Jump to content

My Browser Builds (Part 3)


Recommended Posts

If you're a user of JustOff's "GitHub/GitLab Web Components Polyfill" extension, this has been very recently updated to v1.2.4, to mitigate mainly GitLab introduced breakage...

https://github.com/JustOff/github-wc-polyfill/releases/tag/1.2.4

For those not in the know, this extension is indispensable to users of UXP-based browsers (official ones and forks) to enable them to work properly on these two popular web-hosted code repositories...

After Microsoft acquired GitHub some time ago, they let go the old team of developers and quickly turned that platform (github.com) to target mainly their new "child" of a browser, Chromium-based Edge ("affectionately" also called ChrEdge...); support for "legacy" browser engines was dropped like a hot potato, their introduction of Web Components necessitated the creation of said extension for GitHub (because the venomous :angry: upstream devs couldn't port Google's WC APIs to UXP...).

GitLab resisted more the implementation of WC on their codebase, but it has been steadily happening over the course of 2021; the extension was extended (pun intended :)) to also cover gitlab.com pages on UXP browsers...

In a move reminiscent of the pair Mozilla Firefox and Google Chrome, GitLab is now in hot pursuit of GitHub: not only are they embracing WC for everything (even emojis :angry: ...), they have recently dropped support for Firefox < 78.x.x (breaking GitLab compatibility with WaterFox Classic, SeaMonkey and, of course, UXP-based browsers :( ) ; in doing so, they introduced RegExp with Named Capture Groups (a subset of V8 regexp code), which is only found on Fx >= 78.0 ; the latest extension update to v1.2.4 tries its best to remedy above changes...

JustOff also extends an invitation to our "own" @roytam1 to consider porting RegExp with Named Capture Groups to our UXP fork, since a WaterFox Classic developer already did it there... But, without being a coder, I fear this task presents serious challenges, because WFC is Fx56-based, while UXP only Fx52-based; plus, there's a lot of code to be tested and merged as it is...

Usability details:
Those with the extension already installed in Serpent 52.9.0 should have been already auto-updated; those who want to install it for the first time on Serpent should install it directly from GitHub; those using New Moon 28 must first download the updated .xpi to disk and manually modify its install.rdf file (to target New Moon 28 proper versions) prior to installing it; I have explained all this in detail in relevant previous posts...

The extension works as intended on fairly recent versions of UXP (older versions of Serpent and NM28 - recently put in the spotlight for generating better benchmark results :whistle:- lack needed APIs to sustain the correct functioning of GitHub/GitLab pages).

In the not-so-distant past, the extension could be force-installed on FxESR 52.9.x to restore some percentage of GitHub functionality, but M$ have broken it so much during these last three months, that such an installation is now moot... 
Affected in the same manner is Serpent 55.0.0; its web engine is considerably behind UXP (but better than FxESR 52.9.x's), force-installing the extension there will now only achieve very few github.com parts working as intended... :(

Edited by VistaLover
Link to comment
Share on other sites


1 hour ago, VistaLover said:

JustOff also extends an invitation to our "own" @roytam1 to consider porting RegExp with Named Capture Groups to our UXP fork, since a WaterFox Classic developer already did it there... But, without being a coder, I fear this task presents serious challenges, because WFC is Fx56-based, while UXP only Fx52-based; plus, there's a lot of code to be tested and merged as it is...

upstream had a `new-regexp` branch for this, but I don't know how well it is.

 

