summaryrefslogtreecommitdiffstats
path: root/docs/manual/troubleshooting.html
blob: d3ed14e314226779beca2013b756b58e834f6ee4 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
<!doctype html PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=iso-8859-1">
<meta http-equiv="content-style-type" content="text/css">
<link rel="stylesheet" type="text/css" href="style.css">
<title>ProGuard Troubleshooting</title>
</head>
<body>

<h2>Troubleshooting</h2>

While preparing a configuration for processing your code, you may bump into a
few problems. The following sections discuss some common issues and solutions:

<h3><a href="#processing">Problems while processing</a></h3>
<ul>
<li><a href="#dynamicalclass">Note: can't find dynamically referenced class</a></li>
<li><a href="#dynamicalclasscast">Note: ... calls '(...)Class.forName(variable).newInstance()'</a></li>
<li><a href="#dynamicalclassmember">Note: ... accesses a field/method '...' dynamically</a></li>
<li><a href="#descriptorclass">Note: the configuration keeps the entry point '...', but not the descriptor class '...'</a></li>
<li><a href="#duplicateclass">Note: duplicate definition of program/library class</a></li>
<li><a href="#duplicatezipentry">Warning: can't write resource ... Duplicate zip entry</a></li>
<li><a href="#unresolvedclass">Warning: can't find superclass or interface</a></li>
<li><a href="#unresolvedclass">Warning: can't find referenced class</a></li>
<li><a href="#unresolvedclassmember">Warning: can't find referenced field/method</a></li>
<li><a href="#unresolvedenclosingmethod">Warning: can't find enclosing class/method</a></li>
<li><a href="#dependency">Warning: library class ... depends on program class ...</a></li>
<li><a href="#unexpectedclass">Warning: class file ... unexpectedly contains class ...</a></li>
<li><a href="#mappingconflict1">Warning: ... is not being kept as ..., but remapped to ...</a></li>
<li><a href="#mappingconflict2">Warning: field/method ... can't be mapped to ...</a></li>
<li><a href="#keep">Error: You have to specify '-keep' options</a></li>
<li><a href="#filename">Error: Expecting class path separator ';' before 'Files\Java\...' (in Windows)</a></li>
<li><a href="#macosx">Error: Can't read [.../lib/rt.jar] (No such file or directory) (in MacOS X)</a></li>
<li><a href="#outofmemoryerror">OutOfMemoryError</a></li>
<li><a href="#stackoverflowerror">StackOverflowError</a></li>
<li><a href="#unexpectederror">Unexpected error</a></li>
<li><a href="#otherwise">Otherwise...</a></li>
</ul>

<h3><a href="#afterprocessing">Unexpected observations after processing</a></h3>
<ul>
<li><a href="#disappearingclasses">Disappearing classes</a></li>
<li><a href="#notkept">Classes or class members not being kept</a></li>
<li><a href="#notobfuscated">Variable names not being obfuscated</a></li>
</ul>

<h3><a href="#preverifying">Problems while preverifying for Java Micro Edition</a></h3>

<ul>
<li><a href="#invalidclassexception1">InvalidClassException, class loading error, or verification error</a></li>
</ul>

<h3><a href="#runtime">Problems at run-time</a></h3>
<ul>
<li><a href="#stacktraces">Stack traces without class names or line numbers</a></li>
<li><a href="#noclassdeffounderror">NoClassDefFoundError</a></li>
<li><a href="#classnotfoundexception">ClassNotFoundException</a></li>
<li><a href="#nosuchmethodexception">NoSuchMethodException</a></li>
<li><a href="#missingresourceexception">MissingResourceException or NullPointerException</a></li>
<li><a href="#invalidjarfile">Invalid or corrupt jarfile</a></li>
<li><a href="#invalidclassexception2">InvalidClassException, class loading error, or verification error (in Java Micro Edition)</a></li>
<li><a href="#nosuchfieldormethod">Error: No Such Field or Method, Error verifying method (in a Java Micro Edition emulator)</a></li>
<li><a href="#failingmidlets">Failing midlets (on a Java Micro Edition device)</a></li>
<li><a href="#disappearingloops">Disappearing loops</a></li>
<li><a href="#securityexception">SecurityException: SHA1 digest error</a></li>
<li><a href="#classcastexception">ClassCastException: class not an enum</a></li><li><a href="#classcastexception">IllegalArgumentException: class not an enum type</a></li>
<li><a href="#arraystoreexception">ArrayStoreException: sun.reflect.annotation.EnumConstantNotPresentExceptionProxy</a></li>
<li><a href="#compilererror">CompilerError: duplicate addition</a></li>
<li><a href="#classformaterror">ClassFormatError: repetitive field name/signature</a></li>
<li><a href="#nosuchmethoderror">NoSuchMethodError or AbstractMethodError</a></li>
<li><a href="#verifyerror">VerifyError</a></li>
</ul>


