Re: [RFCv4] install: Create a GC root during install-grub.

2016-03-13 Thread Jookia
On Sun, Mar 13, 2016 at 10:44:38PM +0100, Ludovic Courtès wrote: > Diffs to patches would be hard to read, but if you could simply add a > cover letter with a short summary of the changes compared to previous > versions, that’d be perfect. :-) I think for an RFC at least it'd be best for me to st

Re: [RFCv4] install: Create a GC root during install-grub.

2016-03-13 Thread Ludovic Courtès
Jookia <166...@gmail.com> skribis: > On Fri, Mar 11, 2016 at 03:48:38PM +0100, Ludovic Courtès wrote: >> Jookia <166...@gmail.com> skribis: >> >> > While previously creating a GC root for GRUB's resources was the caller's >> > responsibility, it's much less repetitive to put it in install-grub no

Re: [RFCv4] install: Create a GC root during install-grub.

2016-03-11 Thread Jookia
On Fri, Mar 11, 2016 at 03:48:38PM +0100, Ludovic Courtès wrote: > Jookia <166...@gmail.com> skribis: > > > While previously creating a GC root for GRUB's resources was the caller's > > responsibility, it's much less repetitive to put it in install-grub now that > > it's wrapped by error handling.

Re: [RFCv4] install: Create a GC root during install-grub.

2016-03-11 Thread Ludovic Courtès
Jookia <166...@gmail.com> skribis: > While previously creating a GC root for GRUB's resources was the caller's > responsibility, it's much less repetitive to put it in install-grub now that > it's wrapped by error handling. This also means we can replace the > install-grub* > function with a smal

[RFCv4] install: Create a GC root during install-grub.

2016-03-10 Thread Jookia
While previously creating a GC root for GRUB's resources was the caller's responsibility, it's much less repetitive to put it in install-grub now that it's wrapped by error handling. This also means we can replace the install-grub* function with a small definition inside perform-action named 'insta