The impossible happened, committing suicide

[Gunther Mayer]

Joshua Isom wrote:
> On May 19, 2008, at 10:16 AM, Gunther Mayer wrote:
>
>> Assertion failed: (0 && "The impossible happened, committing 
>> suicide"), function load_plist, file store_txt.c, line 840.
>> Abort trap
>>
>> Obviously a developer's joke, but I'm concerned that there might be a 
>> real problem. Would anybody here have any clue as to why this would 
>> occur?
>
> It looks like portsearch is where the assertion fails.  With a cursory 
> look, you have a bad plist file for some port and portsearch just dies 
> when it gets bad input.  The "best" thing to do would be to patch 
> portsearch to figure out what file is causing the problem, 
> `s->plist_fn`, and either delete it(assuming portsearch updates it, 
> since I seem to have few files starting with plist and it looks as 
> though it's trying to load a file named plist I'm assuming it's 
> controlled by portsearch) or just reinstall portsearch.
You're right, it is portsearch that's giving the problem, but neither 
reinstalling portsearch (upgrading actually) nor a regeneration of its 
files seem to work. When I do an "rm -rf /var/db/portsearch && 
portsearch -u" portsearch completes fine, but any subsequent invocations 
of "portsearch -u" fail with the same problem. It seems like it 
generates the same corrupt data over and over again and next time it 
runs it chokes on its own mess...
>
> But as I said, I just did a cursory look, and I don't use portsearch, 
> so I'm only looking at the source code.
Any portsearch users out there?

Gunther