Azure CI: Migrate from legacy MODEL=32mscoff to MODEL=32

Incl. fixing up the d_do_test.d tool to work properly with MODEL=32
on Windows.

And somewhat improve the documentation of MODEL=32omf, nowadays used
for the 32-bit OMF toolchain.
This commit is contained in:
Martin Kinkelin 2023-12-10 19:39:36 +01:00 committed by The Dlang Bot
parent 0c43ef8254
commit f3ad118c61
5 changed files with 34 additions and 24 deletions

View file

@ -13,7 +13,6 @@ jobs:
strategy:
matrix:
x64:
OS: Win_64
MODEL: 64
ARCH: x64
steps:
@ -29,11 +28,9 @@ jobs:
strategy:
matrix:
x64:
OS: Win_64
MODEL: 64
ARCH: x64
x86-OMF:
OS: Win_32
MODEL: 32omf
ARCH: x86
steps:
@ -52,13 +49,11 @@ jobs:
strategy:
matrix:
x64:
OS: Win_64
MODEL: 64
ARCH: x64
# x64 only because 32bit causes weird linker errors for several tests???
# x86-OMF:
# OS: Win_32
# MODEL: 32omf
# ARCH: x86
steps:
@ -75,17 +70,14 @@ jobs:
strategy:
matrix:
x64_Debug:
OS: Win_64
MODEL: 64
ARCH: x64
CONFIGURATION: Debug
x86-mscoff:
OS: Win_32
MODEL: 32mscoff
MODEL: 32
ARCH: x86
x86-mscoff_MinGW:
OS: Win_32
MODEL: 32mscoff
MODEL: 32
ARCH: x86
C_RUNTIME: mingw
steps: