Обсуждение: hi, trying to compile postgres 8.3.11
hi, i tried to compile postgres 8.3.11 on SCO OpenServer 5.0.7, but ... when the xlog.o is being compiled i have the next error: /usr/tmp/ccihgiYL.s: 1113: syntax error at name f /usr/tmp/ccihgiYL.s: 1113: syntax error at integer constant: 1 i dont have copy of that assembly file because was deleted... What can i do to solve this? by the way i've compiled postgres 8.3.1 and i don't have problem. Regards,erobles.
On Thu, May 27, 2010 at 10:42 AM, erobles <erobles@sensacd.com.mx> wrote: > hi, > i tried to compile postgres 8.3.11 on SCO OpenServer 5.0.7, but ... when > the xlog.o is being compiled i have the next error: > > /usr/tmp/ccihgiYL.s: 1113: syntax error at name f > /usr/tmp/ccihgiYL.s: 1113: syntax error at integer constant: 1 > > i dont have copy of that assembly file because was deleted... > > What can i do to solve this? by the way i've compiled postgres 8.3.1 and > i don't have problem. What assembler are you using? One possibility is to try and upgrade gas if you using it or using gas if you are using the stock assembler. I rate chance of success at about 25%. What compiler are you using? merlin
erobles <erobles@sensacd.com.mx> writes:
> i tried to compile postgres 8.3.11 on SCO OpenServer 5.0.7, but ...
> when the xlog.o is being compiled i have the next error:
> /usr/tmp/ccihgiYL.s: 1113: syntax error at name f
> /usr/tmp/ccihgiYL.s: 1113: syntax error at integer constant: 1
Looks like a compiler bug. Get a newer compiler from SCO if you can,
or consider switching to gcc.
regards, tom lane
i have been using gcc 2.95.2 to compile.... On 05/27/2010 10:02 AM, Tom Lane wrote: > erobles<erobles@sensacd.com.mx> writes: > >> i tried to compile postgres 8.3.11 on SCO OpenServer 5.0.7, but ... >> when the xlog.o is being compiled i have the next error: >> > >> /usr/tmp/ccihgiYL.s: 1113: syntax error at name f >> /usr/tmp/ccihgiYL.s: 1113: syntax error at integer constant: 1 >> > Looks like a compiler bug. Get a newer compiler from SCO if you can, > or consider switching to gcc. > > regards, tom lane > >
erobles <erobles@sensacd.com.mx> writes:
> i have been using gcc 2.95.2 to compile....
At least get onto 2.95.3 ;-). I've been using that version on HPPA
for quite awhile and haven't tripped across any bugs. But in any case
these are stone-age versions.
regards, tom lane
On 27/05/2010 10:42 PM, erobles wrote: > hi, > i tried to compile postgres 8.3.11 on SCO OpenServer 5.0.7 Possibly stupid question: Why? Do you need the Pg server to run on SCO OpenServer? Or just a client? Have you considered running your (presumably SCO-based) client with a network connection to a Pg server running on something a bit more modern and maintainable? If so, is your goal to get a newer libpq built for SCO? Because getting libpq to build might be a whole lot easier than getting the whole Pg client+server to build and run. I maintain a SCO OpenServer 5.0.5 box (thankfully now as a VM) and I'd never, ever, ever run any service I didn't have to on it. It's a fossil, and while I have to keep it around for one app (for now) I cannot imagine voluntarily deploying anything new on it under any circumstances. -- Craig Ringer
On 05/27/2010 10:29 AM, Craig Ringer wrote: > On 27/05/2010 10:42 PM, erobles wrote: >> hi, >> i tried to compile postgres 8.3.11 on SCO OpenServer 5.0.7 > > Possibly stupid question: Why? Do you need the Pg server to run on SCO OpenServer? Yes, i need it :-P Before i have running pg 7.2 after we migrate to 8.3.1 so i want to compile the newest postgres , which has all the fixes until now.
erobles <erobles@sensacd.com.mx> wrote: >> Do you need the Pg server to run on SCO OpenServer? > > Yes, i need it :-P Of course it's none of my business, but whenever I had a supplier insisting on some idosyncratic or obsolete OS I started thinking hard about replacing the supplier and their product. Even worse if the supplier is gone and you're using a totally unsupported product. All IMHO of course, and I've supported some peculiar setups when business requirements made alternatives impossible. Good luck, Giles