing about, and traipsing through the source > > code, and traversing through the configure scripts -- and tracking > > down libraries and library dependencies -- and header files... (you > > get the general idea here) -- > > > > rebuild a server to test a theory, and it appears to hold true -- if > > you install the 'regular' Operating Environment Component, there are > > actually libraries around that are totally botched and messed up -- > > all kinds of flaws in them... swlist HPUX11i-OE will tell you if you > > have this configuration -- if it comes back with information, that's > > what you have installed, and you're kinda up a tree... > > > > IF, however, you installed the HP-UX Technical Computing OE Component > > (HPUX11i-TCOE) those libraries are fixed -- and you are able to > > configure, make, and make install HP with no problems... (until you > > try to access the IPC$ share on the HP host with a Mac OS X system, > > then you get problems, but you can map directly to the specific share > > without any problems at all.. go figure) -- > > > > SO -- if you can get your hands on a TCOE system, you can run > > 'configure --disable-shared --enable-statid' and then you can port > > your /usr/local/samba directory around to the various other servers -- > > it appears to be stable thus far, and I'm just gonna keep my fingers > > crossed > > > > no -- I don't have any clue how you get from OE to TCOE without > > re-installing -- sorry for that one... > > > > later all! > > > > toph > > > > > > ---------- Forwarded message ---------- > > From: topher <[EMAIL PROTECTED]> > > Date: Wed, 28 Jul 2004 13:32:25 -0400 > > Subject: Fwd: Building Samba 3.0.5 on HP-UX 11.11 > > To: [EMAIL PROTECTED] > > > > got a couple of requests for further information -- so here's an update: > > > > yes, I've done the configure and make as root and as a regular user, > > things die with either compiler in the same place I'm afraid > > > > yes, I've run 'autoconf' before running configure, and the only > > difference I've seen is that the errors during configure get more > > 'stuff' wrapped around them to make them more apparent... the errors > > in the log are the same -- > > > > here is the exact error message I get during make using gcc: > > > > Compiling dynconfig.c > > In file included from include/includes.h:109, > > from dynconfig.c:21: > > /usr/include/sys/socket.h:484: error: parse error before "sendfile" > > /usr/include/sys/socket.h:484: error: parse error before "bsize_t" > > /usr/include/sys/socket.h:486: error: parse error before "sendpath" > > /usr/include/sys/socket.h:486: error: parse error before "bsize_t" > > /usr/include/sys/socket.h:493: error: parse error before "sendfile64" > > /usr/include/sys/socket.h:493: error: parse error before "bsize64_t" > > /usr/include/sys/socket.h:495: error: parse error before "sendpath64" > > /usr/include/sys/socket.h:495: error: parse error before "bsize64_t" > > In file included from /usr/include/sys/mp.h:55, > > from /usr/include/sys/malloc.h:59, > > from /usr/include/net/netmp.h:53, > > from /usr/include/net/if.h:59, > > from include/includes.h:275, > > from dynconfig.c:21: > > /usr/include/machine/sys/setjmp.h:45: error: redefinition of `struct label_t' > > In file included from dynconfig.c:21: > > include/includes.h:495: error: conflicting types for `socklen_t' > > /usr/include/sys/socket.h:199: error: previous declaration of `socklen_t' > > make: *** [dynconfig.o] Error 1 > > > > any other thoughts/suggestions are appreciated and MORE than welcome! > > > > toph > > > > > > > > ---------- Forwarded message ---------- > > From: topher <[EMAIL PROTECTED]> > > Date: Wed, 28 Jul 2004 09:32:06 -0400 > > Subject: Building Samba 3.0.5 on HP-UX 11.11 > > To: [EMAIL PROTECTED] > > > > First -- to avoid the common answers to the 'building on HP-UX' > > questions that seem to be available on the list and the web: > > > > 1) using the depot from the Porting Center won't work, it's well > > downrev of 3.0.5 and I need to get the new security fixes (so that the > > security team will go back into it's dark dank den and fear not for > > the security of our environment) > > > > 2) samba.org only has a 3.0.4a binary package for HP-UX -- so again, > > no security fixes > > > > as such, I'm trying to build 3.0.5 on HP-UX 11.11 -- and I'm running > > into the same problem I pretty much always seem to run into: > > > > if I use gcc (gcc 3.3.2 off the porting center), configure runs > > through with several errors that look like: > > > > configure: WARNING: foo/bar.h: present but cannot be compiled > > configure: WARNING: foo/bar.h: check for missing prerequisite headers? > > configure: WARNING: foo/bar.h: proceeding with the preprocessor's result > > > > where foo and bar are any of about a dozen different headers... I > > suspect that if I could find a solution to this one, then I'd be able > > to get somewhere -- the configure.log shows that the problem with the > > headers is: > > > > configure:9644: gcc -c -O -D_HPUX_SOURCE -D_POSIX_SOURCE -D_LARGEFILE64_SOURCE > > -D_ALIGNMENT_REQUIRED=1 -D_MAX_ALIGNMENT=4 -DMAX_POSITIVE_LOCK_OFFSET=0x1ffffff > > ffffLL conftest.c >&5 > > In file included from configure:9622: > > /usr/include/sys/socket.h:484: error: parse error before "sendfile" > > /usr/include/sys/socket.h:484: error: parse error before "bsize_t" > > /usr/include/sys/socket.h:486: error: parse error before "sendpath" > > /usr/include/sys/socket.h:486: error: parse error before "bsize_t" > > /usr/include/sys/socket.h:493: error: parse error before "sendfile64" > > /usr/include/sys/socket.h:493: error: parse error before "bsize64_t" > > /usr/include/sys/socket.h:495: error: parse error before "sendpath64" > > /usr/include/sys/socket.h:495: error: parse error before "bsize64_t" > > configure:9647: $? = 1 > > configure: failed program was: > > #line 9617 "configure" > > > > (that's just one instance of the headers failing -- if you want more, > > lemme know and I'll make a big list for ya -- this error is also the > > error I get if I run a make, and things go downhill from there...) > > > > I read in one posting that you can't use the gcc compiler, but you > > have to use the cc that comes with HP (course I read in another post > > that you have to use an ANSI compiler, so you have to use gcc and > > can't use the cc that comes with HP -- but I digress) -- so another > > configure is run (on a whole new set of the source code to prevent any > > confusion) -- but this time I use: > > > > CC=/usr/ccs/bin/cc ./configure > > > > which appears to work, as the first line of output is: > > checking for gcc... /usr/ccs/bin/cc > > > > and the configure goes on from there, right up until: > > > > Using libraries: > > LIBS = -lgen -lsec -lnsl > > AUTH_LIBS = > > checking configure summary... WARNING: No automated network interface > > determination > > ERROR: no seteuid method available > > configure: error: summary failure. Aborting config > > > > it seems like it gets 'so close', as this appears to about 20 lines > > short of the 'successful' configure using 'gcc' > > > > Please -- any help would be much appreciated -- I'm just about willing > > to try anything to get the security gorillas off my back -- I've got > > this build working on AIX 5.1 and 5.2 and Sun 8 and 9 -- it's just > > this silly HP thing that I can't get finished... I'm begging here -- > > please help! > > > > <grin> > > > > toph > > -- > > To unsubscribe from this list go to the following URL and read the > > instructions: http://lists.samba.org/mailman/listinfo/samba > > > -- To unsubscribe from this list go to the following URL and read the instructions: http://lists.samba.org/mailman/listinfo/samba