<a name="processing">&nbsp;</a>
<h2>Problems while processing</h2>

ProGuard may print out some notes and non-fatal warnings:

<dl>
<dt><a name="dynamicalclass"><b>Note: can't find dynamically referenced class</b></a></dt>

<dd>ProGuard can't find a class or interface that your code is accessing by
    means of introspection. You should check if you want to add the jar that
    contains this class.</dd>

<dt><a name="dynamicalclasscast"><b>Note: ... calls '(...)Class.forName(variable).newInstance()'</b></a></dt>

<dd>ProGuard lists all class casts of dynamically created class instances,
    like "<code>(MyClass)Class.forName(variable).newInstance()</code>".
    Depending on your application, you may need to keep the mentioned classes
    with an option like "<code>-keep class MyClass</code>", or their
    implementations with an option like "<code>-keep class * implements
    MyClass</code>". You can switch off these notes by specifying the
    <a href="usage.html#dontnote"><code>-dontnote</code></a> option.</dd>

<dt><a name="dynamicalclassmember"><b>Note: ... accesses a field/method '...' dynamically</b></a></dt>

<dd>ProGuard lists a number of constructs like
    "<code>.getField("myField")</code>". Depending on your application, you
    may need to figure out where the mentioned class members are defined and
    keep them with an option like "<code>-keep class MyClass { MyFieldType
    myField; }</code>". Otherwise, ProGuard might remove or obfuscate the
    class members, since it can't know which ones they are exactly. It does
    list possible candidates, for your information. You can switch off these
    notes by specifying the <a
    href="usage.html#dontnote"><code>-dontnote</code></a> option.</dd>

<dt><a name="descriptorclass"><b>Note: the configuration keeps the entry point '...', but not the descriptor class '...'</b></a></dt>

<dd>Your configuration contains a <code>-keep</code> option to preserve the
    given method (or field), but no <code>-keep</code> option for the given
    class that is an argument type or return type in the method's descriptor.
    You may then want to keep the class too. Otherwise, ProGuard will
    obfuscate its name, thus changing the method's signature. The method might
    then become unfindable as an entry point, e.g. if it is part of a public
    API. You can switch off these notes by specifying the <a
    href="usage.html#dontnote"><code>-dontnote</code></a> option.</dd>

<dt><a name="duplicateclass"><b>Note: duplicate definition of program/library class</b></a></dt>

<dd>Your program jars or library jars contain multiple definitions of the
    listed classes. ProGuard continues processing as usual, only considering
    the first definitions. The warning may be an indication of some problem
    though, so it's advisable to remove the duplicates. A convenient way to do
    so is by specifying filters on the input jars or library jars. You can
    switch off these notes by specifying the <a
    href="usage.html#dontnote"><code>-dontnote</code></a> option.</dd>

<dt><a name="duplicatezipentry"><b>Warning: can't write resource ... Duplicate zip entry</b></a></dt>

<dd>Your input jars contain multiple resource files with the same name.
    ProGuard continues copying the resource files as usual, skipping any files
    with previously used names. Once more, the warning may be an indication of
    some problem though, so it's advisable to remove the duplicates. A
    convenient way to do so is by specifying filters on the input jars. There
    is no option to switch off these warnings.</dd>

</dl>
<p>

ProGuard may terminate when it encounters parsing errors or I/O errors, or
some more serious warnings:

<dl>
<dt><a name="unresolvedclass"><b>Warning: can't find superclass or interface</b><br/><b>Warning: can't find referenced class</b></a></dt>

<dd>If there are unresolved references to classes or interfaces, you most
    likely forgot to specify an essential library. For proper processing, all
    libraries that are referenced by your code must be specified, including
    the Java run-time library. For specifying libraries, use
    the <a href="usage.html#libraryjars"><code>-libraryjars</code></a> option.
    <p>
    If the class that is reported as missing is a non-public library class,
    you should specify the <a
    href="usage.html#dontskipnonpubliclibraryclasses"><code>-dontskipnonpubliclibraryclasses</code></a>
    option. Common examples are the classes
    <code>javax.swing.TransferHandler$HasGetTransferHandler</code> and
    <code>java.util.zip.ZipConstants</code>, which are used as interfaces in
    some public classes, even though they are only package visible. This
    option is not set by default for reasons of efficiency. Setting it increases
    the processing time a bit, but it won't hurt the output in any way.
    <p>
    If you're missing a library and you're absolutely sure it isn't used
    anyway, you can try your luck with the <a
    href="usage.html#ignorewarnings"><code>-ignorewarnings</code></a> option,
    or even the <a href="usage.html#dontwarn"><code>-dontwarn</code></a>
    option. Only use these options if you really know what you're doing
    though.</dd>

<dt><a name="unresolvedclassmember"><b>Warning: can't find referenced field/method</b></a></dt>

<dd>If there are unresolved references to class members in input classes, your
    class files are most likely inconsistent. Possibly, some class file didn't
    get recompiled properly, or some class file was left behind after its
    source file was removed. Try removing all class files, recompiling them,
    zipping them up, and running ProGuard again.
    <p>
    If the class member that is reported as missing is actually implemented in
    a non-public library class, you should specify the
    <a href="usage.html#dontskipnonpubliclibraryclasses"><code>-dontskipnonpubliclibraryclasses</code></a> option. A common example is
    the method <code>setLength(int)</code> in the public class
    <code>java.lang.StringBuilder</code>. This method is actually defined in
    the package visible superclass
    <code>java.lang.AbstractStringBuilder</code>, which ProGuard ignores by
    default.
    <p>
    If your program classes reside in the same packages as library classes,
    and refer to their package visible class members, then you should specify
    the <a
    href="usage.html#dontskipnonpubliclibraryclassmembers"><code>-dontskipnonpubliclibraryclassmembers</code></a>
    option.</dd>

<dt><a name="unresolvedenclosingmethod"><b>Warning: can't find enclosing class/method</b></a></dt>

<dd>If there are unresolved references to classes that are defined inside
    methods in your input, once more, your class files are most likely
    inconsistent. Possibly, some class file didn't get recompiled properly, or
    some class file was left behind after its source file was removed. Try
    removing all class files, recompiling them, zipping them up, and running
    ProGuard again.</dd>

<dt><a name="dependency"><b>Warning: library class ... depends on program class ...</b></a></dt>

<dd>If any of your library classes depend on your program classes, by
    extending, implementing or just referencing them, your processed code will
    generally be unusable. Program classes can depend on library classes, but
    not the other way around. Program classes are processed, while library
    classes always remain unchanged. It is therefore impossible to adapt
    references from library classes to program classes, for instance if the
    program classes are renamed. You should define a clean separation between
    program code and library code, and try again.</dd>

<dt><a name="unexpectedclass"><b>Warning: class file ... unexpectedly contains class ...</b></a></dt>

<dd>The given class file contains a definition for the given class, but the
    directory name of the file doesn't correspond to the package name of the
    class. ProGuard will accept the class definition, but the current
    implementation will not write out the processed version. Please make sure
    your input classes are packaged correctly. Notably, class files that are
    in the <code>WEB-INF/classes</code> directory in a war should be packaged
    in a jar and put in the <code>WEB-INF/lib</code> directory. If you don't
    mind these classes not being written to the output, you can specify the <a
    href="usage.html#ignorewarnings"><code>-ignorewarnings</code></a> option,
    or even the <a href="usage.html#dontwarn"><code>-dontwarn</code></a>
    option.</dd>

<dt><a name="mappingconflict1"><b>Warning: ... is not being kept as ..., but remapped to ...</b></a></dt>

