35970 commits

This branch

This branch

All branches

Author	SHA1	Message	Date
Andrew Kelley	db622f14c4	Merge pull request #25780 from Vexu/translate-c ... Update Aro and translate-c to latest	2025-11-20 10:24:31 -08:00
Matthew Lugg	8a73fc8d8e	Merge pull request #25981 from mlugg/macos-fuzz-2 ... make the fuzzer vaguely work on macOS	2025-11-20 17:48:35 +00:00
Andrew Kelley	a9568ed296	Merge pull request #25898 from jacobly0/elfv2-progress ... Elf2: more progress	2025-11-20 04:33:04 -08:00
Veikka Tuominen	df50f9e28e	update resinator to Aro changes	2025-11-20 13:12:53 +02:00
Veikka Tuominen	21f3ff2a8d	update Aro and translate-c to latest	2025-11-20 13:12:53 +02:00
Matthew Lugg	a87b533231	std.Io.Writer: fix some bugs	2025-11-20 10:42:21 +00:00
Matthew Lugg	b05fefb9c9	std.http: stop assuming previous chunk state ... The full file may not be written, either due to a previous chunk being in-progress when `sendFile` was called, or due to `limit`.	2025-11-20 10:42:21 +00:00
Matthew Lugg	bc524a2b1a	std.Build: fix crashes running fuzz tests	2025-11-20 10:42:21 +00:00
Matthew Lugg	0f06b5b583	std.debug.MachOFile: handle 'path/to/archive.a(entry.o)' form	2025-11-20 10:42:21 +00:00
Matthew Lugg	e1fa4011fb	fuzz: hack around unknown module structure	2025-11-20 10:42:20 +00:00
Matthew Lugg	010dcd6a9b	fuzzer: account for runtime address slide ... This is relevant to PIEs, which are notably enabled by default on macOS. The build system needs to only see virtual addresses, that is, those which do not have the slide applied; but the fuzzer itself naturally sees relocated addresses (i.e. with the slide applied). We just need to subtract the slide when we communicate addresses to the build system.	2025-11-20 10:42:20 +00:00
Matthew Lugg	0a330d4f94	std.debug.Info: basic Mach-O support	2025-11-20 10:42:20 +00:00
Matthew Lugg	0caca625eb	std.debug: split up Mach-O debug info handling ... Like ELF, we now have `std.debug.MachOFile` for the host-independent parts, and `std.debug.SelfInfo.MachO` for logic requiring the file to correspond to the running program.	2025-11-20 10:42:20 +00:00
Jacob Young	7b325e08c9	Elf2: revert incorrect endian fix	2025-11-19 22:19:57 -05:00
Alex Rønne Petersen	f06adc70b0	ci: bump timeout for s390x-linux-debug to 5 hours on Forgejo Actions	2025-11-20 03:34:42 +01:00
Alex Rønne Petersen	c9afa901f3	ci: bump timeout for s390x-linux-release to 4 hours on Forgejo Actions	2025-11-19 23:47:58 +01:00
Alex Rønne Petersen	5078acf3a3	std.Io.net: disable listen on a unix socket, send bytes, receive bytes on Windows ... https://github.com/ziglang/zig/issues/25983	2025-11-19 21:51:57 +01:00
Alex Rønne Petersen	4ed5bb0bac	ci: bump aarch64-linux-debug timeout to 3 hours on Forgejo Actions	2025-11-19 20:45:36 +01:00
Harold	2f240d0819	std.Build.Step.Compile: add support for '-z defs' flag	2025-11-19 20:13:54 +01:00
Jan200101	bd832ed39a	std.Io.Threaded: add missing statx masks ... statx does not guarantee that the values requested by the mask be present and those not requested be absent which is why this worked.	2025-11-19 20:13:25 +01:00
Alex Rønne Petersen	43371cf388	Merge pull request #25965 from alexrp/s390x ... `s390x-linux` and general big-endian stuff	2025-11-19 19:52:18 +01:00
Matthew Lugg	806470b492	compiler: fix crash if file contents change during update ... When reporting a compile error, we would load the new file, but assume we could apply old AST/token indices (etc) to it, potentially causing crashes. Instead, if the file stat has changed since it was loaded, just emit an error that the file was modified mid-update.	2025-11-19 09:44:22 +00:00
Alex Rønne Petersen	abd05b3819	ci: enable s390x-linux jobs on Forgejo Actions	2025-11-19 09:41:58 +01:00
Alex Rønne Petersen	830831dcba	ci: add s390x-linux scripts	2025-11-19 09:41:55 +01:00
Alex Rønne Petersen	e11901c1a1	test: enable C ABI tests on s390x-linux	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	692c798303	test: disable a bunch of failing C ABI tests on s390x	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	ae0cc8c065	test: disable C ABI tests using i128 on s390x due to an LLVM crash ... https://github.com/llvm/llvm-project/issues/168460	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	a8e77b7a05	test: disable test-link on big-endian hosts ... https://github.com/ziglang/zig/issues/25961	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	a0e6d41331	test: remove complex arithmetic testing from c_compiler standalone test ... This has no business being here. Tests for our compiler-rt routines should be in compiler-rt, and tests for our C ABI compliance should be in `test-c-abi`.	