Project64 1.6 claims to have saved state successfully, without it actually working.
Gheerax IV
Recently it has mysteriously developed a problem: it will sometimes claim
a successful save when F5 is hit, but a subsequent hit of F7 will load a
much older save that should have been saved over *several* times since
then. Using the Load... menu item to view the save directory only shows
the one save, so the save slot didn't get rotated accidentally. Searching
the entire disk drive doesn't turn up any other save files.
This occurs with plenty of space free on every drive and every partition
and without any error message from PJ64 on save (such as "insufficient
disk space"). There's nothing else obvious that it could be, either.
Usual causes of saves failing:
* Not enough space. But there's gigs free on every partition.
* Disk damage. But there're no problems in other applications and the
drives are reasonably new. They don't make odd noises.
* Permissions not set right. But this happens when I play while logged in
as administrator. Also, I can get a successful save and then a silently-
unsuccessful save in the same session. It seems random. Permissions
don't change randomly and on the fly unless a) you're not the admin and
b) the admin is online at the same time and is jerking your chain.
Obviously inapplicable on a single-user machine, which this is.
* Something else (e.g. antivirus) is deleting or quarantining the file.
But then the file would disappear completely, not revert to an older
version of itself.
* Another user, with permission to modify the file, is fooling around
with the file, copying it and then overwriting the newer version with
the older one at random times. But, again, this is a single user
machine.
* The file is "in use". But nothing else should be locking Project 64's
savefiles except Project 64 and I'm not running two PJ64 instances
concurrently. Indexing service a) isn't supposed to lock files, for
obvious reasons (it would cause exactly these kinds of problems for
every application and frequently) and b) has been disabled on this
machine anyway as I find it makes search results have false negatives
AND slows the machine down, and actually slows down search rather than
speeds it up 9 times out of 10.
* The application contains a bug. But this started only this week or so,
and I've used PJ64 1.6 for years. There have been no updates or changes
to it or its plugins, though I have recently reinstalled it. Unless the
developers made changes to the downloadable binary, introducing a bug,
without changing the version number or otherwise disclosing that the 1.6
available from their web site today is not the 1.6 that was there a year
or two ago, which seems unlikely, then this can't be it.
* User finger slipped and PJ64 didn't actually register the F5 keypress.
But I saved at point A in the game, later saved at point B and
definitely saw "Saved ..." appear in the screen corner on that occasion,
reloaded about five minutes later, and found myself at point A. With
only the one file in the save directory.
* The file contents were altered by cosmic rays or some other non-
software means. But these means all produce effectively random
corruption of data. It's improbable that they'd be selective for a
single file on a bunch of disks with, collectively, over half a million
files, and googolplex-to-one-against that that file would have tens of
thousands or more of bits flipped in just such a way as to completely
scramble it *but yield a different, yet perfectly valid PJ64 savefile,
that actually closely resembles a game state I actually saved in hours
prior*.
Note again that if any of the above stopped PJ64 from successfully
writing a save it *should* alert the user with an error dialog, also, and
that's not happening.
So it seems like the evidence eliminates every single alternative
explanation, leaving none. What happened appears to be physically
impossible, yet it did nonetheless happen.
Anyone else here seen anything similar? Is there anything missing from
the above list of alternative explanations?
Gheerax IV
Well?
Mark McDougall
> Well?
Isn't the source available?
--
| Mark McDougall | "Electrical Engineers do it
| <http://members.iinet.net.au/~msmcdoug> | with less resistance!"
Gheerax IV
> On 30/08/2010 1:24 PM, Gheerax IV wrote:
>
>> Well?
>
> Isn't the source available?
Yes, but *I* haven't a clue how to interpret it! I leave that to experts
like the program's developers or anyone hanging out here who knows more
about emulators than I do.