<dd>There is a conflict between a <code>-keep</code> option in the
    configuration, and the mapping file, in the obfuscation step. The given
    class name or class member name can't be kept by its original name, as
    specified in the configuration, but it has to be mapped to the other given
    name, as specified in the mapping file. You should adapt your
    configuration or your mapping file to remove the conflict. Alternatively,
    if you're sure the renaming won't hurt, you can specify the <a
    href="usage.html#ignorewarnings"><code>-ignorewarnings</code></a> option,
    or even the <a href="usage.html#dontwarn"><code>-dontwarn</code></a>
    option.</dd>

<dt><a name="mappingconflict2"><b>Warning: field/method ... can't be mapped to ...</b></a></dt>

<dd>There is a conflict between some new program code and the mapping file, in
    the obfuscation step. The given class member can't be mapped to the given
    name, because it would conflict with another class member that is already
    being mapped to the same name. This can happen if you are performing
    incremental obfuscation, applying an obfuscation mapping file from an
    initial obfuscation step. For instance, some new class may have been added
    that extends two existing classes, introducing a conflict in the name
    space of its class members. If you're sure the class member receiving
    another name than the one specified won't hurt, you can specify the <a
    href="usage.html#ignorewarnings"><code>-ignorewarnings</code></a> option,
    or even the <a href="usage.html#dontwarn"><code>-dontwarn</code></a>
    option. Note that you should always use the <a
    href="usage.html#useuniqueclassmembernames"><code>-useuniqueclassmembernames</code></a>
    option in the initial obfuscation step, in order to reduce the risk of
    conflicts.</dd>

<dt><a name="keep"><b>Error: You have to specify '-keep' options</b></a></dt>

<dd>You either forgot to specify <a
    href="usage.html#keep"><code>-keep</code></a> options, or you mistyped the
    class names. ProGuard has to know exactly what you want to keep: an
    application, an applet, a servlet, a midlet,..., or any combination of
    these. Without the proper seed specifications, ProGuard would shrink,
    optimize, or obfuscate all class files away.</dd>


<dt><a name="filename"><b>Error: Expecting class path separator ';' before 'Files\Java\</b>...<b>'</b> (in Windows)</a></dt>

<dd>If the path of your run-time jar contains spaces, like in "Program Files",
    you have to enclose it with single or double quotes, as explained in the
    section on <a href="usage.html#filename">file names</a>. This is actually
    true for all file names containing special characters, on all
    platforms.</dd>

<dt><a name="macosx"><b>Error: Can't read [</b>...<b>/lib/rt.jar] (No such file or directory)</b> (in MacOS X)</a></dt>

<dd>In MacOS X, the run-time classes may be in a different place than on most
    other platforms. You'll then have to adapt your configuration, replacing
    the path <code>&lt;java.home&gt;/lib/rt.jar</code> by
    <code>&lt;java.home&gt;/../Classes/classes.jar</code>.</dd>

</dl>
<p>

Should ProGuard crash while processing your application:

<dl>
<dt><a name="outofmemoryerror"><b>OutOfMemoryError</b></a></dt>

<dd>You can try increasing the heap size of the Java virtual machine (with the
    usual <code>-Xms</code> and <code>-Xmx</code> options). You can also
    reduce the amount of memory that ProGuard needs by removing unnecessary
    library jars from your configuration, or by filtering out unused library
    packages and classes. Remember that only classes or interfaces that are
    extended or implemented by classes in your input jars are required.</dd>

<dt><a name="stackoverflowerror"><b>StackOverflowError</b></a></dt>

<dd>This error might occur when processing a large code base on Windows
    (surprisingly, not so easily on Linux). In theory, increasing the stack
    size of the Java virtual machine (with the usual <code>-Xss</code> option)
    should help too. In practice however, the <code>-Xss</code> setting
    doesn't have any effect on the main thread, due to <a
    href="http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=4362291">Sun Bug
    #4362291</a>. As a result, this solution will only work when running
    ProGuard in a different thread, e.g. from its GUI.</dd>

<dt><a name="unexpectederror"><b>Unexpected error</b></a></dt>

<dd>ProGuard has encountered an unexpected condition, typically in the
    optimization step. It may or may not recover. You should be able to avoid
    it using the <a
    href="usage.html#dontoptimize"><code>-dontoptimize</code></a> option. In
    any case, please report the problem, preferably with the simplest example
    that causes ProGuard to crash.</dd>