EDIT: tried to build on vs2015, it breaks

 9:10.85 RegExpAPI.cpp
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(504): error C2719: 'value': formal parameter with requested alignment of 8 won't be aligned
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(527): error C2719: 'object': formal parameter with requested alignment of 8 won't be aligned
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(535): error C2719: 'object': formal parameter with requested alignment of 8 won't be aligned
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(548): error C2719: 'value': formal parameter with requested alignment of 8 won't be aligned
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(549): error C2719: 'object': formal parameter with requested alignment of 8 won't be aligned
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(585): error C2719: 'object': formal parameter with requested alignment of 8 won't be aligned
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(636): error C2719: 'value': formal parameter with requested alignment of 8 won't be aligned
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(686): note: see reference to class template instantiation 'v8::internal::Handle<T>' being compiled
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(812): error C2719: 'object': formal parameter with requested alignment of 8 won't be aligned
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(818): error C2719: 'string': formal parameter with requested alignment of 8 won't be aligned
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(915): error C2719: 'object': formal parameter with requested alignment of 8 won't be aligned
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(1055): error C2719: 'value': formal parameter with requested alignment of 8 won't be aligned
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(635): error C2719: 'object': formal parameter with requested alignment of 8 won't be aligned
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(1024): note: see reference to class template instantiation 'v8::internal::Handle<v8::internal::HeapObject>' being compiled
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-shim.h(1119): error C2719: 'object': formal parameterwith requested alignment of 8 won't be aligned
 9:10.85 c:\devel\myUXP-newexp\js\src\new-regexp/regexp.h(174): error C2719: 'key_string': formal parameter with requested alignment of 8 won't be aligned
 9:10.86 c:\devel\myUXP-newexp\js\src\new-regexp/regexp.h(174): error C2719: 'key_pattern': formal parameter with requested alignment of 8 won't be aligned
 9:10.86 c:\devel\myUXP-newexp\js\src\new-regexp/regexp.h(180): error C2719: 'cache': formal parameter with requested alignment of 8 won't be aligned
 9:10.86 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-macro-assembler.h(258): error C2719: 're_code': formal parameter with requested alignment of 8 won't be aligned
 9:10.86 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-macro-assembler.h(274): error C2719: 'input': formal parameter with requested alignment of 8 won't be aligned
 9:10.86 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-macro-assembler.h(274): error C2719: 'regexp': formal parameter with requested alignment of 8 won't be aligned
 9:10.86 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-compiler.h(503): error C2719: 'code': formal parameter with requested alignment of 8 won't be aligned
 9:10.86 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-interpreter.h(50): error C2719: 'code_array': formal parameter with requested alignment of 8 won't be aligned
 9:10.86 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-interpreter.h(50): error C2719: 'subject_string': formal parameter with requested alignment of 8 won't be aligned
 9:10.86 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-interpreter.h(55): error C2719: 'regexp': formal parameter with requested alignment of 8 won't be aligned
 9:10.86 c:\devel\myUXP-newexp\js\src\new-regexp/regexp-interpreter.h(55): error C2719: 'subject_string': formal parameter with requested alignment of 8 won't be aligned
 9:10.86
 9:10.87 In the directory  /c/devel/myUXP-newexp/obj-i686-pc-mingw32/js/src/new-regexp
 9:10.87 The following command failed to execute properly:
 9:10.87 c:/devel/myUXP-newexp/obj-i686-pc-mingw32/_virtualenv/Scripts/python.exe -m mozbuild.action.cl cl.EXE -FoRegExpAPI.obj -c -DNDEBUG=1 -DTRIMMED=1 -DENABLE_SHARED_ARRAY_BUFFER -DEXPORT_JS_API -DJS_HAS_CTYPES -DDLL_PREFIX="" -DDLL_SUFFIX=".dll" -DFFI_BUILDING -D_CRT_RAND_S -DNOMINMAX -DMOZ_HAS_MOZGLUE -Ic:/devel/myUXP-newexp/js/src/new-regexp -Ic:/devel/myUXP-newexp/obj-i686-pc-mingw32/js/src/new-regexp -Ic:/devel/myUXP-newexp/obj-i686-pc-mingw32/js/src -Ic:/devel/myUXP-newexp/js/src -Ic:/devel/myUXP-newexp/obj-i686-pc-mingw32/dist/include -Ic:/devel/myUXP-newexp/obj-i686-pc-mingw32/dist/include/nspr -MD -FI c:/devel/myUXP-newexp/obj-i686-pc-mingw32/js/src/js-confdefs.h -DMOZILLA_CLIENT -TP -nologo -utf-8 -wd4345 -wd4351 -wd4800 -wd4595 -D_CRT_SECURE_NO_WARNINGS -wd5026 -wd5027 -Zc:sizedDealloc- -Zc:threadSafeInit- -D_HAS_EXCEPTIONS=0 -W3 -Gy -Zc:inline -arch:SSE2 -FS -Gw -wd4244 -wd4267 -wd4251 -we4553 -GR- -Zi -O2 -Ob2ity -GS- -GFA -Qpar -Qfast_transcendentals -openmp -Oy- -fp:precise -wd4805 -wd4661 -we4067 -we4258 -we4275 -wd4146 -wd4577 -wd4312 -wd4275 -Fdgenerated.pdb c:/devel/myUXP-newexp/js/src/new-regexp/RegExpAPI.cpp
 9:10.88 c:/devel/myUXP-newexp/config/rules.mk:879: recipe for target 'RegExpAPI.obj' failed

 

Edited by roytam1
Link to comment
Share on other sites

On 9/22/2021 at 12:16 AM, feodor2 said:

Where they took the code, most likely from the mozilla, forgot may be they a thing on makefiles.

