[codenameone-discussions] Re: A new IOS hard crash
I didn't acknowledge anything, so far you are the only guy to report this. -- You received this message because you are subscribed to the Google Groups "CodenameOne Discussions" group. To unsubscribe from this group and stop receiving emails from it, send an email to codenameone-discussions+unsubscr...@googlegroups.com. Visit this group at https://groups.google.com/group/codenameone-discussions. To view this discussion on the web visit https://groups.google.com/d/msgid/codenameone-discussions/51b47875-bcfa-4b78-8bea-9b389a46a7fb%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[codenameone-discussions] Re: A new IOS hard crash
A policy question. As of now, it's been 16 hours since you tacitly ackknowledged that IOS builds are seriously broken by a recent update. I know "stuff happens" but your build servers are still building, still producing the bad builds. Shouldn't there be something in place to either turn off the build servers, or do an emergency rollback, or something? It's just not appropriate to let your customers continue building timebombs and either wasting their own time or (worse) distributing bad builds to their customers. -- You received this message because you are subscribed to the Google Groups "CodenameOne Discussions" group. To unsubscribe from this group and stop receiving emails from it, send an email to codenameone-discussions+unsubscr...@googlegroups.com. Visit this group at https://groups.google.com/group/codenameone-discussions. To view this discussion on the web visit https://groups.google.com/d/msgid/codenameone-discussions/5d9276f9-83c1-4ef7-9a7b-2caf3cefe25b%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[codenameone-discussions] Re: A new IOS hard crash
Meantime, anyone who is using 2d arrays is shipping bombs. -- You received this message because you are subscribed to the Google Groups "CodenameOne Discussions" group. To unsubscribe from this group and stop receiving emails from it, send an email to codenameone-discussions+unsubscr...@googlegroups.com. Visit this group at https://groups.google.com/group/codenameone-discussions. To view this discussion on the web visit https://groups.google.com/d/msgid/codenameone-discussions/1b531c6a-668a-4b0e-abd4-028eb0008b93%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[codenameone-discussions] Re: A new IOS hard crash
Yup, that looks like the right place. Bummer that it was a fix to a purely theoretical problem that I reported. Q/A is a bitch. -- You received this message because you are subscribed to the Google Groups "CodenameOne Discussions" group. To unsubscribe from this group and stop receiving emails from it, send an email to codenameone-discussions+unsubscr...@googlegroups.com. Visit this group at https://groups.google.com/group/codenameone-discussions. To view this discussion on the web visit https://groups.google.com/d/msgid/codenameone-discussions/220fab1b-132f-4115-ad30-140540fb9f1c%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[codenameone-discussions] Re: A new IOS hard crash
Probably this: https://github.com/codenameone/CodenameOne/commit/05b7d1cff7f6811553f7d06c1eaba2cd4938a7b2 -- You received this message because you are subscribed to the Google Groups "CodenameOne Discussions" group. To unsubscribe from this group and stop receiving emails from it, send an email to codenameone-discussions+unsubscr...@googlegroups.com. Visit this group at https://groups.google.com/group/codenameone-discussions. To view this discussion on the web visit https://groups.google.com/d/msgid/codenameone-discussions/a77bcfd0-3a46-4888-bcac-6f4b4525d5e0%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[codenameone-discussions] Re: A new IOS hard crash
I got lucky, my first attempt to produce a small test case worked. https://github.com/codenameone/CodenameOne/issues/1960 -- You received this message because you are subscribed to the Google Groups "CodenameOne Discussions" group. To unsubscribe from this group and stop receiving emails from it, send an email to codenameone-discussions+unsubscr...@googlegroups.com. Visit this group at https://groups.google.com/group/codenameone-discussions. To view this discussion on the web visit https://groups.google.com/d/msgid/codenameone-discussions/d2df612b-0d9a-478b-a732-259bc0a2f5ce%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[codenameone-discussions] Re: A new IOS hard crash
To my inexpert eye, the new code is the same as the old code - and the problem is probably that the value of "self" or of instance variable "board" is corrupt. In any case, this kind of crash is definitely not in the manual; the VM is demonstrably damaged. -- You received this message because you are subscribed to the Google Groups "CodenameOne Discussions" group. To unsubscribe from this group and stop receiving emails from it, send an email to codenameone-discussions+unsubscr...@googlegroups.com. Visit this group at https://groups.google.com/group/codenameone-discussions. To view this discussion on the web visit https://groups.google.com/d/msgid/codenameone-discussions/fdd74d83-d4e2-49a9-83d9-960279254ab6%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[codenameone-discussions] Re: A new IOS hard crash
For reference, here is the previously generated .m code, which worked.
