DCMTK build error

[Ryan Schmidt]

On Jul 5, 2010, at 16:55, John B Brown wrote:

> On 7/5/10 2:19 PM, Ryan Schmidt wrote:
>> 
> 
>> http://trac.macports.org/ticket/22840
>> 
>> Patches were submitted to fix problems on Snow Leopard, which in turn broke Leopard. Someone needs to figure out how to fix the port so it builds properly on both Leopard and Snow Leopard.
> 
> 	A 1984 sysv file tells me the use of the proper ifdef can detect which OS is being host to the build. For example:
> 
> cmd/finc.c:#ifdef vax
> cmd/finc.c:#ifdef u370
> cmd/finc.c:#ifdef u3b2
> cmd/finc.c:#ifdef u370
> cmd/finc.c:#ifdef FsMAGIC
> cmd/finc.c:#ifdef DEBUG
> cmd/finc.c:#ifdef FsMAGIC
> cmd/frec.c:#ifdef FsMAGIC
> cmd/frec.c:#ifdef debug
> cmd/frec.c:#ifdef debug
> cmd/frec.c:#ifdef debug
> cmd/frec.c:#ifdef u370
> cmd/nl.c:#ifdef u370
> 
> 	We still use uname do we not?

Yes of course. I'm saying the patches we applied were obviously not correct for Leopard, so someone just needs to go look at them and revise them. This is what was applied:

http://trac.macports.org/changeset/67316

As you can see, the patch changed the program's configure script. The script used to have a block for Tiger and a block for Leopard, and they both did the same thing. The patch we applied condensed these two blocks into one, and made it apply to Tiger, Leopard and Snow Leopard, and add the -D_DARWIN_C_SOURCE switch in all these cases. My first guess would would be that the -D_DARWIN_C_SOURCE switch is not appropriate for Leopard and earlier and is the cause of the breakage. Volunteers who would like to test this theory and submit a new patch to fix the problem would be appreciated.