Version 1.5.1 Crashing on Start

[GoogleCodeExporter]

This may be a repeat or a related issue to others I find here on this site in 
the issues list, but I don't see it exactly.  It is reported on the Apple App 
Store in Customer Reviews section.  

I am a fan and have very much enjoyed this on my iPod Touch.  I hope there 
could be a fix or a working 1.5 made available. 

What steps will reproduce the problem?
1. I just updated and now it will not start.
2. I only get the Frotz Screen and then it goes back to Home Screen.
3.

What is the expected output or behavior?  What do you see instead?
  To see the IF browser screen or an open IF.

What version of Frotz are you using?
  1.5.1

What device model are you using (iPhone 3GS, iPad, iPod Touch 3rd gen,
etc.)
  iPod Touch 2nd gen

What version of iOS?  3.2, 4.0, 4.3, etc.)
  4.2.1

Please provide any additional information below.
  Is the 1.5 version available somewhere?

Original issue reported on code.google.com by sonny...@gmail.com on 2 Apr 2011 at 1:46

[GoogleCodeExporter]

This sounds like the same issue I reported in 145. If you go into prefs and 
disable auto restore, you will get the story list but it will still crash if 
you try to start a story. 

Original comment by tony.mcc...@gmail.com on 2 Apr 2011 at 2:21

[GoogleCodeExporter]

Yes, that did it.

I briefly read that issue but I didn't realize the preferences were  
in the iPod Settings.  I thought they would be reached through the  
App itself.

Many thanks!

Original comment by sonny...@gmail.com on 2 Apr 2011 at 5:31

[GoogleCodeExporter]

Yep, still crashes when I begin a story.  At least I am getting the list.

Original comment by sonny...@gmail.com on 2 Apr 2011 at 5:36

[GoogleCodeExporter]

Unfortunately this is a hard one for me to diagnose because it isn't happening 
for me and seems to only happen on some devices, none of which I have.   Or 
it's only in iOS 4.2.1 perhaps, and Apple won't let you downgrade the OS.

If anyone in affected by the problem wants to help me test a solution, please 
send me your device ID (UDID) number and I will send you a beta to try as soon 
as I have one.  To find your UDID, connect the device to iTunes, bring up the 
Summary tab, and then in the top section on the last line where is says "Serial 
Number.", CLICK on the serial number and it will magically reveal your UDID.   
I don't think there's anything malicious anyone can do with a UDID, but just in 
case, mail it to me at ifrotz@gmail.com instead of posting it here.   

Thanks in advance to anyone who wants to help, and sorry for the inconvenience 
caused by the bug.  I wish it was easier to make apps run exactly the same on 
all devices and OS versions.

Original comment by spath...@gmail.com on 2 Apr 2011 at 6:11

[GoogleCodeExporter]

Issue 145 has been merged into this issue.

Original comment by spath...@gmail.com on 2 Apr 2011 at 6:12

[GoogleCodeExporter]

I think it might be 4.2.1 related as we have different hardware but older iOS. 
Of course it may also be ram related since both devices have lower ram. I'll 
send my ID next time I backup. I'll try to find some logs for you as well if 
they are generated. 
Tony.

Original comment by tony.mcc...@gmail.com on 2 Apr 2011 at 7:37

[GoogleCodeExporter]

Sending my UDID also.  Someday I'll have better hardware but it is great for on 
the train or sitting in a doctor's office.  Thanks to both of you!

Original comment by sonny...@gmail.com on 2 Apr 2011 at 8:39

[GoogleCodeExporter]

I have tried 3 different os' and i am still having this problem.

Original comment by LOGA...@gmail.com on 3 Apr 2011 at 12:50

[GoogleCodeExporter]

I also have the 2nd Gen. I-touch, could it be the hardware in the iPod?

Original comment by LOGA...@gmail.com on 3 Apr 2011 at 12:56

[GoogleCodeExporter]

@LOGA: What 3 diff OS versions did you try?
I do not have a 2ng gen iPT to try, but some people have reported problems with 
the 3G as well, so  I am installing 4.2.1 to see if it will happen for me on 
that device.

Original comment by spath...@gmail.com on 3 Apr 2011 at 12:58

[GoogleCodeExporter]

I have the same issue on a 1st gen iPhone with iOS 3.1.3.

Original comment by esteban....@gmail.com on 3 Apr 2011 at 5:16

[GoogleCodeExporter]

I have managed to reproduce this on my end finally on by procuring an iPhone 3G 
running 4.2.1.
It's looking like it may actually be a compiler bug in Apple's new LLVM-based 
compiler.
I'm still working on narrowing it down, but all I may have to do is recompile 
with gcc to make the problem go away.

I will still send betas to anyone who sends me a device UDID to try.  I'll 
probably get to this tonight or tomorrow.
The betas will have the additional bonus (provided they work for you!) of 
allowing direct downloads again.

Original comment by ifr...@gmail.com on 3 Apr 2011 at 5:31

[GoogleCodeExporter]

Issue 147 has been merged into this issue.

Original comment by ifr...@gmail.com on 3 Apr 2011 at 7:04

[GoogleCodeExporter]

@Spath: I tried 4.2.1, 4.1 and 4.0.1

Original comment by LOGA...@gmail.com on 3 Apr 2011 at 7:13

[GoogleCodeExporter]

@ifrotz who do I send my UDID to?

