Link with libm dynamically (#19877)

added 1 commit

marked this merge request as ready

approved this merge request

assigned to @marge-bot

I will attempt to batch this MR (!6560 (closed))...

Batch MR !6560 (closed) failed: CI failed! I will retry later...

I will attempt to batch this MR (!6563 (closed))...

Batch MR !6563 (closed) failed: CI failed! I will retry later...

I will attempt to batch this MR (!6564 (closed))...

Batch MR !6564 (closed) failed: CI failed! I will retry later...

I will attempt to batch this MR (!6565 (closed))...

Batch MR !6565 (closed) failed: CI failed! I will retry later...

I will attempt to batch this MR (!6566 (closed))...

Batch MR !6566 (closed) failed: CI failed! I will retry later...

I will attempt to batch this MR (!6568 (closed))...

Batch MR !6568 (closed) failed: CI failed! I will retry later...

I will attempt to batch this MR (!6570 (closed))...

Batch MR !6570 (closed) failed: CI failed! I will retry later...

Should we note somewhere that HAS_LIBM should become an RTS-only define?

I don't understand

I grepped for where is HAS_LIBM used after this, and it seems to only be within the RTS now.

Oh, I'm not sure it's worth it. I guess nobody really cares as it's just to avoid linking on platforms that lack libm.

Well, I mean keeping notes for when we give RTS it's own configure.ac and want to trim down the main one as much as possible.

I also wonder whether it's necessary to GHC to have this logic vs just having the transitive dep via the rts, but we can figure that out later.

I will attempt to batch this MR (!6573 (closed))...

Batch MR !6573 (closed) failed: CI failed! I will retry later...

I will attempt to batch this MR (!6574 (closed))...

Batch MR !6574 (closed) failed: CI failed! I will retry later...

I will attempt to batch this MR (!6575 (closed))...

Batch MR !6575 (closed) failed: CI failed! I will retry later...

I will attempt to batch this MR (!6576 (closed))...

Merged in 92257abd

closed

Merge request reports