> JAVA_VOID
> online_game_gBoard_SetBoardCell___char_int_online_game_cell(CODENAME_ONE_THREAD_STATE,
>
> JAVA_OBJECT __cn1ThisObject, JAVA_CHAR __cn1Arg1, JAVA_INT __cn1Arg2,
> JAVA_OBJECT __cn1Arg3) {
> JAVA_INT ilocals_5_; /* v5 */
> JAVA_INT ilocals_4_; /* v4 */
> JAVA_INT ilocals_7_; /* v7 */
> JAVA_INT ilocals_6_; /* v6 */
> DEFINE_INSTANCE_METHOD_STACK(6, 8, 0, 10363, 10379);
> locals[0].data.o = __cn1ThisObject; locals[0].type =
> CN1_TYPE_OBJECT; JAVA_INT ilocals_1_ = __cn1Arg1;
> JAVA_INT ilocals_2_ = __cn1Arg2;
> locals[3].data.o = __cn1Arg3;
> locals[3].type = CN1_TYPE_OBJECT;
> BC_ALOAD(0);
> /* CustomInvoke */{ JAVA_INT tmpResult =
> virtual_online_game_gBoard_BCtoYindex___char_int_R_int(threadStateData,
> SP[-1].data.o, ilocals_1_, ilocals_2_);
> SP[-1].data.i = tmpResult; SP[-1].type = CN1_TYPE_INT; }
> BC_ISTORE(4);
> BC_ALOAD(0);
> /* CustomInvoke */{ JAVA_INT tmpResult =
> virtual_online_game_gBoard_BCtoXindex___char_int_R_int(threadStateData,
> SP[-1].data.o, ilocals_1_, ilocals_2_);
> SP[-1].data.i = tmpResult; SP[-1].type = CN1_TYPE_INT; }
> BC_ISTORE(5);
> BC_ALOAD(0);
> /* CustomInvoke */{ JAVA_INT tmpResult =
> virtual_online_game_gBoard_XindexToBC___int_int_R_char(threadStateData,
> SP[-1].data.o, ilocals_5_, ilocals_4_);
> SP[-1].data.i = tmpResult; SP[-1].type = CN1_TYPE_INT; }
> BC_ISTORE(6);
> BC_ALOAD(0);
> /* CustomInvoke */{ JAVA_INT tmpResult =
> virtual_online_game_gBoard_YindexToBC___int_int_R_int(threadStateData,
> SP[-1].data.o, ilocals_5_, ilocals_4_);
> SP[-1].data.i = tmpResult; SP[-1].type = CN1_TYPE_INT; }
> BC_ISTORE(7);
> if (ilocals_6_!=ilocals_1_) /* IF_ICMPNE CustomJump */ goto
> label_L11395;
> if (ilocals_7_!=ilocals_2_) /* IF_ICMPNE CustomJump */ goto
> label_L11395;
> PUSH_INT(1); /* ICONST_1 */
> goto label_L315529366;
>
> label_L11395:
> PUSH_INT(0); /* ICONST_0 */
>
> label_L315529366:
> /* LDC: '%s,%s encodes correctly'*/
> PUSH_POINTER(STRING_FROM_CONSTANT_POOL_OFFSET(10380));
> PUSH_INT(2); /* ICONST_2 */
> SP--;
> PUSH_POINTER(__NEW_ARRAY_java_lang_Object(threadStateData,
> SP[0].data.i));
> BC_DUP(); /* DUP */
> PUSH_INT(0); /* ICONST_0 */
> /* CustomInvoke
> */PUSH_OBJ(java_lang_Character_valueOf___char_R_java_lang_Character(threadStateData,
>
> ilocals_1_));
> CHECK_ARRAY_ACCESS(3, SP[-2].data.i); { /* BC_AASTORE */
> JAVA_OBJECT aastoreTmp = SP[-3].data.o;
> ((JAVA_ARRAY_OBJECT*) (*(JAVA_ARRAY)aastoreTmp).data)[SP[-2].data.i] =
> SP[-1].data.o;
> SP -= 3; }
> BC_DUP(); /* DUP */
> PUSH_INT(1); /* ICONST_1 */
> /* CustomInvoke
> */PUSH_OBJ(java_lang_Integer_valueOf___int_R_java_lang_Integer(threadStateData,
>
> ilocals_2_));
> CHECK_ARRAY_ACCESS(3, SP[-2].data.i); { /* BC_AASTORE */
> JAVA_OBJECT aastoreTmp = SP[-3].data.o;
> ((JAVA_ARRAY_OBJECT*) (*(JAVA_ARRAY)aastoreTmp).data)[SP[-2].data.i] =
> SP[-1].data.o;
> SP -= 3; }
> { JAVA_INT tmpResult =
> lib_G_Assert___boolean_java_lang_String_java_lang_Object_1ARRAY_R_boolean(threadStateData,
>
> SP[-3].data.i, SP[-2].data.o, SP[-1].data.o);
> SP-=2;
> SP[-1].data.i = tmpResult; SP[-1].type = CN1_TYPE_INT; }
> SP--; /* POP */
> PUSH_POINTER(get_field_online_game_gBoard_board(__cn1ThisObject));
> (*SP).type = CN1_TYPE_INT; /* ILOAD */
> (*SP).data.i = ilocals_5_;
> SP++;
> CHECK_ARRAY_ACCESS(2, SP[-1].data.i); /* AALOAD */
> SP--; SP[-1].type = CN1_TYPE_INVALID;
> SP[-1].data.o = ((JAVA_ARRAY_OBJECT*)
> (*(JAVA_ARRAY)SP[-1].data.o).data)[(*SP).data.i];
> SP[-1].type = CN1_TYPE_OBJECT;
> (*SP).type = CN1_TYPE_INT; /* ILOAD */
> (*SP).data.i = ilocals_4_;
> SP++;
> BC_ALOAD(3);
> CHECK_ARRAY_ACCESS(3, SP[-2].data.i); { /* BC_AASTORE */
> JAVA_OBJECT aastoreTmp = SP[-3].data.o;
> ((JAVA_ARRAY_OBJECT*) (*(JAVA_ARRAY)aastoreTmp).data)[SP[-2].data.i] =
> SP[-1].data.o;
> SP -= 3; }
> releaseForReturn(threadStateData, cn1LocalsBeginInThread);
> return;
> }
>
>
--
You received this message because you are subscribed to the Google Groups
"CodenameOne Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to codenameone-discussions+unsubscr...@googlegroups.com.
Visit this group at https://groups.google.com/group/codenameone-discussions.
To view this discussion on the web visit
https://groups.google.com/d/msgid/codenameone-discussions/c988d518-a6c2-4f8d-9602-101cd56d6a75%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
[codenameone-discussions] Re: A new IOS hard crash
For reference, here is the .m code that used to be generated (and which
worked)
JAVA_OBJECT
online_game_gBoard_GetBoardCell___char_int_R_online_game_cell(CODENAME_ONE_THREAD_STATE,
JAVA_OBJECT __cn1ThisObject, JAVA_CHAR __cn1Arg1, JAVA_INT __cn1Arg2) {
JAVA_INT ilocals_3_; /* v3 */
JAVA_INT ilocals_4_; /* v4 */
DEFINE_INSTANCE_METHOD_STACK(3, 5, 0, 10740, 10758);
locals[0].data.o = __cn1ThisObject; locals[0].type =
CN1_TYPE_OBJECT; JAVA_INT ilocals_1_ = __cn1Arg1;
JAVA_INT ilocals_2_ = __cn1Arg2;
BC_ALOAD(0);
/* CustomInvoke */{ JAVA_INT tmpResult =
virtual_online_game_gBoard_validBoardPos___char_int_R_boolean(threadStateData,
SP[-1].data.o, ilocals_1_, ilocals_2_);
SP[-1].data.i = tmpResult; SP[-1].type = CN1_TYPE_INT; }
if(POP_INT() == 0) /* IFEQ */ goto label_L103709560;
BC_ALOAD(0);
/* CustomInvoke */{ JAVA_INT tmpResult =
virtual_online_game_gBoard_BCtoYindex___char_int_R_int(threadStateData,
SP[-1].data.o, ilocals_1_, ilocals_2_);
SP[-1].data.i = tmpResult; SP[-1].type = CN1_TYPE_INT; }
BC_ISTORE(3);
BC_ALOAD(0);
/* CustomInvoke */{ JAVA_INT tmpResult =
virtual_online_game_gBoard_BCtoXindex___char_int_R_int(threadStateData,
