Thread: Compilation of pg 7.4.17 fails on HP-UX
Hello, I have just tried to compile postgresql 7.4.17 on a HP-UX 11.11 box (PA-RISC) and the compliation fails in gist. make[4]: Entering directory `/home/adrian/postgresql-7.4.17/src/backend/access/gist' gcc -O2 -fno-strict-aliasing -Wall -Wmissing-prototypes -Wmissing-declarations -I../../../../src/include -D_XOPEN_SOURCE_EXTENDED -c -o gist.o gist.c /var/tmp//ccxYASP0.s: Assembler messages: /var/tmp//ccxYASP0.s:1528: Error: Field not properly aligned [8] (52). /var/tmp//ccxYASP0.s:1528: Error: Invalid operands /var/tmp//ccxYASP0.s:1531: Error: Field not properly aligned [8] (44). /var/tmp//ccxYASP0.s:1531: Error: Invalid operands /var/tmp//ccxYASP0.s:1534: Error: Field not properly aligned [8] (60). /var/tmp//ccxYASP0.s:1534: Error: Invalid operands make[4]: *** [gist.o] Error 1 make[4]: Leaving directory `/home/adrian/postgresql-7.4.17/src/backend/access/gist' make[3]: *** [gist-recursive] Error 2 make[3]: Leaving directory `/home/adrian/postgresql-7.4.17/src/backend/access' make[2]: *** [access-recursive] Error 2 make[2]: Leaving directory `/home/adrian/postgresql-7.4.17/src/backend' make[1]: *** [all] Error 2 make[1]: Leaving directory `/home/adrian/postgresql-7.4.17/src' make: *** [all] Error 2 This doesn't happen in HEAD. $ uname -a HP-UX hpdev B.11.11 U 9000/800 822196241 unlimited-user license $ gcc --version gcc (GCC) 3.3.3 Cheers, Adrian Maier
"Adrian Maier" <adrian.maier@gmail.com> writes:
> I have just tried to compile postgresql 7.4.17 on a HP-UX 11.11 box
> (PA-RISC) and the compliation fails in gist.
I do not think anyone cares about making 7.4.x run on platforms it did
not support before. Use a newer PG release.
regards, tom lane
On Thu, Aug 09, 2007 at 11:46:47AM -0400, Tom Lane wrote: > "Adrian Maier" <adrian.maier@gmail.com> writes: > > I have just tried to compile postgresql 7.4.17 on a HP-UX 11.11 box > > (PA-RISC) and the compliation fails in gist. > > I do not think anyone cares about making 7.4.x run on platforms it did > not support before. Use a newer PG release. Actually, this is PA-RISC, not the numerous emails we've gotten this week about HPUX on ia64 (what is it with the HPUX guys this week?) http://www.postgresql.org/docs/7.4/interactive/supported-platforms.html indicates that HPUX on PA-RISC should work... -- Decibel!, aka Jim Nasby decibel@decibel.org EnterpriseDB http://enterprisedb.com 512.569.9461 (cell)
Decibel! <decibel@decibel.org> writes: > On Thu, Aug 09, 2007 at 11:46:47AM -0400, Tom Lane wrote: >> "Adrian Maier" <adrian.maier@gmail.com> writes: >>> I have just tried to compile postgresql 7.4.17 on a HP-UX 11.11 box >>> (PA-RISC) and the compliation fails in gist. >> >> I do not think anyone cares about making 7.4.x run on platforms it did >> not support before. Use a newer PG release. > Actually, this is PA-RISC, not the numerous emails we've gotten this > week about HPUX on ia64 (what is it with the HPUX guys this week?) Oh, my mistake --- obviously hadn't consumed enough caffeine this morning. [ digs around a bit... ] However, I might have been right for the wrong reasons. I'm thinking Adrian is trying to build for 64-bit HPPA (a beast the 7.4 supported-platforms matrix doesn't know about), and is falling foul of this problem: http://archives.postgresql.org/pgsql-hackers/2004-03/msg01196.php Teodor fixed that in the 8.0 devel cycle, but couldn't back-patch it because it meant an on-disk layout change of gist indexes. What I suspect is that gcc 3.3.3 doesn't have the specific error checks I complained of in the above message, but simply generates bogus assembly code for the incorrect C code :-( regards, tom lane
