f39c399d9d
The exit code for -### is inconsistent. Unrecognized options lead to exit code 1, as expected. However, most others errors (including invalid option value) lead to exit code 0, differing from GCC and most utilities. This is a longstanding quirk of -###, and we didn't fix it because many driver tests need adjustment. Change -### to be similar to -fdriver-only -v and exit with code 1. This requires fixing many driver tests, but the end result gives us stronger tests. * Existing `RUN: %clang -### ...` tests usually don't use `CHECK-NOT: error:` or `--implicit-check-not=error:`. If a change introduces an error, such a change usually cannot be detected. * Many folks contributing new tests don't know `-fdriver-only -v`. To test no driver error/warning for new tests, they can use the familiar `-### -Werror` instead of `-fdriver-only -v -Werror`. An incomplete list of prerequisite test improvement: *2f79bb1046: add -nogpulib to some AMDGPU tests *9155e517e6: add --cuda-path= (test w/ and w/o /usr/local/cuda) *80765ede5b: -mcpu=native may return either 0 or 1, depending on whether `--target=` specifies a native target *abae53f43f: fix -fuse-ld=lld misuses (test w/o and w/o /usr/local/bin/ld.lld) *ab68df505e: add -resource-dir= and -fvisibility=hidden to some -fsanitize=cfi tests *d5ca1602f6: --rtlib=platform without --unwindlib= may fail if CLANG_DEFAULT_UNWINDLIB=unwindlib Reviewed By: jhuber6, yaxunl, dblaikie Differential Revision: https://reviews.llvm.org/D156363
10 lines
434 B
C
10 lines
434 B
C
// RUN: %clang -### -c %s -fmessage-length=80 2>&1 | FileCheck %s
|
|
// CHECK: "-fmessage-length=80"
|
|
|
|
/// Omit -fmessage-length=0 to simplify common CC1 command lines.
|
|
// RUN: %clang -### -c %s -fmessage-length=0 2>&1 | FileCheck --check-prefix=ZERO %s
|
|
// ZERO-NOT: "-fmessage-length=0"
|
|
|
|
// RUN: not %clang -### -c %s -fmessage-length=nan 2>&1 | FileCheck --check-prefix=ERR %s
|
|
// ERR: error: invalid argument 'nan' to -fmessage-length=
|