Original comment by LOGA...@gmail.com on 3 Apr 2011 at 11:15

[GoogleCodeExporter]

@LOGA:  ifrotz@gmail.com.  It was mentioned earlier in the thread, comment #4.

Original comment by ifr...@gmail.com on 4 Apr 2011 at 3:16

[GoogleCodeExporter]

I figured out exactly what's wrong, for anyone who may be interested and groks 
developerspeak:

It is indeed a bug in the clang 2.0/LLVM compiler, specifically in armv6 code 
generation, which is why it only affects older devices.  Nothing to do with the 
OS version at all; it's actually the hardware difference.  XCode, Apple's dev 
IDE, actually produces "fat" or "universal" binaries that have multiple 
variants for different processor sub-architectures, and the armv6 (newer models 
are armv7) supports fewer assembly instructions so the code generated is longer 
and, it turns out, buggy.  It's for some reason incorrectly adjusting the stack 
pointer in the function which lays out text (which is why the crash doesn't 
happen until you start a story), and then the epilog later tries to pop saved 
registers, gets the wrong saved pc and jumps to la-la land and crashes.

The fix is just to switch back to trusty GCC 4.2 for the build!   I'll file a 
bug report w/Apple, too.   They shouldn't even be accepting builds with clang 
yet if it's this unstable.  Maybe I can get an expedited review too since it's 
Apple's fault completely.  

I'll still send beta ipa's to those who sent me their UDIDs as a thank you for 
the help.
Btw, the betas aren't restricted from downloading games directly, so the 
install image will be a lot smaller than the App Store version because it 
doesn't need the IFDB bundled games.

Original comment by spath...@gmail.com on 4 Apr 2011 at 5:14

[GoogleCodeExporter]

Hi there,

I just wanted to add that I'm experiencing the above mentioned behaviour on an 
iphone 4 running ios4.3.1 and Frotz 1.5.1. I don't know if that helps your 
diagnosis, but I thought I'd at least throw my hat in the ring. At least you 
know there's at least one piece of up to date hardware/software that's 
experiencing this issue.

I updated both ios and Frotz in the last few days, and didn't try using Frotz 
between the two updates, so I'm not sure how to isolate the factor that caused 
this.

I'd like to take this opportunity to thank you for making and supporting this 
incredible app. I used to play Infocom games on the Apple IIe, so being able to 
play them once again on the iphone/ipad is a dream come true. Thanks!

Original comment by michaelb...@gmail.com on 4 Apr 2011 at 12:20

[GoogleCodeExporter]

@michaelb:   If you go to the global Preferences app and then under 
Applications -> Frotz, turn off autorestoring of games on launch, do you then 
make it to the Story List, but crash launching ANY story?  If your observed 
behavior isn't EXACTLY this, it may be a different problem.  Please clarify, 
and if there are any differences, file a new bug report.  Thanks.

Original comment by spath...@gmail.com on 4 Apr 2011 at 2:35

[GoogleCodeExporter]

[deleted comment]

[GoogleCodeExporter]

Hi again,

After switching off auto restore, it appears that my issue exactly matches bug 
report 142 rather than this one (146) as I previously thought. Sorry for the 
confusion. I will post in the 142 thread instead. Thanks again!

Original comment by michaelb...@gmail.com on 4 Apr 2011 at 11:38

[GoogleCodeExporter]

@michaelb: Thanks, that makes a lot more sense.  It would be very mysterious 
and disconcerting if you were really seeing this problem on an iPhone 4.

Original comment by spath...@gmail.com on 4 Apr 2011 at 11:45

[GoogleCodeExporter]

Thanks to everyone who reported and helped diagnose this problem.

FYI, I'm submitting the update to Apple today, so I'm no longer accepting UDIDs 
for beta testing.

Original comment by spath...@gmail.com on 5 Apr 2011 at 4:39

[GoogleCodeExporter]

Update:

I have filed a bug report to Apple on the compiler issue which caused this 
problem in Frotz (with a reproducible test case), submitted a Frotz update 
(1.5.2) for review, and officially requested an expedited review because of the 
severity of the problem for affected users.

Hopefully it will be approved and available in considerably less time than the 
usual 7-10 days.

Original comment by spath...@gmail.com on 6 Apr 2011 at 7:30

[GoogleCodeExporter]

Amazingly, Apple approved the "request for expedited review", and Frotz 1.5.2 
is now in the App Store with less than 24 hours review time.  

Woohoo!

(Of course, the email they sent me said basically "we are granting a one time 
exemption and don't expect this in the future.")

Original comment by spath...@gmail.com on 7 Apr 2011 at 2:28

• Changed state: Fixed

[GoogleCodeExporter]

Just a quick note to confirm that the new version in the store is indeed fixed.

Well done on the quick turnaround and thanks again for all your work on this - 
it's a fantastic app.

regards,
tony.

Original comment by tony.mcc...@gmail.com on 8 Apr 2011 at 11:51

[GoogleCodeExporter]

Thanks.

By the way, if anyone here wouldn't mind rating Frotz in iTunes, I'd really 
appreciate it.  It seems to take forever when a new version is published for 
enough ratings to develop for it to display them, and it bugs me.  :-)

Original comment by spath...@gmail.com on 8 Apr 2011 at 2:23