SP[-1].data.o, ilocals_1_, ilocals_2_);
SP[-1].data.i = tmpResult; SP[-1].type = CN1_TYPE_INT; }
BC_ISTORE(4);
PUSH_POINTER(get_field_online_game_gBoard_board(__cn1ThisObject));
(*SP).type = CN1_TYPE_INT; /* ILOAD */
(*SP).data.i = ilocals_4_;
SP++;
CHECK_ARRAY_ACCESS(2, SP[-1].data.i); /* AALOAD */
SP--; SP[-1].type = CN1_TYPE_INVALID;
SP[-1].data.o = ((JAVA_ARRAY_OBJECT*)
(*(JAVA_ARRAY)SP[-1].data.o).data)[(*SP).data.i];
SP[-1].type = CN1_TYPE_OBJECT;
(*SP).type = CN1_TYPE_INT; /* ILOAD */
(*SP).data.i = ilocals_3_;
SP++;
CHECK_ARRAY_ACCESS(2, SP[-1].data.i); /* AALOAD */
SP--; SP[-1].type = CN1_TYPE_INVALID;
SP[-1].data.o = ((JAVA_ARRAY_OBJECT*)
(*(JAVA_ARRAY)SP[-1].data.o).data)[(*SP).data.i];
SP[-1].type = CN1_TYPE_OBJECT;
releaseForReturn(threadStateData, cn1LocalsBeginInThread);
return POP_OBJ();
label_L103709560:
PUSH_POINTER(JAVA_NULL); /* ACONST_NULL */
releaseForReturn(threadStateData, cn1LocalsBeginInThread);
return POP_OBJ();
}
--
You received this message because you are subscribed to the Google Groups
"CodenameOne Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to codenameone-discussions+unsubscr...@googlegroups.com.
Visit this group at https://groups.google.com/group/codenameone-discussions.
To view this discussion on the web visit
https://groups.google.com/d/msgid/codenameone-discussions/a35b4b8d-2d80-4d05-9346-dd1e3154ffbc%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
[codenameone-discussions] Re: A new IOS hard crash
Here is the actual java code corresponding to the culprit function.
// this is used only during board construction, when we populate
> // the board array with cells, or subtypes of cells.
> void SetBoardCell(char col,int row,CELLTYPE con)
> {int y = BCtoYindex(col,row);
> int x = BCtoXindex(col,row);
> char nc = XindexToBC(x,y);
> int nr = YindexToBC(x,y);
> // a little consistency check
> //System.out.println("C "+nc + col);
> G.Assert(((nc==col) && (nr==row)),"%s,%s encodes
> correctly",col,row);
> board[x][y]=con;
> }
>
> The crash corresponds to the internal bounds check on the board[x][y] =
con line.
On Friday, November 11, 2016 at 2:42:31 PM UTC-8, Dave Dyer wrote:
>
>
> A new IOS hard crash has appeared. I reproduced this by building from
> the sources on my mac. It's going to be really
> hard to extract a small test case without some clue what features of the
> java sources are triggering the problem. Note
> that this is not recently changed code, and previous builds worked fine.
>
> Here's the .m code that is crashing. The crash is at the
> "check_array_acccess" near the bottom, the error
> in the IOS simulator is EXC_BAD_ACCESS code-1 address = 0x36.
>
> My working hypothesis would be that some recently tweaked optimization
> has resulted in a mismanaged stack.