<dt><a name="otherwise"><b>Otherwise...</b></a></dt>

<dd>Maybe your class files are corrupt. See if recompiling them and trying
    again helps. If not, please report the problem, preferably with the
    simplest example that causes ProGuard to crash.</dd>

</dl>
<p>

<a name="afterprocessing">&nbsp;</a>
<h2>Unexpected observations after processing</h2>

If ProGuard seems to run fine, but your processed code doesn't look right,
there might be a couple of reasons:

<dl>
<dt><a name="disappearingclasses"><b>Disappearing classes</b></a></dt>

<dd>If you are working on Windows and it looks like some classes have
    disappeared from your output, you should make sure you're not writing your
    output class files to a directory (or unpacking the output jar). On
    platforms with case-insensitive file systems, such as Windows, unpacking
    tools often let class files with similar lower-case and upper-case names
    overwrite each other. If you really can't switch to a different operating
    system, you could consider using ProGuard's <a
    href="usage.html#dontusemixedcaseclassnames"><code>-dontusemixedcaseclassnames</code></a>
    option.
    <p>
    Also, you should make sure your class files are in directories that
    correspond to their package names. ProGuard will read misplaced class
    files, but it will currently not write their processed versions. Notably,
    class files that are in the <code>WEB-INF/classes</code> directory in a
    war should be packaged in a jar and put in the <code>WEB-INF/lib</code>
    directory.</dd>

<dt><a name="notkept"><b>Classes or class members not being kept</b></a></dt>

<dd>If ProGuard is not keeping the right classes or class members, make sure
    you are using fully qualified class names. If the package name of some
    class is missing, ProGuard won't match the elements that you might be
    expecting. It may help to double-check for typos too. You can use the <a
    href="usage.html#printseeds"><code>-printseeds</code></a> option to see
    which elements are being kept exactly.
    <p>
    If you are using marker interfaces to keep other classes, the marker
    interfaces themselves are probably being removed in the shrinking step.
    You should therefore always explicitly keep any marker interfaces.</dd>

<dt><a name="notobfuscated"><b>Variable names not being obfuscated</b></a></dt>

<dd>If the names of the local variables and parameters in your obfuscated code
    don't look obfuscated, because they suspiciously resemble the names of
    their types, it's probably because the decompiler that you are using is
    coming up with those names. ProGuard's obfuscation step does remove the
    original names entirely, unless you explicitly keep the
    <code>LocalVariableTable</code> or <code>LocalVariableTypeTable</code>
    attributes.</dd>

</dl>

<a name="preverifying">&nbsp;</a>
<h2>Problems while preverifying for Java Micro Edition</h2>

If ProGuard seems to run fine, but the external preverifier subsequently
produces errors, it's usually for a single reason:

<dl>
<dt><a name="invalidclassexception1"><b>InvalidClassException</b>, <b>class loading error</b>, or <b>verification error</b></a></dt>

<dd>If you get any such message from the preverifier, you are probably working
    on a platform with a case-insensitive file system, such as Windows. The
    <code>preverify</code> tool always unpacks the jars, so class files with
    similar lower-case and upper-case names overwrite each other. You can use
    ProGuard's <a
    href="usage.html#dontusemixedcaseclassnames"><code>-dontusemixedcaseclassnames</code></a>
    option to work around this problem.
    <p>
    If the above doesn't help, there is probably a bug in the optimization
    step of ProGuard. Make sure you are using the latest version. You should
    be able to work around the problem by using the <a
    href="usage.html#dontoptimize"><code>-dontoptimize</code></a> option. You
    can check the bug database to see if it is a known problem (often with a
    fix). Otherwise, please report it, preferably with the simplest example on
    which you can find ProGuard to fail.</dd>

</dl>

Note that it is no longer necessary to use an external preverifier. With the
<a href="usage.html#microedition"><code>-microedition</code></a> option,
ProGuard will preverify the class files for Java Micro Edition.
<p>

<a name="runtime">&nbsp;</a>
<h2>Problems at run-time</h2>

If ProGuard runs fine, but your processed application doesn't work, there
might be several reasons:

<dl>
<dt><a name="stacktraces"><b>Stack traces without class names or line numbers</b></a></dt>

