Re: MacOS X 10.5.1 and compiling for multiple Architectures

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hey Tom, Shane,

actually, PHP w/out the PG stuff builds just fine with x86_64 and i386 specified at the same time. I didn't bother with the PPC architectures, since I don't have a Leopard capable PPC box any more. My old PowerBook G3/500 point blank refuses to run anything higher than 10.4.11 :).

When I tried the build, including the PG libs for both architectures, I got error messages for missing symbols only for the i386 architecture (since I only built the PG libs for 64-bit). Everything else seemed to be present.

I'm good to run with the x86_64 only binary since my Apache is built using that architecture. I'm also not to worried about any bugs in the 64-bit OS X code, this is for development only, no live data, ever. If it crashes or mangles up my data, so be it, backups must be good for something, right? Would have been nice though, to go and give the completely built libphp5.so module to somebody and say 'here, run it' regardless of what his hardware architecture supports.

best regards,
chris
--
chris ruprecht
network grunt and bit pusher extraordinaíre





On Jan 3, 2008, at 00:41, Tom Lane wrote:

Chris Ruprecht <chris@xxxxxxxxxxxx> writes:
I am trying to build PG 8.3Beta4 for MacOS Leopard.
It works fine when using the default i386 architecture. However, this
is not an option as Apple, in their wisdom, have chosen to build
Apache as a 64-bit application and their PHP 5.2.4 module without PG
support as a Intel 32, Intel 64, PPC 32 and PPC 64 type module. For
those that don't know: Under MacOS, it is possible to build
executables that will run native on multiple architectures, all
contained in a single file - kind of weird, and pretty bloated.

You're not going to have a lot of luck building PG for those four
architectures from a single build run --- you really need a different
pg_config.h for each case.  The specific failure you're seeing seems
to stem from SIZEOF_DATUM not agreeing with sizeof(Datum), but you
can bet your bottom dollar that that's just the tip of the iceberg.

I'd be somewhat interested to see a build approach that supports this,
but I don't want to spend any of my own time on it.  I have a vague
suspicion that Apple probably built PHP four separate times and then
melded the executables after the fact.

			regards, tom lane

Attachment: smime.p7s
Description: S/MIME cryptographic signature


[Index of Archives]     [KVM ARM]     [KVM ia64]     [KVM ppc]     [Virtualization Tools]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite Questions]     [Linux Kernel]     [Linux SCSI]     [XFree86]

  Powered by Linux