[llvm] [llvm-objcopy] Add --change-section-address (PR #98664)

[James Henderson via llvm-commits]

================
@@ -0,0 +1,120 @@
+# RUN: yaml2obj --docnum=1 %s -o %ti
+
+# RUN: llvm-objcopy --adjust-section-vma *+0x20 %ti %to
+# RUN: llvm-readelf --section-headers %to | FileCheck %s --check-prefix=CHECK-ADD-ALL
+
+# RUN: llvm-objcopy --change-section-address *+0x20 %ti %to
+# RUN: llvm-readelf --section-headers %to | FileCheck %s --check-prefix=CHECK-ADD-ALL
+
+# RUN: llvm-objcopy --change-section-address .anotherone-0x30 %ti %to
+# RUN: llvm-readelf --section-headers %to | FileCheck %s --check-prefix=CHECK-SUB-SECTION
+
+# RUN: llvm-objcopy --change-section-address .text*+0x20 %ti %to
+# RUN: llvm-readelf --section-headers %to | FileCheck %s --check-prefix=CHECK-ADD-PATTERN
+
+# RUN: llvm-objcopy --change-section-address .text*=0x10 %ti %to
+# RUN: llvm-readelf --section-headers %to | FileCheck %s --check-prefix=CHECK-SET-PATTERN
+
+# RUN: llvm-objcopy --change-section-address .anotherone-0x30  --change-section-address .anotherone+0x20 %ti %to
+# RUN: llvm-readelf --section-headers %to | FileCheck %s --check-prefix=CHECK-USE-LAST
+
+# RUN: llvm-objcopy --change-section-address .anotherone=0x455 --change-section-address *+0x20 %ti %to
+# RUN: llvm-readelf --section-headers %to | FileCheck %s --check-prefix=CHECK-NOTSUPERSET-SET
+
+# RUN: llvm-objcopy --change-section-address *+0x20 --change-section-address .anotherone=0x455 %ti %to
+# RUN: llvm-readelf --section-headers %to | FileCheck %s --check-prefix=CHECK-SUPERSET-SET
+
+# CHECK-ADD-ALL:          [Nr] Name              Type            Address          Off    Size   ES Flg Lk Inf Al
+# CHECK-ADD-ALL:               .text1
+# CHECK-ADD-ALL-SAME:                                            0000000000000120
----------------
jh7370 wrote:

I might be wrong, but I'd expect my suggestion to make it clearer what has actually failed, if the check did fail, because the check pattern reported would include both the section name and value, rather than just the latter, or worse something like ("match succeeded but not on the same line). As a minor point, it also removes the risk of the 120 matching another column value on the same line (hopefully not a real issue admittedly).

https://github.com/llvm/llvm-project/pull/98664