And the error is the same everywhere,  https://www.programmersought.com/article/37382464851/

tried with EWDK-rs4 based on your guidelines, and it compiles.

just it crash with `cx` becomes `NULL` when starting up, when using new-regexp. maybe it need to check waterfox's changes to see if there is something missed.

I think it hits https://github.com/roytam1/UXP/commit/367fc9432e210834d0b44238939f9ebc5d44d090

BTW with EWDK-rs4 VC 19.13, build time is about 10 minutes longer than vc2015u3.

Edited by roytam1
Link to comment
Share on other sites

12 hours ago, roytam1 said:

BTW with EWDK-rs4 VC 19.13, build time is about 10 minutes longer than vc2015u3.

Does this mean you'll be switching to Visual C++ 2017 (update 6)? Also, how long do builds take with those 2 compilers on your computer?

Edited by ClassicNick
Link to comment
Share on other sites

 

49 minutes ago, ClassicNick said:

Does this mean you'll be switching to Visual C++ 2017 (update 6)?


not for now, unless new-regexp is properly landed.

49 minutes ago, ClassicNick said:

how long do builds take with those 2 compilers on your computer?

a full rebuild with vc2015u3 is about 25 minutes, while EWDK-rs3 vc2017 is about 35 minutes.

build hardware environment: win7 sp1 x64, ryzen 3700x, 32gb ram, sn750 1tb for system and build drive

 

Edited by roytam1
Link to comment
Share on other sites

I'm a little envious of your computer. I'm presently trying to compile Firefox 35 on a computer with 8GB RAM (3GB usable due to Windows XP 32-bit), and an AMD 4400M CPU running at 1.7 GHz. If the build succeeds, I expect it will have taken 3 hours to compile. I plan to get a new computer in late 2023/early 2024 because I want to wait for DDR5 RAM. If I were you, I would start a "new-regexp" branch that requires Visual C++ 2017, but keep a separate branch that is regularly updated that only requires Visual C++ 2015 in order to build.

Link to comment
Share on other sites

On 9/23/2021 at 10:27 AM, ClassicNick said:

If I were you, I would start a "new-regexp" branch that requires Visual C++ 2017, but keep a separate branch that is regularly updated that only requires Visual C++ 2015 in order to build.

it does, checkout my git tree please :)

 

Quote

(spoiler belows)

 

 

 

 

 

 

 

 

people are still having fun with that issue.
https://archived.moe/g/thread/83496704/
https://albusluna.com/

 

Edited by roytam1
Link to comment
Share on other sites

New build of Serpent/UXP for XP!

Test binary:
Win32 https://o.rthost.win/basilisk/basilisk52-g4.8.win32-git-20210925-f94c0da-uxp-9ab5fe727-xpmod.7z
Win64 https://o.rthost.win/basilisk/basilisk52-g4.8.win64-git-20210925-f94c0da-uxp-9ab5fe727-xpmod.7z

source code that is comparable to my current working tree is available here: https://github.com/roytam1/UXP/commits/custom

IA32 Win32 https://o.rthost.win/basilisk/basilisk52-g4.8.win32-git-20210925-f94c0da-uxp-9ab5fe727-xpmod-ia32.7z

source code that is comparable to my current working tree is available here: https://github.com/roytam1/UXP/commits/ia32

NM28XP build:
Win32 https://o.rthost.win/palemoon/palemoon-28.10.4a1.win32-git-20210925-ba47fad4d-uxp-9ab5fe727-xpmod.7z
Win32 SSE https://o.rthost.win/palemoon/palemoon-28.10.4a1.win32-git-20210925-ba47fad4d-uxp-9ab5fe727-xpmod-sse.7z
Win64 https://o.rthost.win/palemoon/palemoon-28.10.4a1.win64-git-20210925-ba47fad4d-uxp-9ab5fe727-xpmod.7z

No official UXP changes since my last build.

No official Basilisk changes since my last build.

No official Pale-Moon changes since my last build.

