1 |
On Sunday 26 January 2014 19:18:55 Fabian Groffen wrote: |
2 |
> On 26-01-2014 16:48:33 +0100, Ruud Koolen wrote: |
3 |
> > Something went seriously wrong here: lots of patches are missing from the |
4 |
> > overlay. Bootstrapping with LATEST_TREE_YES=1 fails when building sed due |
5 |
> > to a missing sed-4.1.4-makeinfo-c-locale.patch, and bash runs into the |
6 |
> > same problem. |
7 |
> > |
8 |
> > I already fixed missing patches in sys-libs/readline, but found out |
9 |
> > immediately afterwards that the problem is more widespread. I don't know |
10 |
> > how widespread exactly, but you probably want to have a look at what |
11 |
> > exactly went wrong in the import process. |
12 |
> |
13 |
> Odd, I took the tree before your commits, and it happily compiles |
14 |
> without problems (nothing missing). That is, for me the Manifest wasn't |
15 |
> broken, and no patches appeared to be missing. Same for sed. |
16 |
|
17 |
Very strange. You can compile sed? I cannot compile sed due to an invalid |
18 |
Manifest on my vanilla prefix install right now (after an emerge --sync), and |
19 |
as I said it breaks for the same reasons when bootstrapping with |
20 |
LATEST_TREE_YES. amd64-linux, debian squeeze. |
21 |
|
22 |
Can you confirm, by manual inspection, that sed-4.1.4-makeinfo-c-locale.patch |
23 |
is mentioned in the Manifest but not present in the tree? |
24 |
|
25 |
> The patch you mention isn't supposed to be in the ebuilds as far as it |
26 |
> appears to me... ?!? |
27 |
|
28 |
Indeed, it looks like the sed patch could in fact be removed without a |
29 |
problem. The readline patch in my commit could not, though -- I think we can |
30 |
agree that |
31 |
http://prefix.gentooexperimental.org/hg/prefix-tree/file/ceb1b3177485/sys-libs/readline |
32 |
uses, but does not contain, 5.0-no_rpath.patch . |