[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Thanks for the responses! Consolidated replies below...
> Also AtomicWord should not be 32bit int on 64bit machines (but I did not look
> at the full code in context as I didn't have an extracted ff52 source tree
> lying around).
Yes, the modified section falls within the "else" of a 64-bit ifdef.
> I am confused by the patch though - you added a comment about arm there,
> but actually lots of architectures use long as intptr_t.
Yeah, that's just sloppy on my part. I didn't write those lines, just
copied them from an alternate atomicops.h that doesn't have the same
> We want a single source version (upstream, patched by us) that will then build
TBH, I wasn't thinking about this from the standpoint of introducing
the patch officially into the pkgsrc, just alerting SPARC users to the
possibility. If it were going to be added to the distro with
cross-platform compatibility in mind, I'd instead suggest the
2021-05-29 21:57:30.851502479 -0600
2021-05-30 09:01:48.791246976 -0600
@@ -80,6 +80,10 @@
// Atomic64 routines below, depending on your architecture.
#if defined(__OpenBSD__) && !defined(GOOGLE_PROTOBUF_ARCH_64_BIT)
typedef Atomic32 AtomicWord;
+// NetBSD/sparc uses long for intptr_t, which fails -fpermissive checks.
+// So explicitly use int32 here instead.
+#elif defined(__NetBSD__) && defined(ARCH_CPU_SPARC_FAMILY) &&
+typedef int32 AtomicWord;
typedef intptr_t AtomicWord;
> I would say Netsurf is the one of best browers for 32bit Sparcs at this point.
Thanks, I'll take a look! Looks like it doesn't have a SPARC binary
package either, so hopefully I don't run into anything similar
Main Index |
Thread Index |