My changes since my last build:
- [CSS] Add calc availability to color functions. (8a0897d23)
- [Basilisk] Fix error when installing signed webextensions (18908d2f0)
- mozbuild: remove tilde char from special character blacklist as win32 short path contains tilde character. (c45d02ceb)
- ffvpx: update ffmpeg to 3.4.9 (9ab5fe727)

* Notice: From now on, UXP rev will point to `custom` branch of my UXP repo instead of now-dead MCP UXP repo, while "official UXP changes" shows only `tracking` branch changes. MCP Basilisk/Pale-Moon rev after datecode will be removed later.

Link to comment
Share on other sites

Link to comment
Share on other sites

New NewMoon 27 Build!

32bit https://o.rthost.win/palemoon/palemoon-27.10.0.win32-git-20210925-45e41b763-xpmod.7z
32bit SSE https://o.rthost.win/palemoon/palemoon-27.10.0.win32-git-20210925-45e41b763-xpmod-sse.7z
32bit noSSE https://o.rthost.win/palemoon/palemoon-27.10.0.win32-git-20210925-45e41b763-xpmod-ia32.7z

64bit https://o.rthost.win/palemoon/palemoon-27.10.0.win64-git-20210925-45e41b763-xpmod.7z

source repo: https://github.com/roytam1/palemoon27

repo changes since my last build:
- import changes from `dev' branch of rmottola/Arctic-Fox:
 - missing bit of Bug 889158 - Fix arrow function lexical arguments binding (6dd07824a)
 - missing parts of Bug 1135708 - Implement es7 exponentiation operator. (87d1b2a09)
 - Bug 1183400 - Add a by-kind switch to perform constant-folding and eventually eliminate by-arity handling. r=efaust (ce6768404)
 - Bug 1171909 - Check for alloc failure initializing AssemblerBufferWithConstantPools and update error handling ExceptionHandlerBailout r=jandem (fb86a97ec)
 - Bug 1182428 - Consider TypeSet stuff to be GCPointers, since they can contain them, r=jonco (2a7841983)
 - Bug 1184199 - Use a Map instead of a plain object to avoid the "constructor" property, r=jonco" (3af559983)
 - Bug 1184199 - Add an explanation only if there is something to explain, r=jonco (882178fcf)
 - Bug 1182428 - Improve naming, r=jonco (24bcf0679)
 - Bug 1182428 - Recognize more rooted type names, r=jonco (75e9d692e)
 - Bug 963738 - Assume(ptr) is not a hazard; Assume(*ptr) is, r=bhackett (e704d13e6)
 - No bug. Add a --function argument for analyzing a single function by name. (d9e8fc4c0)
 - Bug 1172682 - Comment and refactor analysis code, r=terrence The one change this patch makes in functionality is that it properly computes minimumUse in more cases, in anticipation of it being used for hazards (in addition to its current use for unnecessary roots). (a481f942a)
 - Bug 1182428 - Accept more edge types in edgeCanGC, r=jonco (98413c47f)
 - Bug 1170200 - Part 1: Change the API for nsIPermissionManager::Remove() to accept a URI instead of a string; r=ehsan (bd5ef47ed)
 - Add two missing includes to SharedMemoryBasic_mach.cpp, no bug (dbdeeaca1)
 - bug 1175898 - crash in mozilla::CrossProcessMutex::ShareToProcess(int) r=billm (93cab5830)
 - bug 1204985 - make SharedMemoryBasic_mach build on iOS. r=billm (0f7b668c3)
 - bug 1204985 horrible clobber-avoidance bustage fix r=YOLO CLOSED TREE (d46b9b946)
 - Bug 1205164 - Detect message races in Mach Shmem implementation. r=blassey (2b88cc22b)
 - Bug 1221540: OS X IPC timeout retry with a longer interval. r=milan (66d2006fe) (9e300313d)
- import changes from `dev' branch of rmottola/Arctic-Fox:
 - bug 1167604 - don't hide ProxyAccessible::Parent() with DocAccessibleParent::Parent r=lsocks (32e4cac18)
 - Bug 1173418 - Check if a document is top level in DocAccessibleParent::Destroy so that RemoteDocShutdown isn't erroneously called r=tbsaunde (db280c0a1)
 - bug 1184217 - check consistancy of the document tree before removing accessibles in a hide event r=davidb (91d2f38fb)
 - bug 1162543 - make DocAccessibleChild::IDToAccessible correctly get the document r=davidb (be2eb3eb7)
 - bug 1170595 - switch to sending __delete__ from the parent instead of the child r=davidb, smaug (216a90d34) (3a8ef50a8)
- import change from `dev' branch of rmottola/Arctic-Fox:
 - MOZ_OVERRIDE -> override (624f7c712) (57cc49827)
- import changes from `dev' branch of rmottola/Arctic-Fox:
 - Bug 1182373 - Don't constant-fold in the tag component of a tagged-template, because constant-folding evaluates the tag too early (and moght convert a value into a property reference, causing the wrong |this| to be passed). r=jorendorff (f9e41352b)
