Closed
Bug 368859
Opened 18 years ago
Closed 17 years ago
large sharp variable numbers are silently rounded down
Categories
(Core :: JavaScript Engine, defect)
Core
JavaScript Engine
Tracking
()
VERIFIED
FIXED
People
(Reporter: sync2d, Assigned: sync2d)
Details
(Keywords: verified1.8.0.12, verified1.8.1.4)
Attachments
(1 file, 1 obsolete file)
935 bytes,
patch
|
brendan
:
review+
dveditz
:
approval1.8.1.4+
dveditz
:
approval1.8.0.12+
|
Details | Diff | Splinter Review |
Without the patch: $ dbg.obj/js -e "print(function(){ return #65535#, #65536#; });" function () { return #65535#, #0#; } With the patch: $ dbg.obj/js -e "print(function(){ return #65535#, #65536#; });" -e:1: SyntaxError: overlarge sharp variable number: -e:1: print(function(){ return #65535#, #65536#; }); -e:1: ..................................^
Attachment #253494 -
Flags: review?(brendan)
Comment 1•17 years ago
|
||
Comment on attachment 253494 [details] [diff] [review] use the appropriate limit Yikes! Why'd I couple to the wrong limit (no atom index [JOF_CONST format] ops for sharp var numbers, they're 16-bit [JOF_UINT16]). Thanks for patching this, I will check it in now. /be
Attachment #253494 -
Flags: review?(brendan)
Attachment #253494 -
Flags: review+
Attachment #253494 -
Flags: approval1.8.1.3?
Attachment #253494 -
Flags: approval1.8.0.11?
Comment 2•17 years ago
|
||
Safe rider for patch releases. js/src/jsscan.c 3.119 Patch I landed in a sec, for posterity. /be
Status: ASSIGNED → RESOLVED
Closed: 17 years ago
Resolution: --- → FIXED
Comment 3•17 years ago
|
||
Pedantic, but I figured it might help future opcode format readers and revisers. /be
Comment 4•17 years ago
|
||
Comment on attachment 255286 [details] [diff] [review] use UINT16_LIMIT from jsopcode.h This is the one that should ride along. /be
Attachment #255286 -
Flags: review+
Attachment #255286 -
Flags: approval1.8.1.3?
Attachment #255286 -
Flags: approval1.8.0.11?
Updated•17 years ago
|
Attachment #253494 -
Attachment is obsolete: true
Attachment #253494 -
Flags: review+
Attachment #253494 -
Flags: approval1.8.1.3?
Attachment #253494 -
Flags: approval1.8.0.11?
Comment 5•17 years ago
|
||
/cvsroot/mozilla/js/tests/js1_5/extensions/regress-368859.js,v <-- regress-368859.js
Flags: in-testsuite+
Comment 6•17 years ago
|
||
verified fixed 20070217 1.9.0 windows/mac*/linux
Status: RESOLVED → VERIFIED
Comment 7•17 years ago
|
||
Comment on attachment 255286 [details] [diff] [review] use UINT16_LIMIT from jsopcode.h approved for 1.8/1.8.0 branches, a=dveditz for drivers
Attachment #255286 -
Flags: approval1.8.1.4?
Attachment #255286 -
Flags: approval1.8.1.4+
Attachment #255286 -
Flags: approval1.8.0.12?
Attachment #255286 -
Flags: approval1.8.0.12+
Updated•17 years ago
|
Whiteboard: needs branch landing
Updated•17 years ago
|
Whiteboard: needs branch landing → needs branch landing (brendan did trunk)
Comment 8•17 years ago
|
||
Landed on branches: 1.8 - jsscan.c: 3.81.2.28 1.8.0 - jsscan.c: 3.81.2.7.2.10
Keywords: fixed1.8.0.12,
fixed1.8.1.4
Comment 9•17 years ago
|
||
For 1.8.0, we don't have the UINT16_LIMIT define, so I am just using ((uintN)1 << 16) directly. jsscan.c: 3.81.2.7.2.11
Comment 10•17 years ago
|
||
verified fixed windows, macppc, linux 1.8.0, 1.8.1 20070420 shell.
Whiteboard: needs branch landing (brendan did trunk)
You need to log in
before you can comment on or make changes to this bug.
Description
•