Does JPE Discovery Mode not work on UPX'd exe ?

Message

grannyGeek · #1 Post by **grannyGeek** » Mon Sep 17, 2007 11:45 pm

I get error message "executable not available".

Just checking before I start beating my head against a wall of unknown causes.

Thanks for any info.
gG

redllar · #2 Post by **redllar** » Tue Sep 18, 2007 8:48 pm

I get error message "executable not available".

This sounds like the "launched app is not on the drive that jpe is on" bug. Open up the properties of the launchpad item and see if the Path field starts with a question mark. If it does then replace it with the drive it should be and then OK out of the dialog box and you should be good to go.

To answer your question, AFAIK JPE hasn't had any problems with UPX'd apps, running in discovery mode or otherwise.

grannyGeek · #3 Post by **grannyGeek** » Tue Sep 18, 2007 10:09 pm

Yes, the path was what caused the error, thank you much.

However, I found a new glitch. If anybody else can reproduce this, you could maybe file it away for future reference if you find it interesting.
Using JPE v 015 in Discovery Mode, I can successfully open and run the app. But upon closing, get message
"PDFXCview.exe - application error
The instruction at "0x7c90256c" referenced memory at "0x0000000". The memory could not be read."

Likewise with PeaZip. Same address listed in the message.
This happened with multiple attempts.

The discovery ini was created just fine though, so tha

PDFxc I upx'd myself. PeaZip arrived already packed.

I don't have UPX available on this pc, so I can't pack any other exe files to test again. I will try some others tonight at home.
This might just be a coincidence, though.

Maybe you should just hit me upside the head with a 2x4 and tell me to stay off the computer for a month.

Absolutely not an urgent situation, and only posted as FYI in case you find it interesting.

crownixx · #4 Post by **crownixx** » Wed Sep 19, 2007 12:38 am

"PDFXCview.exe - application error
The instruction at "0x7c90256c" referenced memory at "0x0000000". The memory could not be read."

I also got that error message last time when testing google talk with 0161 version. shutdown my pc, then the next day without touching any runtime ini settings or change the version, suddenly the error disapeared. Now i dont have those error message anymore for more testing..

redllar · #5 Post by **redllar** » Wed Sep 19, 2007 8:38 am

That's most likely caused by the api hooks being unhooked at app-end in a multi-threaded app. It only occurs at the tail-end of the app closing down, so it shouldn't cause any problems re: file damage, etc.

The 015, 0160, 0161, and now 0162, jpe dlls all have different attempts I made at stopping the crash, but it's a sensitive situation that teeters between deadlocking and crashing. Hopefully one of these versions I'll find the perfect balance.

grannyGeek · #6 Post by **grannyGeek** » Wed Sep 19, 2007 6:15 pm

A newbish observation -- dual core machines seem to have no problem.
My home pc is dual-core, XP Pro SP2, and I get no error.

Work pcs are not dual core, also on XP Pro SP2, and I always get the "crash" upon shut-down of the packed apps in Discovery Mode. But as you said, no damage or corruption, so it's all good.

redllar · #7 Post by **redllar** » Wed Sep 19, 2007 7:33 pm

Thanks for the info gG. I just uploaded an 0163 to try and address the app-end crashes. The download link's at http://portablefreeware.com/forums/viewtopic.php?t=2176 if you'd like to try it out.

BTW, I'm at the other end of the spectrum pc-wise. Mine's probably too slow to run into the problem and yours is probably too fast. Wanna trade? :)

grannyGeek · #8 Post by **grannyGeek** » Wed Sep 19, 2007 8:23 pm

great, a new jpe version.

I'll trade pc's with you if you throw in your programming knowledge.
On the other hand, mine comes with a lazy tomcat who sleeps on top of the warm monitor.
Still wanna trade?