<dd>If your stack traces don't contain any class names or lines numbers,
    even though you are keeping the proper attributes, make sure this debugging
    information is present in your compiled code to start with. Notably the Ant
    javac task has debugging information switched off by default.</dd>

<dt><a name="noclassdeffounderror"><b>NoClassDefFoundError</b></a></dt>

<dd>Your class path is probably incorrect. It should at least contain all
    library jars and, of course, your processed program jar.</dd>

<dt><a name="classnotfoundexception"><b>ClassNotFoundException</b></a></dt>

<dd>Your code is probably calling <code>Class.forName</code>, trying to create
    the missing class dynamically. ProGuard can only detect constant name
    arguments, like <code>Class.forName("mypackage.MyClass")</code>. For
    variable name arguments like <code>Class.forName(someClass)</code>, you
    have to keep all possible classes using the appropriate <a
    href="usage.html#keep"><code>-keep</code></a> option, e.g. "<code>-keep
    class mypackage.MyClass</code>" or "<code>-keep class * implements
    mypackage.MyInterface</code>".</dd>

<dt><a name="nosuchmethodexception"><b>NoSuchMethodException</b></a></dt>

<dd>Your code is probably calling something like
    <code>myClass.getMethod</code>, trying to find some method dynamically.
    Since ProGuard isn't detecting this (yet), you have to keep the missing
    method in using the appropriate <a
    href="usage.html#keep"><code>-keep</code></a> option, e.g. "<code>-keep
    class mypackage.MyClass { void myMethod(); }</code>".</dd>

<dt><a name="missingresourceexception"><b>MissingResourceException</b> or <b>NullPointerException</b></a></dt>

<dd>Your processed code may be unable to find some resource files. ProGuard
    simply copies resource files over from the input jars to the output jars.
    Their names and contents remain unchanged, unless you specify the options
    <a
    href="usage.html#adaptresourcefilenames"><code>-adaptresourcefilenames</code></a>
    and/or <a
    href="usage.html#adaptresourcefilecontents"><code>-adaptresourcefilecontents</code></a>.
    <p>
    Furthermore, directory entries in jar files aren't copied, unless you
    specify the option <a
    href="usage.html#keepdirectories"><code>-keepdirectories</code></a>.</dd>

<dt><a name="invalidjarfile"><b>Invalid or corrupt jarfile</b></a></dt>

<dd>You are probably starting your application with the java option
    <code>-jar</code> instead of the option <code>-classpath</code>. The java
    virtual machine returns with this error message if your jar doesn't
    contain a manifest file (<code>META-INF/MANIFEST.MF</code>), if the
    manifest file doesn't specify a main class (<code>Main-Class:</code> ...),
    or if the jar doesn't contain this main class. You should then make sure
    that the input jar contains a valid manifest file to start with, that this
    manifest file is the one that is copied (the first manifest file that is
    encountered), and that the main class is kept in your configuration,</dd>

<dt><a name="invalidclassexception2"><b>InvalidClassException</b>, <b>class loading error</b>, or <b>verification error</b> (in Java Micro Edition)</a></dt>

<dd>If you get such an error in Java Micro Edition, you may have forgotten to
    specify the <a
    href="usage.html#microedition"><code>-microedition</code></a> option, so
    the processed class files are preverified properly.</dd>

<dt><a name="nosuchfieldormethod"><b>Error: No Such Field or Method</b>, <b>Error verifying method</b> (in a Java Micro Edition emulator)</a></dt>

<dd>If you get such a message in a Motorola or Sony Ericsson phone emulator,
    it's because these emulators don't like packageless classes and/or
    overloaded fields and methods. You can work around it by not using the
    options <code><a href="usage.html#repackageclasses">-repackageclasses</a>
    ''</code> and <a
    href="usage.html#overloadaggressively"><code>-overloadaggressively</code></a>.
    If you're using the JME WTK plugin, you can adapt the configuration
    <code>proguard/wtk/default.pro</code> that's inside the
    <code>proguard.jar</code>.</dd>

<dt><a name="failingmidlets"><b>Failing midlets</b> (on a Java Micro Edition device)</a></dt>

