IRC channel logs


back to list of logs

<alirio>civodul: I forgot to tell that '(system* "make" "mrproper" "headers_check")' delete the .config
<civodul>alirio: aah right, so it's not completely non-sensical
<civodul>it does make a difference
<civodul>i haven't yet try allyesconfig followed by allmodconfig
<civodul>(do i remember right?)
<alirio>localyesconfig will make a little working kernel
<alirio>so you can do a "grep -v 'is not set'" and concatenate after allmodconfig
<alirio>and oldconfig to merge
<civodul>so make localyesconfig && mv .config small && make allmodconfig && cat small >> .config && make oldconfig ?
<civodul>i'll try that
<civodul>unless you do ;-)
<alirio>with the grep part*
<alirio>but we will add a .config to patches?
<alirio>localyesconfig will do a different result everywhere
<civodul>you mean on each architecture?
<alirio>no, on each computer
<civodul>aah, because it depends on the hardware that's available, right?
<civodul>then we can't use that
<alirio>actually it depends on the code loaded, even if the hardware is available but the code to handle it isn't loaded it won't be included
<alirio>we can use allmodconfig instead of defconfig on the recipe and manually add the =y's from hardware where it doesn't work consulting localyesconfig output
<alirio>or copy some =y's from somewhere else (nixpkgs?), or handle errors case by case
<civodul>hmm, i really don't want to add a 1000 line .config to the repo
<alirio>to my ex-hardware it was 2 lines
<civodul>ah, well, if there are things that need to be loaded early for specific hardware, that can be done from the initrd
<civodul>no need to tweak the kernel config for that
<alirio>isn't harder to make initrd's?
<civodul>in our case it's OK
<civodul>(i'd say "simple" but i'm biased ;-))
*civodul attempts to move gcc's libdir to a separate output
<phant0mas>I think I am onto something regarding the headers from libpthread