This bug is to be used to track the work required to release another release candidate for Spidermonkey 1.5 based on the MOZILLA_1_8_BRANCH. The plan is to release very soon (like this week) unless there are bugs blocking it. If you know of bugs which should block the release please add them to the depends list here.

Horse, barn door, gone -- ba-bye horsie! (as my daughter Natalie would say). JS_VERSION is now 160 by default (jsconfig.h), so this should be 1.6 RC1. This time, we *will* stop RC'ing soon, and have a final SpiderMonkey 1.6 release (after which I guess we'll have to do dot releases, if necessary). SpiderMonkey 1.5 went through RCs till now, without a final release. Can we retrospectively dub one of them "final"? Hrm, we didn't do one in quite a while, so missed out on some fixes for bugs people hit often enough (see the newsgroup). The last RC, IIRC was RC6a, on 2004-06-09. We could dub the last trunk stabilization tag or branch-tag cut before E4X went in as the "final" 1.5 release (candidate). That tag would be MOZILLA_1_8a4_RELEASE. It's still ugly, because that was old enough that people will want fixes done more recently, although perhaps without wanting the E4X changes. Thoughts? /be

Brendan has okayed a plan where we will make a mini-branch soon (like next week) where selected fixes can be landed for the Spidermonkey 1.6 release even if they don't make Firefox 1.5/Gecko 1.8.

Just wanted to note that these cherry-picked fixes that don't make mozilla1.8 should not be visible to JS users. The UTF-8 compile-time support is a perfect example: an API change that we won't compile into Firefox, XULRunner, or other 1.8-based apps. It may well be that no other such exceptions will arise. /be

(In reply to comment #3) > Just wanted to note that these cherry-picked fixes that don't make mozilla1.8 > should not be visible to JS users. And (beating a dead horse?) this is because we want "JS1.6" to be the same in 1.8 and derived products such as Firefox 1.5, and in SpiderMonkey 1.6 and its products and embeddings. /be

It was suggested in bug 316134 that the directory created from the tar ball contain the version number, e.g. js-1.6.tar.gz would create a directory js-1.6 instead of just js. Make it so.

This patch rolls up the following patches which were determined from the checkins to either the 1.8 branch or trunk. It will be applied to the source tagged FIREFOX_1_5_RELEASE after I create a branch. I'll begin testing soon, so if you would like other bugs fixed in spidermonkey 1.6, please let me know. Bonsai Date Bug(s) 2005-11-04-17-06 232182 2005-11-10-14-19 315925 2005-11-11-12-20 315990 2005-11-30-18-26 315436 2005-12-01-13-05 315436 2005-12-02-18-31 215173 2005-12-02-23-32 318873 2005-12-05-13-35 318767 2005-12-10-22-44 319264 2005-12-10-22-49 319264 2005-12-19-17-20 318922-319391 2006-01-05-11-26 321971 2006-01-06-08-47 316862 2006-01-06-14-57 315797 2006-01-09-11-56 320032 2006-01-09-12-49 321549 2006-01-10-11-29 322499 2006-01-10-16-22 316885-317714 2006-01-10-17-15 322045 2006-01-10-21-02 313929 2006-01-18-09-17 323765 2006-02-07-00-25 325521 2006-02-17-23-13 315974 2006-02-19-09-52 318402 2006-02-19-23-44 232182 2006-02-22-12-40 323529 2006-02-22-12-46 326453 2006-02-22-12-47 327564 2006-02-22-15-39 325269 2006-02-22-15-42 328037 2006-02-27-17-55 321874 2006-02-27-18-00 324650 2006-03-02-12-51 328897 2006-03-02-23-03 232182 2006-03-10-20-15 328589

Comment on attachment 214789 [details] [diff] [review] rollup patch obviously missed something: c:/work/mozilla/builds/ff/spidermonkey-1.6/mozilla/js/src/jsinterp.c(2948) : error C2181: illegal else without matching if c:/work/mozilla/builds/ff/spidermonkey-1.6/mozilla/js/src/jsinterp.c(3385) : warning C4013: 'DO_NEXT_OP' undefined; ass uming extern returning int

