Component
Guest Additions/Other
What problem are you facing?
I'm helping out with packaging Guest Additions for Fedora. There's a plan to get rid of YASM as it's unmaintained, and the Additions is one of the few packages using it right now. Straightforward build attempt with NASM instead of YASM fails due to the fact that not everything in the code is NASM-ready.
I have a patch proposal that builds and is tested to produce working Additions. I would appreciate your feedback on:
- Whether the approach I used makes sense from the technical standpoint, or if there are any better ones
- Whether you'd be interested in taking it upstream
How can we fix this?
Here is the main patch. Fundamentally, it simply wraps a few additional things for NASM vs. YASM, mostly replacing the verbatim wrt rip piece NASM doesn't support with an already-existing xWrtRIP define, appropriately adjusted for the NASM case (emptied out + default rel set instead).
It goes together with the RPM spec adjustment, setting appropriate options for kmk (DONT_USE_YASM=1 and TOOL_NASM_AS=nasm).
What alternatives or workarounds exist?
I can't think of anything else, and the changes are localized enough in my opinion, but I'd be grateful for any feedback on whether there's any better solution.
Anything else we should know?
As mentioned, I'm interested in your opinion on technical viability of this fix and whether any better options exist + your opinion on whether you'd be willing to accept this patch upstream.
Component
Guest Additions/Other
What problem are you facing?
I'm helping out with packaging Guest Additions for Fedora. There's a plan to get rid of YASM as it's unmaintained, and the Additions is one of the few packages using it right now. Straightforward build attempt with NASM instead of YASM fails due to the fact that not everything in the code is NASM-ready.
I have a patch proposal that builds and is tested to produce working Additions. I would appreciate your feedback on:
How can we fix this?
Here is the main patch. Fundamentally, it simply wraps a few additional things for NASM vs. YASM, mostly replacing the verbatim
wrt rippiece NASM doesn't support with an already-existingxWrtRIPdefine, appropriately adjusted for the NASM case (emptied out +default relset instead).It goes together with the RPM spec adjustment, setting appropriate options for
kmk(DONT_USE_YASM=1andTOOL_NASM_AS=nasm).What alternatives or workarounds exist?
I can't think of anything else, and the changes are localized enough in my opinion, but I'd be grateful for any feedback on whether there's any better solution.
Anything else we should know?
As mentioned, I'm interested in your opinion on technical viability of this fix and whether any better options exist + your opinion on whether you'd be willing to accept this patch upstream.