>
> JAVA_VOID
>> online_game_gBoard_SetBoardCell___char_int_online_game_cell(CODENAME_ONE_THREAD_STATE,
>>
>> JAVA_OBJECT __cn1ThisObject, JAVA_CHAR __cn1Arg1, JAVA_INT __cn1Arg2,
>> JAVA_OBJECT __cn1Arg3) {
>> JAVA_INT ilocals_5_; /* v5 */
>> JAVA_INT ilocals_4_; /* v4 */
>> JAVA_INT ilocals_7_; /* v7 */
>> JAVA_INT ilocals_6_; /* v6 */
>> DEFINE_INSTANCE_METHOD_STACK(6, 8, 0, 11210, 11226);
>> locals[0].data.o = __cn1ThisObject; locals[0].type =
>> CN1_TYPE_OBJECT; JAVA_INT ilocals_1_ = __cn1Arg1;
>> JAVA_INT ilocals_2_ = __cn1Arg2;
>> locals[3].data.o = __cn1Arg3;
>> locals[3].type = CN1_TYPE_OBJECT;
>> BC_ALOAD(0);
>> /* CustomInvoke */{ JAVA_INT tmpResult =
>> virtual_online_game_gBoard_BCtoYindex___char_int_R_int(threadStateData,
>> SP[-1].data.o, ilocals_1_, ilocals_2_);
>> SP[-1].data.i = tmpResult; SP[-1].type = CN1_TYPE_INT; }
>> BC_ISTORE(4);
>> BC_ALOAD(0);
>> /* CustomInvoke */{ JAVA_INT tmpResult =
>> virtual_online_game_gBoard_BCtoXindex___char_int_R_int(threadStateData,
>> SP[-1].data.o, ilocals_1_, ilocals_2_);
>> SP[-1].data.i = tmpResult; SP[-1].type = CN1_TYPE_INT; }
>> BC_ISTORE(5);
>> BC_ALOAD(0);
>> /* CustomInvoke */{ JAVA_INT tmpResult =
>> virtual_online_game_gBoard_XindexToBC___int_int_R_char(threadStateData,
>> SP[-1].data.o, ilocals_5_, ilocals_4_);
>> SP[-1].data.i = tmpResult; SP[-1].type = CN1_TYPE_INT; }
>> BC_ISTORE(6);
>> BC_ALOAD(0);
>> /* CustomInvoke */{ JAVA_INT tmpResult =
>> virtual_online_game_gBoard_YindexToBC___int_int_R_int(threadStateData,
>> SP[-1].data.o, ilocals_5_, ilocals_4_);
>> SP[-1].data.i = tmpResult; SP[-1].type = CN1_TYPE_INT; }
>> BC_ISTORE(7);
>> if (ilocals_6_!=ilocals_1_) /* IF_ICMPNE CustomJump */ goto
>> label_L1749010499;
>> if (ilocals_7_!=ilocals_2_) /* IF_ICMPNE CustomJump */ goto
>> label_L1749010499;
>> PUSH_INT(1); /* ICONST_1 */
>> goto label_L863296557;
>>
>> label_L1749010499:
>> PUSH_INT(0); /* ICONST_0 */
>>
>> label_L863296557:
>> /* LDC: '%s,%s encodes correctly'*/
>> PUSH_POINTER(STRING_FROM_CONSTANT_POOL_OFFSET(11227));
>> PUSH_INT(2); /* ICONST_2 */
>> SP--;
>> PUSH_POINTER(__NEW_ARRAY_java_lang_Object(threadStateData,
>> SP[0].data.i));
>> BC_DUP(); /* DUP */
>> PUSH_INT(0); /* ICONST_0 */
>> /* CustomInvoke
>> */PUSH_OBJ(java_lang_Character_valueOf___char_R_java_lang_Character(threadStateData,
>>
>> ilocals_1_));
>> CHECK_ARRAY_ACCESS(3, SP[-2].data.i); { /* BC_AASTORE */
>> JAVA_OBJECT aastoreTmp = SP[-3].data.o;
>> ((JAVA_ARRAY_OBJECT*) (*(JAVA_ARRAY)aastoreTmp).data)[SP[-2].data.i]
>> = SP[-1].data.o;
>> SP -= 3; }
>> BC_DUP(); /* DUP */
>> PUSH_INT(1); /* ICONST_1 */
>> /* CustomInvoke
>> */PUSH_OBJ(java_lang_Integer_valueOf___int_R_java_lang_Integer(threadStateData,
>>
>> ilocals_2_));
>> CHECK_ARRAY_ACCESS(3, SP[-2].data.i); { /* BC_AASTORE */
>> JAVA_OBJECT aastoreTmp = SP[-3].data.o;
>> ((JAVA_ARRAY_OBJECT*) (*(JAVA_ARRAY)aastoreTmp).data)[SP[-2].data.i]
>> = SP[-1].data.o;
>> SP -= 3; }
>> { JAVA_INT tmpResult =
>> lib_G_Assert___boolean_java_lang_String_java_lang_Object_1ARRAY_R_boolean(threadStateData,
>>
>> SP[-3].data.i, SP[-2].data.o, SP[-1].data.o);
>> SP-=2;
>> SP[-1].data.i = tmpResult; SP[-1].type = CN1_TYPE_INT; }
>> SP--; /* POP */
>> PUSH_POINTER(get_field_online_game_gBoard_board(__cn1ThisObject));
>> (*SP).type = CN1_TYPE_INT; /* ILOAD */
>> (*SP).data.i = ilocals_5_