(In reply to comment #7) > (From update of attachment 214789 [details] [diff] [review] [edit]) > obviously missed something: > c:/work/mozilla/builds/ff/spidermonkey-1.6/mozilla/js/src/jsinterp.c(2948) : > error C2181: illegal else without matching > if > c:/work/mozilla/builds/ff/spidermonkey-1.6/mozilla/js/src/jsinterp.c(3385) : > warning C4013: 'DO_NEXT_OP' undefined; ass > uming extern returning int Bad merge -- you don't want any of the DO_NEXT_OP or related stuff from bug 121414. Revert all such to the corresponding break or goto out from the same place in the 1.8 branch code. /be

applying 2006-01-10 16:29 brendan%mozilla.org mozilla/js/src/jsinterp.c 3.181.2.21 MOZILLA_1_8_BRANCH 6/18 Oops. fixed it.

rollup patch v2 for: checkin date bug(s) 2005-11-04 17:06 232182 2005-11-10 14:19 315925 2005-11-11 12:20 315990 2005-11-30 18:26 315436 2005-12-01 13:05 315436 2005-12-02 18:31 215173 2005-12-02 23:32 318873 2005-12-05 13:35 318767 2005-12-10 22:44 319264 2005-12-10 22:49 319264 2005-12-19 17:20 318922-319391 2006-01-05 11:26 321971 2006-01-06 08:47 316862 2006-01-06 14:57 315797 2006-01-09 11:56 320032 2006-01-09 12:49 321549 2006-01-10 11:29 322499 2006-01-10 16:22 316885-317714 2006-01-10 16:29 316885-317714 2006-01-10 17:15 322045 2006-01-10 21:02 313929 2006-01-18 09:17 323765 2006-02-07 00:25 325521 2006-02-17 23:13 315974 2006-02-19 09:52 318402 2006-02-19 23:44 232182 2006-02-22 12:40 323529 2006-02-22 12:46 326453 2006-02-22 12:47 327564 2006-02-22 15:39 325269 2006-02-22 15:42 328037 2006-02-27 17:55 321874 2006-02-27 18:00 324650 2006-03-02 12:51 328897 2006-03-02 23:03 232182 2006-03-10 20:15 328589 2006-03-12 01:35 329749 2006-03-13 09:00 329530 |SAVE_SP_AND_PC| replaced by |SAVE_SP(fp); fp->pc = pc;| 2006-03-15 01:23 330169 2006-03-25 11:26 331664 2006-03-25 13:54 331678 2006-03-25 19:51 331719 2006-03-25 21:02 331719 This builds but I haven't tested it yet. Still need: bug 317250 patch needs to be ported to 1.8 bug 312385 needs to be decided upon and fixed or removed from blocker list. bug 326281 needs to be fixed with |SAVE_SP_AND_PC| replaced by |SAVE_SP(fp); fp->pc = pc;|

the branch is being cut this weekend for a release of rc1 early next week. If you have anything you want in rc1 please land it either on the trunk or the 1.8 branch soonest.

2005-11-04 17:06 232182 2005-11-10 14:19 315925 2005-11-11 12:20 315990 2005-11-30 18:26 315436 2005-12-01 13:05 315436 2005-12-02 18:31 215173 2005-12-02 23:32 318873 2005-12-05 13:35 318767 2005-12-10 22:44 319264 2005-12-10 22:49 319264 2005-12-19 17:20 318922-319391 2006-01-05 11:26 321971 2006-01-06 08:47 316862 2006-01-06 14:57 315797 2006-01-09 11:56 320032 2006-01-09 12:49 321549 2006-01-10 11:29 322499 2006-01-10 16:22 316885-317714 2006-01-10 16:29 316885-317714 2006-01-10 17:15 322045 2006-01-10 21:02 313929 2006-01-18 09:17 323765 2006-02-07 00:25 325521 2006-02-17 23:13 315974 2006-02-19 09:52 318402 2006-02-19 23:44 232182 2006-02-22 12:40 323529 2006-02-22 12:46 326453 2006-02-22 12:47 327564 2006-02-22 15:39 325269 2006-02-22 15:42 328037 2006-02-27 17:55 321874 2006-02-27 18:00 324650 2006-03-02 12:51 328897 2006-03-02 23:03 232182 2006-03-10 20:15 328589 2006-03-12 01:35 329749 2006-03-13 09:00 329530 |SAVE_SP_AND_PC(fp)| replaced by |SAVE_SP(fp); fp->pc = pc;| 2006-03-15 01:23 330169 2006-03-25 11:26 331664 2006-03-25 13:54 331678 2006-03-25 19:51 331719 2006-03-25 21:02 331719 2006-03-27 15:54 331793 2006-03-30 17:28 331787 2006-03-31 16:01 332415 2006-04-02 12:08 332472 2006-04-03 10:18 326281 |SAVE_SP_AND_PC(fp)| replaced by |SAVE_SP(fp); fp->pc = pc;|

|SAVE_SP_AND_PC(fp)| replaced by |SAVE_SP(fp); fp->pc = pc;| Don't need to set fp->pc in the branch versions -- it's set for all opcodes at the top of the interpreter loop. /be

created tag SPIDERMONKEY_1_6_0_BASE as a synonym for mozilla/js for the FIREFOX_1_5_RELEASE tag and branch SPIDERMONKEY_1_6_0_BRANCH where this will be checked in. final version of the patch against SPIDERMONKEY_1_6_0_BRANCH without the fp->pc = pc;

