Hubert Tong via llvm-dev
2019-Mar-20 22:46 UTC
[llvm-dev] DWARF debug info unit tests: Tests for wrong triple?
The DWARF debug info unit tests in llvm/unittests/DebugInfo/DWARF/ forms a generator based on the host triple. This is problematic for cross-compile configurations where LLVM does not fully support code generation for the host platform. I have found no rationale for why the host triple was chosen (ref: https://reviews.llvm.org/D27326?id=80002#inline-236026). I would imagine that it is more useful to test for the default target rather than the host. -- HT -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20190320/5b8abd12/attachment.html>
Eric Christopher via llvm-dev
2019-Mar-20 23:28 UTC
[llvm-dev] DWARF debug info unit tests: Tests for wrong triple?
Adding in aprantl here to make sure he's aware. Making sure I understand - you're building a compiler on machine a that llvm doesn't support code generation to and running the llvm tests on that platform? That's probably a bit out of what any of us expected here as the typical one is to use build llvm on a platform that it already supports. I do agree that default target would also work here as a suitable test. -eric On Wed, Mar 20, 2019 at 3:47 PM Hubert Tong via llvm-dev <llvm-dev at lists.llvm.org> wrote:> > The DWARF debug info unit tests in llvm/unittests/DebugInfo/DWARF/ forms a generator based on the host triple. This is problematic for cross-compile configurations where LLVM does not fully support code generation for the host platform. > > I have found no rationale for why the host triple was chosen (ref: https://reviews.llvm.org/D27326?id=80002#inline-236026). > > I would imagine that it is more useful to test for the default target rather than the host. > > -- HT > _______________________________________________ > LLVM Developers mailing list > llvm-dev at lists.llvm.org > https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
Adrian Prantl via llvm-dev
2019-Mar-20 23:32 UTC
[llvm-dev] DWARF debug info unit tests: Tests for wrong triple?
Am I right in assuming that the default target is the host triple for a typical non-cross compiler that most of us (or rather our bots) typically build? If yes then it sounds like changing the default triple to the default target would be a noop for most of us and thus a reasonable change? -- adrian> On Mar 20, 2019, at 4:28 PM, Eric Christopher <echristo at gmail.com> wrote: > > Adding in aprantl here to make sure he's aware. > > Making sure I understand - you're building a compiler on machine a > that llvm doesn't support code generation to and running the llvm > tests on that platform? > > That's probably a bit out of what any of us expected here as the > typical one is to use build llvm on a platform that it already > supports. I do agree that default target would also work here as a > suitable test. > > -eric > > On Wed, Mar 20, 2019 at 3:47 PM Hubert Tong via llvm-dev > <llvm-dev at lists.llvm.org> wrote: >> >> The DWARF debug info unit tests in llvm/unittests/DebugInfo/DWARF/ forms a generator based on the host triple. This is problematic for cross-compile configurations where LLVM does not fully support code generation for the host platform. >> >> I have found no rationale for why the host triple was chosen (ref: https://reviews.llvm.org/D27326?id=80002#inline-236026). >> >> I would imagine that it is more useful to test for the default target rather than the host. >> >> -- HT >> _______________________________________________ >> LLVM Developers mailing list >> llvm-dev at lists.llvm.org >> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
Hubert Tong via llvm-dev
2019-Mar-21 04:04 UTC
[llvm-dev] DWARF debug info unit tests: Tests for wrong triple?
On Wed, Mar 20, 2019 at 7:29 PM Eric Christopher <echristo at gmail.com> wrote:> Adding in aprantl here to make sure he's aware. > > Making sure I understand - you're building a compiler on machine a > that llvm doesn't support code generation to and running the llvm > tests on that platform? >Yes, or rather the support is in-progress, but the host platform should be good for hosting a cross-compiling LLVM toolchain.> > That's probably a bit out of what any of us expected here as the > typical one is to use build llvm on a platform that it already > supports. I do agree that default target would also work here as a > suitable test. > > -eric > > On Wed, Mar 20, 2019 at 3:47 PM Hubert Tong via llvm-dev > <llvm-dev at lists.llvm.org> wrote: > > > > The DWARF debug info unit tests in llvm/unittests/DebugInfo/DWARF/ forms > a generator based on the host triple. This is problematic for cross-compile > configurations where LLVM does not fully support code generation for the > host platform. > > > > I have found no rationale for why the host triple was chosen (ref: > https://reviews.llvm.org/D27326?id=80002#inline-236026). > > > > I would imagine that it is more useful to test for the default target > rather than the host. > > > > -- HT > > _______________________________________________ > > LLVM Developers mailing list > > llvm-dev at lists.llvm.org > > https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20190321/3fe30da4/attachment.html>