Date: Sat, 02 Oct 2021 14:50:07 +0000 From: bugzilla-noreply@freebsd.org To: ports-bugs@FreeBSD.org Subject: [Bug 258846] devel/dyncall: unbreak on arm Message-ID: <bug-258846-7788-IY3ZPIfFa6@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-258846-7788@https.bugs.freebsd.org/bugzilla/> References: <bug-258846-7788@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D258846 --- Comment #6 from Tassilo Philipp <tphilipp@potion-studios.com> --- Alright, I looked at it and ran some tests on 4 different arm platforms, an= d I cannot accept it as-is, but it's close. The main issue is it only unbreaks = arm w/ the clang integrated assembler when not using THUMB (in other words, the file you pointed out is not the only one to patch). I'll come up with a new patch. Details and/or other things while looking at it: - the %-prefixes were probably used b/c of x86 habits, they should be remov= ed indeed, thanks! - this is not the first time we have issues with clang's integrated assembl= er, we needed to change some PPC code also for v1.2 to make it happy - about the UAL/.syntax unified: this doesn't actually help for dyncall, b/c the ARM and THUMB calling conventions differ enough to need their own implementations, anyways.... yes, it would probably also be harmless to use those instruction names, but pre-UAL toolchains might break (and we do actu= ally use dyncall on some of those); I'm surprised that you listed it under "issu= es that prevent it from building on armv6/armv7", b/c I cannot reproduce that.= .. - depending on further tests, until v1.3 is out with the real fix, I might = make use of clang's -no-integrated-as to unbreak the build for the FreeBSD port.= .. I'll have a clearer picture, soon - I don't think I'll add TEST_TARGET, for already pointed out reasons; I understand that this isn't run automatically for the pkg building process, = but there is no point in running `make test` to only build (but not run) the te= sts - the target you probably wanted isn't 'tests' but 'run-tests', and for tho= se I should fix some exit codes, first; as said, this is on my to do list for dyncall 1.3 More to come, soon, thanks again for bringing this to my attention :) --=20 You are receiving this mail because: You are the assignee for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-258846-7788-IY3ZPIfFa6>