Comment on attachment 217059 [details] [diff] [review] roll up patch v4 could have sworn I put brendan in the reviewer field...

Comment on attachment 217059 [details] [diff] [review] roll up patch v4 rs=me, my eyes glazed over at some point. Tell you what: check it into the new branch, then diff MOZILLA_1_8_BRANCH against that branch for all the js*.[cth]* files and put that unified diff in this bug. That diff should be nearly empty, and we should understand why every hunk in it is there. /be

checked in the roll up patch plus the change for the js version from bug 67111.

cvs -z3 -q diff -pu10 -r SPIDERMONKEY_1_6_0_BRANCH -r MOZILLA_1_8_BRANCH js*.[cth] > SPIDERMONKEY_1_6_0_BRANCH-MOZILLA_1_8_BRANCH.diff

<http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=all&branch=SPIDERMONKEY_1_6_0_BRANCH&branchtype=match&dir=mozilla%2Fjs%2Fsrc&file=&filetype=regexp&who=&whotype=regexp&sortby=Date&hours=2&date=day&mindate=&maxdate=&cvsroot=%2Fcvsroot>

Wow, lots of stuff in 1.8 that we want in JS1.6. At least one bad jsinterp.c merge (not fatal, just redundant). Everything except the JS_C_STRINGS_ARE_UTF8 changes, I think, should go into JS1.6, since it's going into 1.8.0.3 (a diff between the 1.8 and 1.8.0 branches would be helpful to prove this). Here's the blow by blow: [items below are JS1.6 but not 1.8 branch unless noted] js.c my_*ErrorReporter removal JS_C_STRINGS_ARE_UTF8 and related exception message Unicode changes JS_YieldRequest fix JS_GetImplementationVersion [1.8 but not JS1.6] JS_SetGCParameter changes JS_SealObject unlocking fix jsapi.c trailing whitespace removal jsdhash.h regparm #if !defined(XP_MACOSX) jsemit.c diff looks like a missing fix, not sure about where missing [1.8 but not JS1.6] call_enumerate and fun_enumerate bugfixes [1.8 but not JS1.6] LENGTH_PROP_ATTRS and fun_getProperty fixes [1.8 but not JS1.6] fun_call and fun_apply JS_ValueToString error checking (should use js_ValueToString!) [1.8 but not JS1.6] Function uses js_Function_str instead of "Function" [1.8 but not JS1.6] js_ValueToFunctionObject checks principals access [1.8 but not JS1.6] jsinterp.c inline_call: fix to put call object, based on jsobj.c obj_eval change to get call object do_nonint_fast_global_incop: in the JS1.6 branch has redundant const JSCodeSpec *cs local and len assignment wrongly merged from trunk threaded interpreter -- REMOVE [1.8 but not JS1.6] script leak on access check failure just before js_Execute call in obj_eval [1.8 but not JS1.6] js_HasOwnPropertyHelper outerObject fix [1.8 but not JS1.6] js_NewObject GC safety fix [1.8 but not JS1.6] jsopcode.c VarPrefix fix Decompile 'else ok = JS_FALSE' after js_GetPrinterOutput failure fix js_DestroyPrinter call mislocated in js_DecompileValueGenerator fix [1.8 but not JS1.6] deprecated with statement warning removed from Statement in jsparse.c [1.8 but not JS1.6] pn2->pn_arity instead of pn->pn_arity fix to FoldXMLConstants [1.8 but not JS1.6] HOLD_REGEXP relocation and fix in jsstr.c match_or_replace [1.8 but not JS1.6] jsscan.c GrowStringBuffer fix [1.8 but not JS1.6] js_GetToken XMLONLYMODE fix [1.8 but not JS1.6] JS_SetGCParameter related fixes to account gcMallocBytes in jsscope.c [1.8 but not JS1.6] match_or_replace error swallowing and DROP_REGEXP bugfixes [1.8 but not JS1.6] jsxml.c bugfixes we want Need to cvs annotate and id the above bugs. I don't have time right now. Help appreciated. /be

