440424 - auto upgrade from 1.6 to 1.6.1 installs bad binary

Status: RESOLVED DUPLICATE of bug 471922

(Camino Graveyard :: General, defect)

Description

[benc]

Background:

I have three systems that run Camino as the default browser. During the last week of March I noticed that there were site-availability problems, and also had problems getting a the upgrade to download (via the auto=upgrade feature).

One one system, an incomplete binary was loaded (apparently). 

STEPS:

Running Camino 1.6, say yes to an auto upgrade. The download runs, the application quits, and then the system dialog appears:

"The application Camino quit unexpectedly.

If you hit relaunch, it always fails. 

When I looked at the dock, I see a generic icon. When I show in finder, I see it is 26.8 MB (wow!).

"Report" shows:

Process:         Camino [47131]
Path:            /Applications/mozilla.org/Camino.app/Contents/MacOS/Camino
Identifier:      org.mozilla.camino
Version:         ??? (???)
Code Type:       PPC (Native)
Parent Process:  launchd [35339]

Date/Time:       2008-06-19 07:36:03.422 -0700
OS Version:      Mac OS X 10.5.3 (9D34)
Report Version:  6

Exception Type:  EXC_BREAKPOINT (SIGTRAP)
Exception Codes: 0x0000000000000001, 0x000000008fe0105c
Crashed Thread:  0

Dyld Error Message:
  Library not loaded: @executable_path/libxpcom.dylib
  Referenced from: /Applications/mozilla.org/Camino.app/Contents/MacOS/Camino
  Reason: image not found

--

So, my thinking is that the download failed in some nasty way that allowed upgrade to continue. I've searched bugzilla for information on the implementation, because I wanted to capture more files and info. I've gotten to the point now, where I really need my default browser to work (when I tell the Safari thing to change default browser to another verison of Camino, Mail still tries to run the broken version.

Comment 1

[Chris Lawson (gone)]

Is there anything in Console from around the time that happened? Sparkle, the update framework we're using, is pretty good about logging useful error messages.

In the meantime, why not just delete the bad binary and manually download a new one?

Comment 2

[Stuart Morgan]

The download couldn't have failed, because we validate the signature of our downloads; the chance of a corrupt download passing signature validation is effectively 0. Much more likely is that either the image mounted incorrectly somehow, or the copy died in the middle.

Comment 3

[benc]

Chris: I'll look (probably tomorrow morning...). I moved the binary today, but originally I left everything alone so I could study the scene of the crime, so to speak...

Stuart: Hmm. I have no idea how to confirm that, but what your saying makes sense (I was puzzled by this failure because I had assumed you had some download checking mechanism...)

Comment 4

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

benc, any update on this?

Comment 5

[Tyler Downer [He/Him]]

@Reporter, we have not heard back from you in a while, so I am closing this bug as INCOMPLETE. You can reopen this bug if more information becomes available. Some helpful information you can provide us is found at http://quality.mozilla.org/bug-writing-guidelines. You should also use a recent version of Firefox, from http://www.getfirefox.com.

Comment 6

[Samuel Sidler (old account; do not CC)]

Tyler: Please do *not* triage Camino bugs.

Comment 7

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

benc, do you still have the "bad" binary? We continue to see sporadic Talkback reports of crashes on startup after upgrading, and bug 471922 has also been filed since you originally reported this.

Comment 8

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

We've heard 0 reports of anything like this since 2.0, so given bug 502257 comment 4, duping to close this out.