404499 - js1_5/Regress/regress-168347.js FAIL - browser|shell

Status: VERIFIED FIXED

(Core :: JavaScript Engine, defect)

Description

[Bob Clary [:bc] (inactive)]

BUGNUMBER: 168347
STATUS: Testing F.toString()
FAILED! expected: [reported from test()] Expected value 'functionF(){varf=arguments.callee;f.i=0;try{f.i=f.i+1;print("i=i+1succeededi="+f.i);}catch(e){print(', Actual value 'functionF(){varf=arguments.callee;f.i=0;try{f.i=f.f+1;print("i=i+1succeededi="+f.f);}catch(e){print('
FAILED! expected: [reported from test()] Expected value '"i=i+1failedwith"+e+"i="+f.i);}try{++f.i;print("++isucceededi="+f.i);}catch(e){print("++ifailedwith"', Actual value '"i=i+1failedwith"+e+"i="+f.f);}try{++f.i;print("++isucceededi="+f.f);}catch(e){print("++ifailedwith"'
FAILED! expected: [reported from test()] Expected value '+e+"i="+f.i);}try{f.i++;print("i++succeededi="+f.i);}catch(e){print("i++failedwith"+e+"i="+f.i);}try', Actual value '+e+"i="+f.f);}try{f.i++;print("i++succeededi="+f.f);}catch(e){print("i++failedwith"+e+"i="+f.f);}try'
FAILED! expected: [reported from test()] Expected value '{--f.i;print("--isucceededi="+f.i);}catch(e){print("--ifailedwith"+e+"i="+f.i);}try{f.i--;print("i--', Actual value '{--f.i;print("--isucceededi="+f.f);}catch(e){print("--ifailedwith"+e+"i="+f.f);}try{f.i--;print("i--'
FAILED! expected: [reported from test()] Expected value 'succeededi="+f.i);}catch(e){print("i--failedwith"+e+"i="+f.i);}}', Actual value 'succeededi="+f.f);}catch(e){print("i--failedwith"+e+"i="+f.f);}}'
PASSED! Section 6 of test -
PASSED! Section 7 of test -
PASSED! Section 8 of test -
PASSED! Section 9 of test - 

regressed in the last day, maybe the hidden properties bug 398609. I'll try to get a better range after I take care of some other stuff.

Comment 1

[Bob Clary [:bc] (inactive)]

I think js1_8/decompilation/regress-381504.js is the same issue:

Expected value ' function _4 ( ) { var bar ; return foo [ bar . baz ] ; } ', Actual value ' function _4 ( ) { var bar ; return foo [ bar . bar ] ; } '

Comment 2

[Igor Bukanov]

I have not seen the regression because the driver does not report the failures even if the output comes with errors:

~/m/trunk/mozilla/js/tests $ ./jsDriver.pl -e smdebug -l js1_5/Regress/regress-168347.js -K
-#- Executing 1 test(s).
-#- Wrote results to 'results-2007-11-20-161256-smdebug.html'.
-#- 0 test(s) failed

For regression testing I typically run the test suite twice, before and after the patch, using:

./jsDriver.pl -e smdebug -L slow-n.tests performance.tests

and then diff the list of failures. Since the failures where not reported as such, I was not aware about them.

Comment 3

[Igor Bukanov]

Attachment: fix v1

The bug was in BEGIN_LITOPX_CASE(JSOP_GETARGPROP, ARGNO_LEN) case in the decompiler. There I assigned the atom overwriting the value loaded by the macro. To prevent such bugs in future I replaced all BEGIN_LITOPX_CASE with explicit case: LOAD_ATOM(); sequences.

Comment 4

[Bob Clary [:bc] (inactive)]

(In reply to comment #2)

> 
> and then diff the list of failures. Since the failures where not reported as
> such, I was not aware about them.

I'll file a bug and fix that. Use -R for now since that will show the failures more reliably I think.

Comment 5

[Igor Bukanov]

Ping for a review: I think this regression busted few extensions.

Comment 6

[Ronny Perinke]

The patch solved the problems with my extensions (see bug 398609 comment 47)

Comment 7

[ronin.achilles]

(In reply to comment #6)
> The patch solved the problems with my extensions (see bug 398609 comment 47)
> 

Thats good news. Hope this one lands soon so I can test the other extensions as well.

Comment 8

[Igor Bukanov]

I checked in the patch from comment 3 to the CVS trunk:

Checking in js/src/jsopcode.c;
/cvsroot/mozilla/js/src/jsopcode.c,v  <--  jsopcode.c
new revision: 3.272; previous revision: 3.271
done

Comment 9

[Igor Bukanov]

Fore references here is the bonsai link for the check in:

http://bonsai.mozilla.org/cvsquery.cgi?module=PhoenixTinderbox&branch=HEAD&cvsroot=%2Fcvsroot&date=explicit&mindate=1195718700&maxdate=1195718763&who=igor%mir2.org

Comment 10

[Bob Clary [:bc] (inactive)]

verified fixed.