hqemu: init at 2.5.2 #75705
hqemu: init at 2.5.2 #75705
Conversation
heavily based on existing qemu package
|
Looks like a qemu derivative work, which would be GPL licensed. Don't know if the content of |
|
That's a very good point - hadn't noticed that. Do we have a "license: unclear"? (also forgot this had approval and now I have committer permissions I could technically merge this myself, but won't until we have some clarity) |
|
I don't think there is |
|
The upstream project is basically an academic paper and it doesn't seem to really be maintained anymore. The license sounds like one of those "university ip policy" disclaimers and I don't see much hope in getting clarity from "upstream". |
|
@risicle do you still want to get this in? |
|
Well, the latest status is I'm unclear hqemu's license |
|
I marked this as stale due to inactivity. → More info |
stale
bot
added
the
2.status: stale
|
I think it'd be OK to add this package to NixOS, with the |
stale
bot
removed
the
2.status: stale
| else null) | ||
| }: | ||
|
|
||
| with stdenv.lib; |
There was a problem hiding this comment.
| with stdenv.lib; | |
| with lib; |
| qemu-system-i386 = "bin/qemu-system-i386"; | ||
| }; | ||
|
|
||
| meta = with stdenv.lib; { |
There was a problem hiding this comment.
| meta = with stdenv.lib; { | |
| meta = with lib; { |
|
TBH I think I'm going to close this PR because I don't want to have to maintain this, or even keep its list of known vulnerabilities up to date. I won't delete the branch, because this at least serves a purpose as a piece of software archaeology - a project that's tricky to build that at least now has a "known good" point in time that it can be reliably built. |
|
I marked this as stale due to inactivity. → More info |
stale
bot
added
the
2.status: stale
stale
bot
removed
the
2.status: stale
|
Closing because of lack of activity from the author. Please re-open if you plan to work on this. |
Motivation for this change
hqemu (http://csl.iis.sinica.edu.tw/hqemu/) is an interesting research project that grafts LLVM on to qemu's TCG binary translation, allowing it to optimize hotspots. I've managed to get some pretty good results from it, tripling execution speeds in some of my tests.
Package is heavily based on the qemu one, with some features removed and disabled for darwin - I get the feeling linux is the only host that has ever seen any attention, and my brief attempts at getting it to build on macos yielded nothing but pain.
Also note it has
knownVulnerabilitiesset because it's not as though this is a security-maintained application, so isn't really appropriate for production use.Things done
sandboxinnix.confon non-NixOS linux)nix-shell -p nix-review --run "nix-review wip"./result/bin/)nix path-info -Sbefore and after)Notify maintainers
cc @