Upgrade frontend & libs to early v2.098.0 #3821
Conversation
Fixing ldc-developers#3520 and adding (revised) tests from ldc-developers#3517.
…tializer As the initializer might just as well not be computable at compile-time. Fixes such a case in `bug6985()` in dmd-testsuite's `compilable/interpret3.d`.
The linkage now (2.098) also affects the size of empty structs - e.g.,
1 for extern(C++), 0 for Posix extern(C), 4 for MSVC extern(C).
This also affects beauties like `extern(C) struct S { double[0] a; }`,
as tested by dmd-testsuite's runnable/ldc_cabi1.d, which is hereby
fixed for Posix - don't try to GEP into an empty LL struct.
Fixes dmd-testsuite's compilable/jsonNoOutFile.d.
…ive RangeErrors Fixing dmd-testsuite's runnable/testbounds.d.
|
Looking good, almost done. For x86, apparently just some importC stuff to be fixed (not sure I can be bothered at this early stage). Some remaining druntime regressions for Android, some usual compatibility trouble with ltsmaster, and druntime integration tests hanging on AArch64 (probably new tests using the concurrent GC). |
E.g., "456" typed as `char[4]*` encountered in dmd-testsuite's runnable/test22070_2.c, and a similar constant case in compilable/testcstuff1.c.
Instead of ICE'ing; e.g., for dmd-testsuite's runnable/cstuff2.c. See ldc-developers#3825.
Instead of ICE'ing; e.g., for dmd-testsuite's runnable/bitfields*.c. See ldc-developers#3825.
Due to a regression with Azure image 20210907.4 - unrelated to the v2.098 bump, also affecting master (verified via a manual CI run).
As e.g. the assert_fail test seems flaky.
E.g., preferring local imports for less merge conflict potential.
|
@kinke I just noticed that this did not update all files in |
|
I obviously can't rule out that I missed some parts, but v2.098 root/man.d hasn't been changed for a year: https://github.com/dlang/dmd/blob/v2.098.0/src/dmd/root/man.d |
Shouldn't git merge always update all the files, no way to miss it? Our |
|
It's not old, it just features |
No description provided.