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	41f7f3d4d5	test: disable an error trace test on optimized s390x-linux	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	0d104a645a	test: fix glibc_compat test for s390x which does not have local atexit	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	508f676bb4	std.os.linux.IoUring: disable bind/listen/connect on s390x ... https://github.com/ziglang/zig/issues/25956	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	e179335bee	std.zon.parse: disable zon vector on s390x ... https://github.com/ziglang/zig/issues/25957	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	c82884542f	langref: work around s390x LLVM compilation crash in test_defining_variadic_function ... https://github.com/ziglang/zig/issues/21350#issuecomment-3543006475	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	4dd05b5a64	langref: work around s390x LLVM miscompilation in runtime_shrExact_overflow ... https://github.com/ziglang/zig/issues/24304	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	04853f4052	std-docs: read/write messages as little endian	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	8817fc8958	incr-check: read/write messages as little endian	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	acfb88e9a5	std.Build.Step.CheckObject: make ELF reading endianness-aware	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	e09ba67161	std.Build.Step.Run: read/write messages as little endian	2025-11-19 01:42:45 +01:00
Matthew Lugg	0922990367	std.Build.Step: send messages to compiler as little-endian ... Little-endian is what `std.zig.Server` expects, but the old logic just send the raw bytes of the struct, so sent in native endian (causing a crash on big-endian targets).	2025-11-19 01:42:45 +01:00
Matthew Lugg	cd7d8dff26	std.zig.Server: read error bundle as little-endian ... Again, `std.zig.Server` expects little-endian. This is easy; we just use a `Reader.fixed` instead of directly `@ptrCast`ing data out of the buffer.	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	989e05d93b	stage1: update zig1.wasm ... Signed-off-by: Alex Rønne Petersen <alex@alexrp.com>	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	4b99e3718b	compiler: don't use self-hosted backends on big-endian hosts ... https://github.com/ziglang/zig/issues/25961	2025-11-19 01:42:45 +01:00
Alex Rønne Petersen	959a3612c2	aro: define arch macros for s390x	2025-11-19 01:42:45 +01:00
Matthew Lugg	a319211eee	llvm: fix lowering of packed struct constants ... The big-endian logic here was simply incorrect. Luckily, it was also overcomplicated; after calling `Value.writeToPackedMemory`, there is a method on `std.math.big.int.Mutable` which just does the correct endianness load for us.	2025-11-19 01:42:37 +01:00
Frank Denis	4ea4728084	Align ML-KEM code with ML-DSA (#25964) ... This will facilitate maintainance and code sharing between primitives.	2025-11-18 15:39:58 +00:00
Matthew Lugg	891f187032	cbe: fix big-endian unnatural integer bitcast ... Integers with padding bits on big-endian targets cannot quite be bitcast with a trivial memcpy, because the padding bits (which are zext or sext) are the most-significant, so are at the *lowest* addresses. So to bitcast to something which doesn't have padding bits, we need to offset past the padding. The logic I've added here definitely doesn't handle all possibilities correctly; I think that would actually be quite complicated. However, it handles a common case, and so prevents the Zig compiler itself from being miscompiled on big-endian targets (hence fixing a bootstrapping problem on big-endian).	2025-11-18 11:10:52 +01:00
Lukas Lalinsky	73f863a6fb	Fix AI/NI flag definitions for BSD systems ... Add missing AI flags for NetBSD and OpenBSD: - NetBSD: Add AI.SRV flag at bit 11 (0x00000800) - OpenBSD: Add AI.EXT flag at bit 3 and AI.FQDN flag at bit 5 Add missing NI (getnameinfo) flag definitions for all BSDs and Darwin: - FreeBSD/Haiku: NOFQDN, NUMERICHOST, NAMEREQD, NUMERICSERV, DGRAM, NUMERICSCOPE - DragonFly/NetBSD: Same flags with NUMERICSCOPE at bit 6 (deprecated WITHSCOPEID at bit 5 is skipped) - OpenBSD: NUMERICHOST, NUMERICSERV, NOFQDN, NAMEREQD, DGRAM (no NUMERICSCOPE) - macOS/Darwin: NOFQDN, NUMERICHOST, NAMEREQD, NUMERICSERV, DGRAM, NUMERICSCOPE at bit 8 (deprecated WITHSCOPEID at bit 5 is skipped) References: - NetBSD: https://github.com/NetBSD/src/blob/trunk/include/netdb.h - OpenBSD: https://github.com/openbsd/src/blob/master/include/netdb.h - FreeBSD: https://github.com/freebsd/freebsd-src/blob/master/include/netdb.h - DragonFly BSD: https://github.com/DragonFlyBSD/DragonFlyBSD/blob/master/include/netdb.h - Haiku: https://github.com/haiku/haiku/blob/master/headers/posix/netdb.h - macOS/Darwin: https://github.com/apple-oss-distributions/Libinfo/blob/main/lookup.subproj/netdb.h	2025-11-17 19:01:39 +01:00
Tobias Simetsreiter	f0a3df98d3	replace @panic with b.addFail in standalone test build.zig	2025-11-17 17:02:53 +01:00
xdBronch	d2b1aa48ab	sema: fix UAF in type mismatch error	2025-11-17 10:37:22 +02:00