<dd>If your midlet runs in an emulator and on some devices, but not on some
    other devices, this is probably due to a bug in the latter devices. For
    some older Motorola and Nokia phones, you might try specifying the <a
    href="usage.html#useuniqueclassmembernames"><code>-useuniqueclassmembernames</code></a>
    option. It avoids overloading class member names, which triggers a bug in
    their java virtual machine.
    <p>
    You might also try using the <a
    href="usage.html#dontusemixedcaseclassnames"><code>-dontusemixedcaseclassnames</code></a>
    option. Even if the midlet has been properly processed and then
    preverified on a case-sensitive file system, the device itself might not
    like the mixed-case class names. Notably, the Nokia N-Gage emulator works
    fine, but the actual device seems to exhibit this problem.</dd>

<dt><a name="disappearingloops"><b>Disappearing loops</b></a></dt>

<dd>If your code contains empty busy-waiting loops, ProGuard's optimization
    step may remove them. More specifically, this happens if a loop
    continuously checks the value of a non-volatile field that is changed in a
    different thread. The specifications of the Java Virtual Machine require
    that you always mark fields that are accessed across different threads
    without further synchronization as <code>volatile</code>. If this is not
    possible for some reason, you'll have to switch off optimization using the
    <a href="usage.html#dontoptimize"><code>-dontoptimize</code></a>
    option.</dd>

<dt><a name="securityexception"><b>SecurityException: SHA1 digest error</b></a></dt>

<dd>You may have forgotten to sign your program jar <i>after</i> having
    processed it with ProGuard.</dd>

<dt><a name="classcastexception"><b>ClassCastException: class not an enum</b>, or <br><b>IllegalArgumentException: class not an enum type</b></a></dt>

<dd>You should make sure you're preserving the special methods of enumeration
    types, which the run-time environment calls by introspection. The required
    options are shown in the <a
    href="examples.html#enumerations">examples</a>.</dd>

<dt><a name="arraystoreexception"><b>ArrayStoreException: sun.reflect.annotation.EnumConstantNotPresentExceptionProxy</b></a></dt>

<dd>You are probably processing annotations involving enumerations. Again, you
    should make sure you're preserving the special methods of the enumeration
    type, as shown in the examples.</dd>

<dt><a name="compilererror"><b>CompilerError: duplicate addition</b></a></dt>

<dd>You are probably compiling or running some code that has been obfuscated
    with the <a
    href="usage.html#overloadaggressively"><code>-overloadaggressively</code></a>
    option. This option triggers a bug in
    <code>sun.tools.java.MethodSet.add</code> in Sun's JDK 1.2.2, which is
    used for (dynamic) compilation. You should then avoid this option.</dd>

<dt><a name="classformaterror"><b>ClassFormatError: repetitive field name/signature</b></a></dt>

<dd>You are probably processing some code that has been obfuscated before with
    the <a
    href="usage.html#overloadaggressively"><code>-overloadaggressively</code></a>
    option. You should then use the same option again in the second processing
    round.</dd>

<dt><a name="nosuchmethoderror"><b>NoSuchMethodError</b> or <b>AbstractMethodError</b></a></dt>

<dd>You should make sure you're not writing your output class files to a
    directory on a platform with a case-insensitive file system, such as
    Windows. Please refer to the section about <a
    href="#disappearingclasses">disappearing classes</a> for details.
    <p>
    Furthermore, you should check whether you have specified your program jars
    and library jars properly. Program classes can refer to library classes,
    but not the other way around.
    <p>
    If all of this seems ok, perhaps there's a bug in ProGuard (gasp!). If so,
    please report it, preferably with the simplest example on which you can
    find ProGuard to fail.</dd>

<dt><a name="verifyerror"><b>VerifyError</b></a></dt>

<dd>Verification errors when executing a program are almost certainly the
    result of a bug in the optimization step of ProGuard. Make sure you are
    using the latest version. You should be able to work around the problem by
    using the <a href="usage.html#dontoptimize"><code>-dontoptimize</code></a>
    option. You can check the bug database to see if it is a known problem
    (often with a fix). Otherwise, please report it, preferably with the
    simplest example on which ProGuard fails.</dd>

</dl>

<hr>
<address>
Copyright &copy; 2002-2009
<a href="http://www.graphics.cornell.edu/~eric/">Eric Lafortune</a>.
</address>
</body>
</html>