 - Bug 1183400 - Constant-fold nullary nodes entirely according to their kind, not their arity. r=efaust (9c6248dd8)
 - Bug 1183400 - Fold typeof nodes by examining their kind. r=efaust (0dc443122)
 - Bug 1183400 - Fold |void <expr>| expressions by kind. r=efaust (d872286ad)
 - Bug 1183400 - Fold delete nodes by kind. r=efaust (a30634947)
 - Bug 1164774 - Remove unused code handling binary-arity PNK_ADD nodes in constant-folding. r=shu (996db297b)
 - Bug 1183400 - Fold !/~/+/- by kind. r=efaust (bf5dd5984)
 - Bug 1183400 - Pass a Parser reference to Fold, not separate handler/options, preparing for another patch that wants to have access to a parser instance. r=efaust (5bbb167be)
 - Bug 1183400 - Fold increment/decrement operations by kind. r=efaust (2728e23a7)
 - Bug 1183400 - Fold ?: expressions by kind. r=efaust (64d6ab853)
 - Bug 1183400 - Fold |if| nodes by kind. r=efaust (0f701a6bb)
 - Bug 1183400 - Fold various simple unary cases by kind. r=efaust (590b24acb)
 - Bug 1183400 - Fold and/or expressions. r=efaust (1f114d2b2)
 - Bug 1183400 - Fold function nodes by kind. r=efaust (0920010c0)
 - Bug 1183400 - Fold binary arithmetic operations by kind, not arity. r=efaust (a3723a1ec)
 - Bug 1183400 - Remove a now-unused variable, following up on the ?:/if-node folding changes. r=trivial, r=orange in a CLOSED TREE (522584dc0)
 - Bug 1183400 - Inline FoldBinaryNumeric into its sole caller and simplify code accordingly. r=efaust (043ccf349)
 - Bug 1183400 - Fold various list nodes not given, nor requiring, special treatment. r=shu (7de28f276)
 - Bug 1183400 - Constant-fold yield/yield*/return by kind and not arity. r=shu (7cae87421)
 - Bug 1183400 - Fold try/catch by kind. r=shu (5df2a97fc)
 - Bug 1183400 - Fold class nodes by kind. r=shu (67fffcc45)
 - Bug 1183400 - Fold element accesses by kind. r=efaust (82538a7a2)
 - Bug 1183400 - Fold addition by kind. r=efaust (68a0c6b49)
 - Bug 1183400 - Fold function calls and tagged templates by kind, not arity. r=efaust (0a368e8f9)
 - Bug 1183400 - Fold various binary nodes with two non-null halves by kind. r=efaust (de48c675b)
 - Bug 1183400 - Fold while and do-while loops by kind, not arity. r=efaust (998bd13cc)
 - Bug 1183400 - Remove dead fold-by-arity code. r=efaust (3b5f719b9)
 - Bug 1183400 - Remove SyntacticContext::Delete, now addressed by modifying how delete nodes are folded. r=efaust (8129f5f7c)
 - Bug 1183400 - Constant-fold switch/default by kind, not arity. r=efaust (e9430d603)
 - Bug 1183400 - Fold for-in, for-of, and for(;;) loops by kind, and classify miscellaneous straightforward kinds for folding. r=efaust (59d09a110)
 - Bug 1183400 - Replace callee-based condition-constant folding with caller-specified condition-constant folding. r=efaust (2090ea750)
 - Bug 1183400 - Remove special |void| handling by making Boolish recognize |void| expressions as falsy, when they're obviously so. r=efaust (d230cbb0d)
 - Bug 1183400 - Remove SyntacticContext::Condition, now handled context-sensitively by callers specifically requesting condition-targeted folding. r=efaust (8be79cd7a)
 - Bug 1183400 - Fold the last few nodes (including PN_NAME nodes) by kind, not arity. r=efaust (4e8e911a4)
 - Bug 1183400 - Remove SyntacticContext completely. r=efaust (745d4df15)
 - Bug 1183400 - Common up some is-effectless testing. r=efaust (ed25885e6)
 - Bug 1135322 - receiveMessage() should convert logins to nsILoginInfo objects. r=MattN (35da0f066)
 - Bug 1140242 - Convert nsILoginInfo to vanilla JS objects before trying to send them in messages. r=dolske (f952e9ab5)
 - Bug 1134846 - Add a module to support per-site password manager recipes. r=dolske (d9d686818)
 - Bug 1135451 - fillForm() cleanup part A: remove unused return type, kill E10S unfriendly fillForm from nsILoginManager, kill passwordmgr-found-form notification, largely a backout of bug 439365. r=MattN (3825eb1dd)
 - Bug 1120129 - Allow per-site recipes to adjust the username/password field detection for autofill. r=dolske (423202980)
 - Bug 1144981 - username/password recipe for www.anthem.com. r=mattn (55ce95a2b)
 - Bug 1146065 - Logins captured but not filled on discover.com. r=mattn (4b19c1a01)
 - reorder after strange misspatch (706cfd4ea)
 - Bustage fix for bug 1146065. (4b0a7b5f6)
 - fix some mispatches and also put back some "backports" lost with Bug 1182969 (f61e63baf) (45e41b763)

