[PPC] Minefield won't load any sites using latest trunk nightly

RESOLVED WORKSFORME

Status

()

Toolkit
Safe Browsing
--
critical
RESOLVED WORKSFORME
7 years ago
3 years ago

People

(Reporter: marcia, Unassigned)

Tracking

({regression, relnote})

Trunk
PowerPC
Mac OS X
regression, relnote
Points:
---
Dependency tree / graph

Firefox Tracking Flags

(blocking2.0 -)

Details

(Reporter)

Description

7 years ago
Seen while running Mozilla/5.0 (Macintosh; PPC Mac OS X 10.5; rv:2.0b4pre) Gecko/20100816 Minefield/4.0b4pre

STR:
1. Try to load any web site.
2. Receive this error: Error: uncaught exception: [Exception... "Component returned failure code: 0x80570016 (NS_ERROR_XPC_GS_RETURNED_FAILURE) [nsIJSCID.getService]"  nsresult: "0x80570016 (NS_ERROR_XPC_GS_RETURNED_FAILURE)"  location: "JS frame :: chrome://browser/content/safebrowsing/sb-loader.js :: anonymous :: line 48"  data: no]

Although I believe I have heard the PPC has an uncertain future, would it be weird if this was broken in B4 yet worked in B3?
(Reporter)

Comment 1

7 years ago
I also found I cannot create a new profile on that machine as well, so I could only test with an existing profile that I had that worked fine on Friday.
Summary: Minefield won't load any sites using latest trunk nightly → [PPC] Minefield won't load any sites using latest trunk nightly
(Reporter)

Updated

7 years ago
Keywords: relnote
(Reporter)

Comment 2

7 years ago
I thought I nominated this yesterday, but doing so now so release drivers can evaluate it.
(Reporter)

Comment 3

7 years ago
Error: (void 0) is undefined
Source File: file:///Users/mozilla/Desktop/Minefield.app/Contents/MacOS/components/nsPlacesAutoComplete.js
Line: 842 is the other error that shows up consistently in the console using the latest nightly from 8/17.

Updated

7 years ago
Hardware: x86 → PowerPC

Comment 4

7 years ago
(In reply to comment #0)
> Although I believe I have heard the PPC has an uncertain future, would it be
> weird if this was broken in B4 yet worked in B3?

That would not be strange as there are tm-merges during each beta release. We might want to cc some JS folks.
(Reporter)

Updated

7 years ago
blocking2.0: --- → ?

Comment 5

7 years ago
I bet this is from YARR. sigh.

Comment 6

7 years ago
Worryingly, this means we have precisely zero automated coverage of PPC. So, we really have no idea what we are shipping to those users.
I really don't think we should ship PPC to users for FF4; I think we're just waiting for the x64 stuff to be ready, so that we can switch to x32/x64 from PPC/x32.  I will...investigate that!
(In reply to comment #6)
> Worryingly, this means we have precisely zero automated coverage of PPC. So, we
> really have no idea what we are shipping to those users.

As usual, automated PPC unittest runs are on http://tinderbox.mozilla.org/showbuilds.cgi?tree=GeriatricMachines. Not sure what specific test suite would have shown this problem, but all the unittests are being run as fast as these old machines will allow.

Let me know if you need further details.
Blocks: 588412
(In reply to comment #8)
> As usual, automated PPC unittest runs are on
> http://tinderbox.mozilla.org/showbuilds.cgi?tree=GeriatricMachines. Not sure
> what specific test suite would have shown this problem, but all the unittests
> are being run as fast as these old machines will allow.

Well, if the builds didn't start up, I would have expected the tests to go red as they wouldn't have been able to start. Those tests aren't reported to the main Tinderbox, is what I think Rob is saying, so we're not paying attention to them.

Good to know they're there.

I filed bug 588412 on ensuring we don't offer Beta 4 updates to PPC users.