On 8/10/07, Tom Lane <tgl@sss.pgh.pa.us> wrote: > Decibel! <decibel@decibel.org> writes: > > On Thu, Aug 09, 2007 at 11:46:47AM -0400, Tom Lane wrote: > >> "Adrian Maier" <adrian.maier@gmail.com> writes: > >>> I have just tried to compile postgresql 7.4.17 on a HP-UX 11.11 box > >>> (PA-RISC) and the compliation fails in gist. > >> > >> I do not think anyone cares about making 7.4.x run on platforms it did > >> not support before. Use a newer PG release. > > > Actually, this is PA-RISC, not the numerous emails we've gotten this > > week about HPUX on ia64 (what is it with the HPUX guys this week?) > > Oh, my mistake --- obviously hadn't consumed enough caffeine this > morning. > > [ digs around a bit... ] However, I might have been right for the > wrong reasons. I'm thinking Adrian is trying to build for 64-bit > HPPA (a beast the 7.4 supported-platforms matrix doesn't know about), > and is falling foul of this problem: > http://archives.postgresql.org/pgsql-hackers/2004-03/msg01196.php > Teodor fixed that in the 8.0 devel cycle, but couldn't back-patch it > because it meant an on-disk layout change of gist indexes. > > What I suspect is that gcc 3.3.3 doesn't have the specific error checks > I complained of in the above message, but simply generates bogus > assembly code for the incorrect C code :-( So it looks like this is simply a known issue that hasn't been backpatched. I'll simply use 8.2 . Thanks for your answers, Adrian Maier
"Adrian Maier" <adrian.maier@gmail.com> writes: > On 8/10/07, Tom Lane <tgl@sss.pgh.pa.us> wrote: >> [ digs around a bit... ] However, I might have been right for the >> wrong reasons. I'm thinking Adrian is trying to build for 64-bit >> HPPA (a beast the 7.4 supported-platforms matrix doesn't know about), >> and is falling foul of this problem: >> http://archives.postgresql.org/pgsql-hackers/2004-03/msg01196.php > So it looks like this is simply a known issue that hasn't been backpatched. You didn't confirm in so many words: were you trying to build 64-bit? I'd have expected a 32-bit build to work. regards, tom lane
On Fri, Aug 10, 2007 at 08:08:55AM +0300, Adrian Maier wrote: > So it looks like this is simply a known issue that hasn't been backpatched. > I'll simply use 8.2 . Ok, I'll bite... if you can use 8.2, why were you trying with 7.4 instead? It sounds like you're paying new-car-prices for something that's 5 years old... -- Decibel!, aka Jim Nasby decibel@decibel.org EnterpriseDB http://enterprisedb.com 512.569.9461 (cell)
On 8/10/07, Decibel! <decibel@decibel.org> wrote: > On Fri, Aug 10, 2007 at 08:08:55AM +0300, Adrian Maier wrote: > > So it looks like this is simply a known issue that hasn't been backpatched. > > I'll simply use 8.2 . > > Ok, I'll bite... if you can use 8.2, why were you trying with 7.4 > instead? It sounds like you're paying new-car-prices for something > that's 5 years old... The first attempt was with 7.4 because our application is still running on that version in production. An upgrade is something that has to be done, but it will take a little more time to test and do it properly. Right now I am just doing some tests: i happen to have access to this PA-RISC server and was curious to see how is the application behaving if the database is running on such a machine. But there are no chances to use such a server in production, so my tests are purely ... educative . Cheers, Adrian Maier