--

ffmpeg library lite 3.4.9 for NM27/KMG76:
64bit https://o.rthost.win/palemoon/lav64-dll-lite-3.4.9.7z

32bit https://o.rthost.win/palemoon/lav-dll-lite-3.4.9.7z
32bit nosse https://o.rthost.win/palemoon/lav-dll-lite-ia32-3.4.9.7z
32bit noasm https://o.rthost.win/palemoon/lav-dll-lite-noasm-3.4.9.7z

Edited by roytam1
Link to comment
Share on other sites

New regular/weekly KM-Goanna release:
https://o.rthost.win/kmeleon/KM76.4.5-Goanna-20210925.7z

Changelog:

In-tree changes:
- permissions: hackfix CPermission::del() to give out nsIURI instead of const char* as "Bug 1170200 - Part 1: Change the API for nsIPermissionManager::Remove() to accept a URI instead of a string; r=ehsan" is landed in platform tree (036b2296)
- bump KM version for API change (as of NM27 rev 45e41b763) (afd6e427)

Out-of-tree changes:
* update Goanna3 to git 9e8144b17...45e41b763:
- import changes from `dev' branch of rmottola/Arctic-Fox:
 - missing bit of Bug 889158 - Fix arrow function lexical arguments binding (6dd07824a)
 - missing parts of Bug 1135708 - Implement es7 exponentiation operator. (87d1b2a09)
 - Bug 1183400 - Add a by-kind switch to perform constant-folding and eventually eliminate by-arity handling. r=efaust (ce6768404)
 - Bug 1171909 - Check for alloc failure initializing AssemblerBufferWithConstantPools and update error handling ExceptionHandlerBailout r=jandem (fb86a97ec)
 - Bug 1182428 - Consider TypeSet stuff to be GCPointers, since they can contain them, r=jonco (2a7841983)
 - Bug 1184199 - Use a Map instead of a plain object to avoid the "constructor" property, r=jonco" (3af559983)
 - Bug 1184199 - Add an explanation only if there is something to explain, r=jonco (882178fcf)
 - Bug 1182428 - Improve naming, r=jonco (24bcf0679)
 - Bug 1182428 - Recognize more rooted type names, r=jonco (75e9d692e)
 - Bug 963738 - Assume(ptr) is not a hazard; Assume(*ptr) is, r=bhackett (e704d13e6)
 - No bug. Add a --function argument for analyzing a single function by name. (d9e8fc4c0)
 - Bug 1172682 - Comment and refactor analysis code, r=terrence The one change this patch makes in functionality is that it properly computes minimumUse in more cases, in anticipation of it being used for hazards (in addition to its current use for unnecessary roots). (a481f942a)
 - Bug 1182428 - Accept more edge types in edgeCanGC, r=jonco (98413c47f)
 - Bug 1170200 - Part 1: Change the API for nsIPermissionManager::Remove() to accept a URI instead of a string; r=ehsan (bd5ef47ed)
 - Add two missing includes to SharedMemoryBasic_mach.cpp, no bug (dbdeeaca1)
 - bug 1175898 - crash in mozilla::CrossProcessMutex::ShareToProcess(int) r=billm (93cab5830)
 - bug 1204985 - make SharedMemoryBasic_mach build on iOS. r=billm (0f7b668c3)
 - bug 1204985 horrible clobber-avoidance bustage fix r=YOLO CLOSED TREE (d46b9b946)
 - Bug 1205164 - Detect message races in Mach Shmem implementation. r=blassey (2b88cc22b)
 - Bug 1221540: OS X IPC timeout retry with a longer interval. r=milan (66d2006fe) (9e300313d)
- import changes from `dev' branch of rmottola/Arctic-Fox:
 - bug 1167604 - don't hide ProxyAccessible::Parent() with DocAccessibleParent::Parent r=lsocks (32e4cac18)
 - Bug 1173418 - Check if a document is top level in DocAccessibleParent::Destroy so that RemoteDocShutdown isn't erroneously called r=tbsaunde (db280c0a1)
 - bug 1184217 - check consistancy of the document tree before removing accessibles in a hide event r=davidb (91d2f38fb)
 - bug 1162543 - make DocAccessibleChild::IDToAccessible correctly get the document r=davidb (be2eb3eb7)
 - bug 1170595 - switch to sending __delete__ from the parent instead of the child r=davidb, smaug (216a90d34) (3a8ef50a8)
- import change from `dev' branch of rmottola/Arctic-Fox:
 - MOZ_OVERRIDE -> override (624f7c712) (57cc49827)
- import changes from `dev' branch of rmottola/Arctic-Fox:
 - Bug 1182373 - Don't constant-fold in the tag component of a tagged-template, because constant-folding evaluates the tag too early (and moght convert a value into a property reference, causing the wrong |this| to be passed). r=jorendorff (f9e41352b)
 - Bug 1183400 - Constant-fold nullary nodes entirely according to their kind, not their arity. r=efaust (9c6248dd8)