(In reply to comment #20) The following bugs where included in the patch I applied to js16 but have not had a checkin to 1.8. You can query the dependencies for this bug with blocking1.8.1[?+] Bug 215173 - JS_ReportErrorNumber doesn't raise the user-specified exception Bug 232182 - Can't display non-ascii characters in JS exceptions Bug 315436 - Setting a read-only property does not generate a warning in strict mode Bug 315925 - Fix some TestUtf8 nits Bug 315974 - JSprintf functions cannot print jschar characters and strings Bug 315990 - "this.statement.causes.an.error" is ignored instead of throwing Bug 318402 - Rename JS_STRINGS_ARE_UTF8 to JS_C_STRINGS_ARE_UTF8 Bug 318767 - js_ReportOutOfMemory needs to support enhanced error reporting Bug 318873 - Crash in js_GetLocalizedErrorMessage following 215173 checkin Bug 319264 - Double-free reporting errors Bug 323765 - Thread unsafe VarPrefix from jsopcode.c Bug 325521 - JS_YieldRequest() allows the garbage collector to race Bug 328589 - VC8 doesn't like copyright sign in jscompat.h Bug 329530 - Out of memory crash when calling fn.toString where fn is a deeply nested function (jsopcode.c patch only) Bug 329749 - Did JS_SealObject mean to call JS_UNLOCK_OBJ instead of JS_UNLOCK_SCOPE ? Bug 330169 - ParseNodeToXML() leaves local root stack under certain circumstances > Wow, lots of stuff in 1.8 that we want in JS1.6. At least one bad jsinterp.c > merge (not fatal, just redundant). Everything except the JS_C_STRINGS_ARE_UTF8 this was Bug 318402 which you nominated for js16. > changes, I think, should go into JS1.6, since it's going into 1.8.0.3 (a diff > between the 1.8 and 1.8.0 branches would be helpful to prove this). > I'll attach that in a moment. It looks to me that this is a complete mess. What do you say about I revert my patch to the SPIDERMONKEY_1_6_0_BRANCH, then we get all of the fixes listed above landed on the 1.8 branch and I create a new rollup patch and we move on from there?

This contains my estimates of the bugs/patches requested. Take it with a cow-sized salt block.

(In reply to comment #6) > tagged FIREFOX_1_5_RELEASE after I create a branch. I'll begin testing soon, so > if you would like other bugs fixed in spidermonkey 1.6, please let me know. I have reported 335652(request for 213257 in .tar) and 336616 Perhaps I am too late? But I have not found a js-1.6.tar online In any case many thanks for your help Bill

Is this happening? Embedders are still working with 1.5-Final which is from Sept. 24, 2004 according to the release notes... or is there somewhere else to get something newer/more stable?

Should this be WONTFIXed?

This patch applies the changes from bug 232182, 315925, 315974, 318402 to the MOZILLA_1_8_0_BRANCH and is intended to add the JS_C_STRINGS_ARE_UTF8 support to the 1.8.0 branch. I ended up adding an additional message JSMSG_USER_DEFINED_ERROR and moving JSMSG_WRONG_CONSTRUCTOR to make them match 1.8. I don't know if that was kosher or not. I've build SpiderMonkey shell opt and debug and Firefox opt and debug with these changes, but haven't run the test suite yet.

fyi, no regressions found in comparing js test suite runs for base 1.8.0 and patched 1.8.0.

Thanks, your patch looks good from a number of angles (I haven't read it all the way through, but I've applied it and interdiffed it against 1.8 branch). This is: * a bit of cleanup that followed the JS_C_STRINGS_ARE_UTF8 patch in; * patch(1) rejected a js.msg hunk so I did that by hand and avoided moving stuff * jsapi.c needed warning fixes for x86_64 and other platforms where size_t!=uintN. /be

Comment on attachment 243710 [details] [diff] [review] tweaked utf8 code to match trunk/1.8, untweaked js.msg applies cleanly to 1.8.0 and there are no signficant differences in the test suite results for winxp browser|windows. I would like to get this in for 1.8.0.9 along with the version change and release sm16 immediately after ff1.5.0.8.

