On 11/5/25 11:42 PM, Peter0x44 wrote:
Nov 5, 2025 11:40:46 PM Jonathan Yong <[email protected]>:
On 10/31/25 4:28 PM, LIU Hao wrote:
在 2025-10-31 19:53, Richard Biener 写道:
On Fri, Oct 31, 2025 at 12:34 PM Peter0x44 <[email protected]> wrote:
On 2025-10-31 08:40, Richard Biener wrote:
On Fri, Oct 31, 2025 at 2:25 AM Peter Damianov <[email protected]>
wrote:
…
Thanks for working on this! Can you state how you tested this? In
particular,
does LTO now work with BigObj objects (and regular COFF objects still
work
with LTO)?
Thanks,
Richard.
I tested this by linking some programs with LTO using both -Wa,-mbig-obj
and without it.
They all worked.
I'm also building a w64devkit with this patch:
https://github.com/Peter0x44/w64devkit/
commit/42452901a1eb31768d3c27d824e8a741125f4c04
w64devkit patches binutils to always emit bigobj, so this should
exercise it well.
Afterwards, I can test if it functions on the windows host too.
The patch is OK. Please leave the maintainers you CCed the chance
to chime in before pushing though (until early next week).
I have bootstrapped GCC on {i686,x86_64}-w64-mingw32 with `-Wa,-mbig- obj` in
{C,CXX}FLAGS, and tried building some C and C++ test programs with `-flto
-Wa,-mbig-obj` (which failed without this patch), and seen no issue so far.
I will push soon if there are no more comments.
The commit message itself needed a change to mention the PR number. Otherwise
it's fine. Maybe I should try to get approved to push commits myself?
Feel free to resend a v2 with an updated message.
