This is the grub build for the efi system. it is named to egrub (like intel named it originally or to conform to elilo) the support for efi can't be added to the grub ebuild as the efi verison is build differently grub and egrub installs the same files so they should block each other Reproducible: Always
Created attachment 254273 [details] sys-boot/egrub/egrub-0.97.ebuild the grub based on grub-fedora the patchset that is applied is the patschset fron the grub package in the fedora fc14 source packge this patchset dont have any of the gentoo patches (extract the efi patches from the fedora verison and integrate into the gentoo patchset ?)
Created attachment 254275 [details, diff] sys-boot/egrub/files/grub-0.97-grub-set-default-1.patch add the missing set-grub-default script
Created attachment 254277 [details, diff] sys-boot/egrub/files/grub-1.98-efi_std_dir-1.patch set the default install dir for grub-install
Created attachment 254279 [details] sys-boot/egrub/egrub-1.98.ebuild the grub ebuild for the grub2
Created attachment 254283 [details] sys-boot/egrub/egrub-1.98.ebuild correct a arror
Created attachment 254409 [details] sys-boot/egrub/egrub-0.97.ebuild Correct some small typos that still where s/grub/egrub/ s/BOOT/EFI ....
Created attachment 254411 [details] sys-boot/egrub/egrub-1.98.ebuild correct some typos that still where s/grub/egrub/ s/BOOT/EFI/ ...
so, let me see if i understand you ... you're copying the whole grub ebuild and adding one or two flags to it so that it builds grub to support EFI ? sounds like this is a dupe of Bug 308385 and you should work on that. i'm not going to merge a dupe of another ebuild.
(for legacy grub) Did you just do a regex replace? - There is no mount-efi.eclass. - Why are you excluding ALL the rest of the patchseries and reverting the numbering to 1.0? - You've dropped all of the amd64 build stuff for the grub-static packaging or users that need a static grub regardless. - You've dropped all of the grub netboot support - why --without-curses? - the IA32_EMULATION warning is important on amd64. - the reiser4 patches break grub still. - your grub-set-default-1 patch seems either reversed or useless. Update ALL of this to be a patch implementing USE=efi on the existing grub legacy build please.
(In reply to comment #9) > (for legacy grub) > Did you just do a regex replace? no not that simple just corrected a wrong assumption on the fedora patchset > - There is no mount-efi.eclass. but i think should be plz follow the dend bug > - Why are you excluding ALL the rest of the patchseries and reverting the > numbering to 1.0? its not the gentoo patchset its the fedora patchset (the efi patches are so deep burried and depend on specific fedora patches that i was not able to extract it from the fedora patchset (i am working on it); so to get it working and have something for my next patch for genkernel) > - You've dropped all of the amd64 build stuff for the grub-static packaging or > users that need a static grub regardless. it dose not make sense as the fedora patch set compiles grub native in the efi version > - You've dropped all of the grub netboot support jep efi has its own netboot support its part of the specificaions efi-shell netboot > - why --without-curses? the efi patchset dose not support curses at the moment und will not ever just get autotools not detect it in efi mode > - the IA32_EMULATION warning is important on amd64. not for efi build its native an the 2 supported archs x86 amd64 > - the reiser4 patches break grub still. its not part of the fedora patchset > - your grub-set-default-1 patch seems either reversed or useless. for some ubscure reasons the fedora patchset removes this from the build > > > Update ALL of this to be a patch implementing USE=efi on the existing grub > legacy build please. >
(In reply to comment #8) > so, let me see if i understand you ... you're copying the whole grub ebuild and > adding one or two flags to it so that it builds grub to support EFI ? > > sounds like this is a dupe of Bug 308385 and you should work on that. i'm not > going to merge a dupe of another ebuild. > yep its right but the main reason for its own ebuild is that i also wanted to get mounting the efi partition right mount-efi the grub package only handles mounting /boot yes i can wrap everything into if use efi then inherit mount-efi ... else inherit mount-boot ... fi ... the main reason for the extra package at the moment is uefi/efi is only suported on x86 ia64 amd64 (itanium, macs, new hp ore dell notebooks, gigabyte motherboards, ...) so i put this in its own package but if you want it in the grub package nop will do my best to get it this way but extractiong the efi support from the fedora patchset would be a really hard work for the legacy grub
how Robbat wants to handle grub 1 is up to him ... maybe he will want a sep ebuild. but grub 2 isnt going to get its own. that's what Bug 308385 is for.
so i am closing it.
feel free to work on Bug 308385 then ... that needs love & attention
(In reply to comment #14) > feel free to work on Bug 308385 then ... that needs love & attention > yep its on my todo (real live needs some attention University) by the way one questions is it possible to wrap a inherit into a use flag ?
(In reply to comment #15) > by the way one questions is it possible to wrap a inherit into a use flag ? No, that would cause the global scope environment calculation to be different depending on the USE flag, which is prohibited.