Please, be sure to release updated self-tests as well. (Bug 303857). Thanks!

bump 1.8.1.1/trunk JS_GetImplementationVersion to JavaScript-C 1.7 pre-release 2 2006-11-19

bump 1.8.0 JS_GetImplementationVersion to JavaScript-C 1.6 2006-11-19

Comment on attachment 245861 [details] [diff] [review] jsapi.c 1.8.0 patch approved for 1.8.0 branch once brendan gives r=, a=dveditz

Comment on attachment 245860 [details] [diff] [review] jsapi.c trunk|1.8 patch approved for 1.8 branch, a=dveditz for drivers

Comment on attachment 243710 [details] [diff] [review] tweaked utf8 code to match trunk/1.8, untweaked js.msg approved for 1.8.0, a=dveditz for drivers

attachment 243710 [details] [diff] [review], attachment 245861 [details] [diff] [review] checked into MOZILLA_1_8_BRANCH Checking in js.c; /cvsroot/mozilla/js/src/js.c,v <-- js.c new revision: 3.93.2.2.2.6; previous revision: 3.93.2.2.2.5 done Checking in js.msg; /cvsroot/mozilla/js/src/js.msg,v <-- js.msg new revision: 3.43.8.2.2.3; previous revision: 3.43.8.2.2.2 done Checking in jsapi.c; /cvsroot/mozilla/js/src/jsapi.c,v <-- jsapi.c new revision: 3.214.2.11.2.10; previous revision: 3.214.2.11.2.9 done Checking in jsapi.h; /cvsroot/mozilla/js/src/jsapi.h,v <-- jsapi.h new revision: 3.107.2.4.2.3; previous revision: 3.107.2.4.2.2 done Checking in jsatom.c; /cvsroot/mozilla/js/src/jsatom.c,v <-- jsatom.c new revision: 3.65.4.1.4.1; previous revision: 3.65.4.1 done Checking in jscntxt.c; /cvsroot/mozilla/js/src/jscntxt.c,v <-- jscntxt.c new revision: 3.60.4.2.2.1; previous revision: 3.60.4.2 done Checking in jsdbgapi.c; /cvsroot/mozilla/js/src/jsdbgapi.c,v <-- jsdbgapi.c new revision: 3.56.2.1.4.7; previous revision: 3.56.2.1.4.6 done Checking in jsobj.c; /cvsroot/mozilla/js/src/jsobj.c,v <-- jsobj.c new revision: 3.208.2.12.2.19; previous revision: 3.208.2.12.2.18 done Checking in jsopcode.c; /cvsroot/mozilla/js/src/jsopcode.c,v <-- jsopcode.c new revision: 3.89.2.8.2.10; previous revision: 3.89.2.8.2.9 done Checking in jsprf.c; /cvsroot/mozilla/js/src/jsprf.c,v <-- jsprf.c new revision: 3.19.30.1; previous revision: 3.19 done Checking in jsprf.h; /cvsroot/mozilla/js/src/jsprf.h,v <-- jsprf.h new revision: 3.8.28.1; previous revision: 3.8 done Checking in jspubtd.h; /cvsroot/mozilla/js/src/jspubtd.h,v <-- jspubtd.h new revision: 3.51.8.2.2.1; previous revision: 3.51.8.2 done Checking in jsscan.c; /cvsroot/mozilla/js/src/jsscan.c,v <-- jsscan.c new revision: 3.81.2.7.2.9; previous revision: 3.81.2.7.2.8 done Checking in jsstr.c; /cvsroot/mozilla/js/src/jsstr.c,v <-- jsstr.c new revision: 3.108.2.5.2.6; previous revision: 3.108.2.5.2.5 done Checking in jsstr.h; /cvsroot/mozilla/js/src/jsstr.h,v <-- jsstr.h new revision: 3.27.20.1; previous revision: 3.27 done Checking in jsxml.c; /cvsroot/mozilla/js/src/jsxml.c,v <-- jsxml.c new revision: 3.50.2.15.2.28; previous revision: 3.50.2.15.2.27 done

(In reply to comment #38) should have read: > attachment 243710 [details] [diff] [review], attachment 245861 [details] [diff] [review] checked into MOZILLA_1_8_0_BRANCH

attachment 245860 [details] [diff] [review] checked into MOZILLA_1_8_BRANCH Checking in jsapi.c; /cvsroot/mozilla/js/src/jsapi.c,v <-- jsapi.c new revision: 3.214.2.33; previous revision: 3.214.2.32 done

Adding missing keywords. This is why you got the e-mail from Jay just a little bit ago. :)

jsapi.c trunk|1.8 patch attachment 245860 [details] [diff] [review] checked into trunk: /cvsroot/mozilla/js/src/jsapi.c,v <-- jsapi.c new revision: 3.292; previous revision: 3.291

finally fixed.