1 |
On Tuesday 22 July 2014 20:41:34 Bill Kenworthy wrote: |
2 |
> On 22/07/14 19:48, Dale wrote: |
3 |
> > Bill Kenworthy wrote: |
4 |
> >> On 22/07/14 19:03, Dale wrote: |
5 |
> >>> J. Roeleveld wrote: |
6 |
> >>>> On Tuesday, July 22, 2014 05:05:43 PM Bill Kenworthy wrote: |
7 |
> >>>>> I have a couple of systems with flash that are always a pain to update |
8 |
> >>>>> because the checksums fail so you have to manually force a manifest |
9 |
> >>>>> rebuild first. As I have to update them anyway, is there a ways to |
10 |
> >>>>> override the portage checksums and say install anyway? Because this |
11 |
> >>>>> package always fails anyway, I cant see any security gain by having a |
12 |
> >>>>> manual update every-time anyway. |
13 |
> >>>> |
14 |
> >>>> I would be more interested in finding out why it fails? |
15 |
> >>>> I use adobe flash myself and never experience a checksum issue with it. |
16 |
> >>>> |
17 |
> >>>> -- |
18 |
> >>>> Joost |
19 |
> >>>> |
20 |
> >>> Same here. I have it installed here and don't recall ever having a |
21 |
> >>> digest issue. It could be that something is off somewhere. If so, I'd |
22 |
> >>> rethink bypassing the checks. |
23 |
> >>> |
24 |
> >>> Dale |
25 |
> >>> |
26 |
> >> Hmm, that's interesting. |
27 |
> >> |
28 |
> >> Caused me to look closer ... I am pulling from http-replicator which |
29 |
> >> doesnt update the package if it cant see a name change (and adobe don't |
30 |
> >> change the name on the package - just the directory its pulled from) so |
31 |
> >> of course it fails checksum. |
32 |
> >> |
33 |
> >> Thanks for the hints to track this down. |
34 |
> >> |
35 |
> >> BillK |
36 |
> > |
37 |
> > Welcome. I wonder if http-replicator needs to check more than the |
38 |
> > name? I use it at times when I have more than one rig running and |
39 |
> > sounds like maybe it needs a new feature. |
40 |
> > |
41 |
> > Dale |
42 |
> > |
43 |
> The saving grace is that I have only seen the behaviour with this one |
44 |
> package so its something easily dealt with - now I know. Plus flash is |
45 |
> dieing so I might be able to do away with it before much longer - |
46 |
> unfortunately the OSS packages just are not as good. I've used |
47 |
> http-replicator for distfiles since it came out in ~2004 and its always |
48 |
> just worked. Oh well ... |
49 |
> |
50 |
> BillK |
51 |
|
52 |
I can't use any of the other packages because I use the BBC's radio streaming |
53 |
service every day, and none of them work with it (as far as I know). |
54 |
|
55 |
I have the same problem every time adobe-flash is updated. Last time it |
56 |
happened I had a conversation with the maintainer about it [1]. He said the |
57 |
problem was caused by Adobe's unconventional version numbering (which sounds |
58 |
like the same thing as Bill found), together with any caching proxy in |
59 |
between. That's http-replicator in my case too. Solved with wget --no-proxy. |
60 |
Or I suppose just deleting the tarball from the proxy's cache should do it. |
61 |
|
62 |
Forcing a re-manifest is not the thing to do, as that would just lead to |
63 |
reinstalling the version you have already. |
64 |
|
65 |
[1] https://bugs.gentoo.org/show_bug.cgi?id=509874 |
66 |
|
67 |
-- |
68 |
Regards |
69 |
Peter |