 - Bug 1183400 - Fold typeof nodes by examining their kind. r=efaust (0dc443122)
 - Bug 1183400 - Fold |void <expr>| expressions by kind. r=efaust (d872286ad)
 - Bug 1183400 - Fold delete nodes by kind. r=efaust (a30634947)
 - Bug 1164774 - Remove unused code handling binary-arity PNK_ADD nodes in constant-folding. r=shu (996db297b)
 - Bug 1183400 - Fold !/~/+/- by kind. r=efaust (bf5dd5984)
 - Bug 1183400 - Pass a Parser reference to Fold, not separate handler/options, preparing for another patch that wants to have access to a parser instance. r=efaust (5bbb167be)
 - Bug 1183400 - Fold increment/decrement operations by kind. r=efaust (2728e23a7)
 - Bug 1183400 - Fold ?: expressions by kind. r=efaust (64d6ab853)
 - Bug 1183400 - Fold |if| nodes by kind. r=efaust (0f701a6bb)
 - Bug 1183400 - Fold various simple unary cases by kind. r=efaust (590b24acb)
 - Bug 1183400 - Fold and/or expressions. r=efaust (1f114d2b2)
 - Bug 1183400 - Fold function nodes by kind. r=efaust (0920010c0)
 - Bug 1183400 - Fold binary arithmetic operations by kind, not arity. r=efaust (a3723a1ec)
 - Bug 1183400 - Remove a now-unused variable, following up on the ?:/if-node folding changes. r=trivial, r=orange in a CLOSED TREE (522584dc0)
 - Bug 1183400 - Inline FoldBinaryNumeric into its sole caller and simplify code accordingly. r=efaust (043ccf349)
 - Bug 1183400 - Fold various list nodes not given, nor requiring, special treatment. r=shu (7de28f276)
 - Bug 1183400 - Constant-fold yield/yield*/return by kind and not arity. r=shu (7cae87421)
 - Bug 1183400 - Fold try/catch by kind. r=shu (5df2a97fc)
 - Bug 1183400 - Fold class nodes by kind. r=shu (67fffcc45)
 - Bug 1183400 - Fold element accesses by kind. r=efaust (82538a7a2)
 - Bug 1183400 - Fold addition by kind. r=efaust (68a0c6b49)
 - Bug 1183400 - Fold function calls and tagged templates by kind, not arity. r=efaust (0a368e8f9)
 - Bug 1183400 - Fold various binary nodes with two non-null halves by kind. r=efaust (de48c675b)
 - Bug 1183400 - Fold while and do-while loops by kind, not arity. r=efaust (998bd13cc)
 - Bug 1183400 - Remove dead fold-by-arity code. r=efaust (3b5f719b9)
 - Bug 1183400 - Remove SyntacticContext::Delete, now addressed by modifying how delete nodes are folded. r=efaust (8129f5f7c)
 - Bug 1183400 - Constant-fold switch/default by kind, not arity. r=efaust (e9430d603)
 - Bug 1183400 - Fold for-in, for-of, and for(;;) loops by kind, and classify miscellaneous straightforward kinds for folding. r=efaust (59d09a110)
 - Bug 1183400 - Replace callee-based condition-constant folding with caller-specified condition-constant folding. r=efaust (2090ea750)
 - Bug 1183400 - Remove special |void| handling by making Boolish recognize |void| expressions as falsy, when they're obviously so. r=efaust (d230cbb0d)
 - Bug 1183400 - Remove SyntacticContext::Condition, now handled context-sensitively by callers specifically requesting condition-targeted folding. r=efaust (8be79cd7a)
 - Bug 1183400 - Fold the last few nodes (including PN_NAME nodes) by kind, not arity. r=efaust (4e8e911a4)
 - Bug 1183400 - Remove SyntacticContext completely. r=efaust (745d4df15)
 - Bug 1183400 - Common up some is-effectless testing. r=efaust (ed25885e6)
 - Bug 1135322 - receiveMessage() should convert logins to nsILoginInfo objects. r=MattN (35da0f066)
 - Bug 1140242 - Convert nsILoginInfo to vanilla JS objects before trying to send them in messages. r=dolske (f952e9ab5)
 - Bug 1134846 - Add a module to support per-site password manager recipes. r=dolske (d9d686818)
 - Bug 1135451 - fillForm() cleanup part A: remove unused return type, kill E10S unfriendly fillForm from nsILoginManager, kill passwordmgr-found-form notification, largely a backout of bug 439365. r=MattN (3825eb1dd)
 - Bug 1120129 - Allow per-site recipes to adjust the username/password field detection for autofill. r=dolske (423202980)
 - Bug 1144981 - username/password recipe for www.anthem.com. r=mattn (55ce95a2b)
 - Bug 1146065 - Logins captured but not filled on discover.com. r=mattn (4b19c1a01)
 - reorder after strange misspatch (706cfd4ea)
 - Bustage fix for bug 1146065. (4b0a7b5f6)
 - fix some mispatches and also put back some "backports" lost with Bug 1182969 (f61e63baf) (45e41b763)

