Bug 190007 – [PATCH] bootstrap-prefix.sh does pass MAKEOPTS to all make calls

• Status: RESOLVED FIXED
• Product: Gentoo/Alt

[Description Stefan Hoelldampf 2007-08-24 08:10:28 UTC]

The attached patch adds ${MAKEOPTS} to all make calls.

[Comment 1 Stefan Hoelldampf 2007-08-24 08:11:04 UTC]

Created attachment 129049 [details, diff]
bootstrap-prefix.sh.MAKEOPTS.patch

[Comment 2 Stefan Hoelldampf 2007-09-27 10:50:56 UTC]

ping

[Comment 3 Fabian Groffen 2007-09-27 13:24:41 UTC]

what if you insert <random crap> into MAKEOPTS and bootstrapping fails due to that?  I'm a bit hesitant here because it opens up opportunities for new random failures during bootstrapping...

Also, I'm not sure if all packages can be built with -j2 for instance, meaning suddenly that needs to be tested.  If you really want it, you better hack around by setting MAKE to "make -j2" I guess...

haubi: opinions here?

[Comment 4 Stefan Hoelldampf 2007-09-27 17:34:43 UTC]

And what about bootstrap_gcc() and bootstrap_python()? Is it intended that they already use ${MAKEOPTS}?

[Comment 5 Fabian Groffen 2007-09-28 14:51:49 UTC]

if they use it, then that's a bug.

[Comment 6 Fabian Groffen 2007-10-08 18:20:06 UTC]

Ok, using both makes no sense.  For GCC/python it really makes a difference and I'd like to keep it in.  So I applied your patch for consistency.

Thanks!

[Comment 7 Jeremy Olexa (darkside) (RETIRED) 2007-10-15 15:20:50 UTC]

I ran into a bug since you added this in. 

Situation: bootstrapping make with a non-GNU make will error if the native make doesn't understand -j2.

Recommended fix: Check to see if the make is GNU make before setting MAKEOPTS. Thoughts? 

(I would like to reopen this bug but I can't)

[Comment 8 Fabian Groffen 2007-10-15 15:39:19 UTC]

Then I'll close/wontfix it asap, as I just forgot to remove that default of -j2.  Only people that have an idea what they're doing should mess with that variable.