Comment 10

7 years ago
See also bug 587771.
We have decided to drop support for PPC/OSX for Firefox 4, so this doesn't block.
blocking2.0: ? → -
(In reply to comment #11)
> We have decided to drop support for PPC/OSX for Firefox 4, so this doesn't
> block.

Mike is there a plan to announce that more widely ?

Comment 13

7 years ago
(In reply to comment #12)
> Mike is there a plan to announce that more widely ?

Especially since the patch in bug 587612 might fix the issue here as well?
This has now been announced more widely: http://wiki.mozilla.org/Firefox/4/Platforms

Updated

7 years ago
Blocks: 587771

Updated

7 years ago
Blocks: 600965

Comment 15

7 years ago
Strange that this bug is in safebrowsing as apparently we are seeing this in SeaMonkey as well.
(Reporter)

Comment 16

7 years ago
Now that this has been announced, can we close out this bug?
This bug isn't a blocker, but if someone comes along with a patch that fixes it or its dependencies we would probably take them.
Did this get resolved by bug 587612?

Updated

7 years ago
Duplicate of this bug: 611980
Yes, it was fixed by bug 587612 (G5 running 4.0b8pre).

Comment 21

7 years ago
(In reply to comment #14)
> This [drop of PPC support] has now been announced more widely:
> http://wiki.mozilla.org/Firefox/4/Platforms

This says “universal binary” and on the other hand “CPU architecture: i386, x86-64”. That's a contradiction!
(In reply to comment #21)
> (In reply to comment #14)
> > This [drop of PPC support] has now been announced more widely:
> > http://wiki.mozilla.org/Firefox/4/Platforms
> 
> This says “universal binary” and on the other hand “CPU architecture: i386,
> x86-64”. That's a contradiction!

from https://secure.wikimedia.org/wikipedia/en/wiki/Universal_binary

With the release of Mac OS X Snow Leopard and the move to 64-bit architecture, Universal can also mean an application that runs natively on i386 (32-bit Intel) and x86_64 systems.

Comment 23

7 years ago
This bug is listed as a known issue on the RC, but comments seem to indicate that this might be fixed. Could anyone confirm either way? We don't want to claim we don't have PPC support if we do.

Comment 24

7 years ago
(In reply to comment #23)
> This bug is listed as a known issue on the RC, but comments seem to indicate
> that this might be fixed. Could anyone confirm either way? We don't want to
> claim we don't have PPC support if we do.

See comment #11. Also, the RC binary is an i386/x86_64 universal build, so it won't run on a PPC machine.

Comment 25

7 years ago
I am aware that the current Mac-build is incompatible with PPC. but Comment 17 made it sound like a PPC version would be offered in some form if this bug was fixed. And Comment 19 says it was.

Perhaps he only meant that it might be fixed for a contributed build?
If you mean me (comment 20), I'm referring to both the El Furbe PPC builds and TenFourFox, both of which do run (this post is from TenFourFox on an iBook G4).

Comment 27

7 years ago
(In reply to comment #26)

Yes, I meant you. Sorry. So my question becomes: Would it be possible for either of those builds to be under "contributed builds"? It would seem to be the same type of thing as allowing Solaris builds there.

If not, I understand, and thank Mr. Kaiser for telling me where I can get a PPC build. I have been "spreading the gospel" about Firefox 4, and now I know I can share with my friends who still use PPC-based Macs.
marking wfm based on comment#20
Status: NEW → RESOLVED
Last Resolved: 5 years ago
Resolution: --- → WORKSFORME

Comment 29

5 years ago
Note for the future: this bug having been fixed doesn't mean that official builds of Firefox run on PPC (e.g. G3, G4, G5) Macs. TenFourFox is the only way to use as much of the latest Firefox code as possible, and, even there, current builds no longer allow the use of plugins.
(Assignee)

Updated

3 years ago
Component: Phishing Protection → Phishing Protection
Product: Firefox → Toolkit
You need to log in before you can comment on or make changes to this bug.