| 1 |
On Sat, Jun 07, 2008 at 06:29:08PM -0700, Zac Medico wrote: |
| 2 |
> felix@×××××××.com wrote: |
| 3 |
> > Pardon me if I am wrong, but I want this too, and to me the simplest |
| 4 |
> > analogy is to make. If some targets fail, make keeps on building the |
| 5 |
> > ones it can, until it is left with no buildable targets. I too have |
| 6 |
> > developed a wrapper using --resume --skipfirst, but that feels like a |
| 7 |
> > kluge. What I would like is an emerge option to generate a Makefile |
| 8 |
> > for output (-M looks available on a quick man check). Then instead of |
| 9 |
> > emerge -ptuvDN, I would use -MtuvDN, and run make at my leisure. I am |
| 10 |
> > incredibly naive about the inner workings of portage, but since it has |
| 11 |
> > dependencies, it does not seem entirely farfetched to generate a Makefile. |
| 12 |
> > |
| 13 |
> |
| 14 |
> Why not just take the features from make and add them to portage |
| 15 |
> itself? That way you can just treat the portage configurations as |
| 16 |
> the "Makefile". |
| 17 |
|
| 18 |
If you mean make the entire portage tree into one big Makefile, how do |
| 19 |
you represent sources which change to trigger making? Chaning USE |
| 20 |
flags is another trigger, and various options to emerge have their own |
| 21 |
effect on which builds are needed. |
| 22 |
|
| 23 |
If you mean imitate make's build process in portage, that seems like a |
| 24 |
lot of wheel reinventing. |
| 25 |
|
| 26 |
An advantage of generating Makefile output is it can replace the -p |
| 27 |
pretend option, and it lets you fine tune it by editing. For |
| 28 |
instance, my system right now wants to throw out mDNSmumble needed by |
| 29 |
KDE something in favor of avahi needed by the latest pidgin. So I |
| 30 |
save emerge output, edit away the avahi build and pidgin rebuild, and |
| 31 |
have a specific list to build. I cannot let emerge have its way; it |
| 32 |
blocks on avahi and mDNSmumble. A Makefile would have the same |
| 33 |
conflict, but make would ignore that part and do the rest, or I could |
| 34 |
edit out the conflict. |
| 35 |
|
| 36 |
-- |
| 37 |
... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._. |
| 38 |
Felix Finch: scarecrow repairman & rocket surgeon / felix@×××××××.com |
| 39 |
GPG = E987 4493 C860 246C 3B1E 6477 7838 76E9 182E 8151 ITAR license #4933 |
| 40 |
I've found a solution to Fermat's Last Theorem but I see I've run out of room o |
| 41 |
-- |
| 42 |
gentoo-portage-dev@l.g.o mailing list |
Archives