1 |
On Monday, 16 September 2019 08:47:33 BST Peter Humphrey wrote: |
2 |
> On Sunday, 15 September 2019 22:35:05 BST Ian Zimmerman wrote: |
3 |
> > On 2019-09-15 16:21, Walter Dnes wrote: |
4 |
> > > On Sun, Sep 15, 2019 at 09:08:24AM -0700, Ian Zimmerman wrote |
5 |
> > > |
6 |
> > > > Is this for real or is it a mistake to be reverted soon? I do not |
7 |
> > > > enjoy the thought of rebuilding firefox twice in a row. |
8 |
> > > > |
9 |
> > > The USE flags seem to have been around for a while. I haven't updated |
10 |
> > > |
11 |
> > > my desktop for a few weeks (I'll do it tomorrow). "emerge -pv firefox" |
12 |
> > > gives firefox-60.8.0 with... |
13 |
> > > |
14 |
> > > [ebuild N ] www-client/firefox-60.8.0::gentoo USE="gmp-autoupdate |
15 |
> > > screenshot -bindist -clang -custom-cflags -custom-optimization -dbus |
16 |
> > > -debug -eme-free -geckodriver -hardened -hwaccel -jack (-neon) |
17 |
> > > -pulseaudio (-selinux) -startup-notification -system-harfbuzz |
18 |
> > > -system-icu -system-jpeg -system-libevent -system-libvpx |
19 |
> > > -system-sqlite -test -wifi" |
20 |
> > |
21 |
> > But this is what I mean. The _defaults_ for these flags have changed, |
22 |
> > so now (with version 68) the default is to build with the system libs. |
23 |
> > You'll see that too after you update. |
24 |
> > |
25 |
> > > This might not apply to Firefox, but it probably does. APIs change |
26 |
> > > with versions. By using the internal libs, you know that Firefox will |
27 |
> > > be using the library versions it's expecting to see. |
28 |
> > |
29 |
> > I trust the gentoo maintainers with that judgment - if it is in fact a |
30 |
> > judgment and not a typo. |
31 |
> |
32 |
> In my case, building with all those system libraries causes firefox to fail |
33 |
> on startup: it complains thus: |
34 |
> |
35 |
> $ firefox |
36 |
> 1568619005920 addons.manager ERROR Exception calling provider |
37 |
> GMPProvider.startup: [Exception... "Component returned failure code: |
38 |
> 0x80520012 (NS_ERROR_FILE_NOT_FOUND) [nsIStringBundle.GetStringFromName]" |
39 |
> nsresult: "0x80520012 (NS_ERROR_FILE_NOT_FOUND)" location: "JS frame :: |
40 |
> resource://gre/modules/addons/GMPProvider.jsm :: buildPluginList :: line |
41 |
> 844" data: no] Stack trace: |
42 |
> buildPluginList()@resource://gre/modules/addons/ GMPProvider.jsm:844 |
43 |
> startup()@resource://gre/modules/addons/GMPProvider.jsm:721 |
44 |
> callProvider()@resource://gre/modules/AddonManager.jsm:213 |
45 |
> _startProvider()@resource://gre/modules/AddonManager.jsm:649 |
46 |
> startup()@resource://gre/modules/AddonManager.jsm:873 |
47 |
> startup()@resource://gre/modules/AddonManager.jsm:3469 |
48 |
> observe()@resource://gre/modules/addonManager.js:70 |
49 |
> 1568619006132 Marionette FATAL <window id="main-window" |
50 |
> ^ |
51 |
> JavaScript error: resource://gre/modules/AutoCompletePopup.jsm, line 113: |
52 |
> NS_ERROR_ILLEGAL_VALUE: Component returned failure code: 0x80070057 |
53 |
> (NS_ERROR_ILLEGAL_VALUE) [nsIObserverService.removeObserver] |
54 |
> |
55 |
> Pro tem I've negated all the system-* flags in package.use. |
56 |
|
57 |
I've noticed a similar failure to launch on one system here. It works when |
58 |
launched with --safe-mode, so I assumed some of the many addons on this system |
59 |
are causing the failure. Other systems do not have such problems. |
60 |
|
61 |
-- |
62 |
Regards, |
63 |
|
64 |
Mick |