Issue 894 in zxing: Crash on some Deutsche Post Datamatrix

[zx...@googlecode.com]

Status: New
Owner: ----

New issue 894 by maxr...@gmail.com: Crash on some Deutsche Post Datamatrix
http://code.google.com/p/zxing/issues/detail?id=894

What steps will reproduce the problem?
1.
Scan http://www.doppelnutzen.de/layout/premiumadress_muster_high.gif
2.
Crash.

What is the expected output? What do you see instead?
Hard to tell... I can see it.

What version of the product are you using? On what operating system?
Android 2.3.4, Nexus S GSM

Please provide any additional information below.
Those Datamatices printed by Deutsche Post often include tons of
non-printable characters.

[zx...@googlecode.com]

Updates:
Cc: dswit...@google.com

Comment #1 on issue 894 by sro...@gmail.com: Crash on some Deutsche Post
Datamatrix
http://code.google.com/p/zxing/issues/detail?id=894

That's weird. The crash is due to a verify error in the bytecode. This
means there was a problem with the Proguard optimization. But I can't
reproduce when building from the latest source code. Daniel, I guess just
make sure you build with the latest Proguard too next time there's a
release? It's 3.6 now.

That fixes the crash at least but it doesn't decode yet. Let me look
further.

[zx...@googlecode.com]

Comment #2 on issue 894 by maxr...@gmail.com: Crash on some Deutsche Post
Datamatrix
http://code.google.com/p/zxing/issues/detail?id=894

Actually this could be a duplicate of 882.

[zx...@googlecode.com]

Comment #3 on issue 894 by sro...@gmail.com: Crash on some Deutsche Post
Datamatrix
http://code.google.com/p/zxing/issues/detail?id=894

It's not the same issue -- that one did not involve a crash, and was
detected.

[zx...@googlecode.com]

Updates:
Labels: Type-Defect

Comment #4 on issue 894 by sro...@gmail.com: Crash on some Deutsche Post
Datamatrix
http://code.google.com/p/zxing/issues/detail?id=894

(No comment was entered for this change.)

[zx...@googlecode.com]

Comment #5 on issue 894 by dacrus...@gmail.com: Crash on some Deutsche Post
Datamatrix
http://code.google.com/p/zxing/issues/detail?id=894

I'm running into (maybe the same) troubles, decoding a Datamatrix
from "Deutsche Post".
The Datamatrix is 26x26 Modules and Contains 0x00-Values.
I'm Using IOS based the OBJ-C Sample-Project.

The output I'm expecting for the provided sample.tif is:
44 45 41 12 18 01 2a 1c 83 1b 00 5a 32 d3 25 1e 00 00 67 33 00 61 01 00 03
30 30 30 30 30 30 30 30 34 31 30 35 39 35 36 34 33

Will any fix be applied to c-Version and Iphone as well?
It does not work with the lates svn-Trunk for me.
I Always get "Error locator degree does not match number of roots"
Thanks

Attachments:
sample.tif 373 KB

[zx...@googlecode.com]

Comment #6 on issue 894 by sro...@gmail.com: Crash on some Deutsche Post
Datamatrix
http://code.google.com/p/zxing/issues/detail?id=894

There are never particular plans to port changes to other languages. You'd
have to port it again yourself and contribute the patch if you like.

[zx...@googlecode.com]

Comment #7 on issue 894 by dennis.k...@googlemail.com: Crash on some

Deutsche Post Datamatrix
http://code.google.com/p/zxing/issues/detail?id=894

If you need an other example:
<http://img13.imageshack.us/img13/7479/imag0016xn.jpg>.
Package name: com.google.zxing.client.android
Package version: 70
Package version name: 3.6
Device: bravo (HTC Desire)
Release: 2.3.3

Result: Application Crash

[zx...@googlecode.com]

Comment #8 on issue 894 by maxr...@gmail.com: Crash on some Deutsche Post
Datamatrix
http://code.google.com/p/zxing/issues/detail?id=894

Doesn't crash for me, but doesn't get recognized.

[zx...@googlecode.com]

Comment #9 on issue 894 by l.dob...@gmail.com: Crash on some Deutsche Post
Datamatrix
http://code.google.com/p/zxing/issues/detail?id=894

Crashes with the Market version.
No crash from Subversion, but the code is not recognized no matter what!
Probably something that was «fixed» in Subversion.
Attaching the stack trace, may help in detecting why the code was killing
the application, and why it's not detecting it now. I suspect it gets
detected, but decoding fails.

Attachments:
DM-NASE.txt 1.5 KB

[zx...@googlecode.com]

Comment #10 on issue 894 by dennis.k...@googlemail.com: Crash on some

Deutsche Post Datamatrix
http://code.google.com/p/zxing/issues/detail?id=894

IMHO the lines
<code>
// We're seeing NegativeArraySizeException errors from users.
if (count < 0) {
throw FormatException.getFormatInstance();
}
</code>
will fix the Error. Without the "if-clause" int count can be negative in
the following for-loop. The error might be fixed in Subversion, but it
remains at least in market version:
---%<---
http://in-market.googlecode.com/svn/trunk/InMarket_android/src/com/google/zxing/datamatrix/decoder/DecodedBitStreamParser.java
---%<---

<http://www.google.com/codesearch#search/&q=DecodedBitStreamParser.java%20count%20250%20d1&type=cs>

[zx...@googlecode.com]

Comment #11 on issue 894 by dennis.k...@googlemail.com: Crash on some

Deutsche Post Datamatrix
http://code.google.com/p/zxing/issues/detail?id=894

Compare lines 454ff in
<http://zxing.googlecode.com/svn/trunk/core/src/com/google/zxing/datamatrix/decoder/DecodedBitStreamParser.java>
to line 434ff in
<http://in-market.googlecode.com/svn/trunk/InMarket_android/src/com/google/zxing/datamatrix/decoder/DecodedBitStreamParser.java>

[zx...@googlecode.com]

Updates:
Status: Fixed
Owner: sro...@gmail.com
Labels: Priority-Medium

Comment #12 on issue 894 by sro...@gmail.com: Crash on some Deutsche Post
Datamatrix
http://code.google.com/p/zxing/issues/detail?id=894

Finally looked again and spotted the easy fix. It long since didn't crash
but now it will decode too.