* pass argv inputs directly to the config command line
* define each option's value type in the switch table so comma splitting
is applied only to list typed switches
Cache immutable section and relocation metadata before color refinement,
then reuse generated-tagged concurrent tables.
Parallelize grouping and fold marking.
Include associative COMDAT descendants in contribution hashes and
canonicalize relocation targets to selected COMDAT leaders.
Stop treating COMDAT section-definition symbols as symlink anchors.
Keep zero-sized COMDAT sections only when they have an owned symlink,
so referenced empty COMDAT symbols still resolve while marker sections
like .gsspr/.gssep are skipped.
Update COMDAT/ICF tests to cover zero-sized COMDAT relocation behavior
and the adjusted removed-COMDAT handling.
Previously, COMDAT symlinks were assigned as each object batch was loaded.
Weak symbol replacement could later change the selected leader, leaving
objects from earlier passes with stale COMDAT leader refs. Library search
does not need symlinks to be valid, so solve this by doing a single sweep
assignment of COMDAT leaders after a fixed point is found.
Add identical COMDAT folding using a color-refinement over foldable
read-only sections. Hash section contents, relocation targets, and target
colors until partitions stabilize, then redirect folded COMDAT symlinks
and discard duplicate sections.
Honor .llvm_addrsig via /LLVM_ADDRSIG so address-significant sections
are not folded, keep separate color space for code, unwind info, and
MSVC vftables, and update COMDAT symbol/section mapping after folding.
Add linker tests covering function folding, relocation-sensitive folds,
alignment, color-space separation, symlink chains, .llvm_addrsig,
pdata/xdata, and C++ ICF cases.
currently we tag pointers to solve ABA problem however if the list
has more than max_U16 items, the tag will overflow and cause serious issues.
To fix this, we need to use 64-bit tags together with 128-bit interlocked
compare-exchange intrinsic