* Notice: the changelog above may not always applicable to XULRunner code which K-Meleon uses.

A goanna3 source tree that has kmeleon adaption patch applied is available here: https://github.com/roytam1/palemoon27/tree/kmeleon76

--

ffmpeg library lite 3.4.9 for NM27/KMG76:
32bit https://o.rthost.win/palemoon/lav-dll-lite-3.4.9.7z
32bit nosse https://o.rthost.win/palemoon/lav-dll-lite-ia32-3.4.9.7z
32bit noasm https://o.rthost.win/palemoon/lav-dll-lite-noasm-3.4.9.7z

Edited by roytam1
Link to comment
Share on other sites

On 9/22/2021 at 3:58 PM, roytam1 said:

BTW with EWDK-rs4 VC 19.13, build time is about 10 minutes longer than vc2015u3.

This a newer EWDK than i used, so likely more bloated.  Previously i used only EWDKs compiler and with win8 sdk from vs2013, not allowed by default, for what i patched stuff in the python mozbuild, i shall remember if needed.

10 hours ago, roytam1 said:

- ffvpx: update ffmpeg to 3.4.9 (9ab5fe727)

I did update to 4.3.1 not sure if its better, but there where begin to develop own av1 decoder, as default av1 decoder code buggy under x86 msvc including upstream too.

https://tools.woolyss.com/html5-audio-video-tester/?u=woolyss.com/f/caminandes-1-llama-drama-av1-opus.webm

it stutters on x86 msvc builds, but good on mingw builds

 

Edited by feodor2
link to check av1
Link to comment
Share on other sites

Hi,

With basilisk52-g4.8.win32-git-20210925-f94c0da-uxp-9ab5fe727-xpmod.7z any add-on I try and install is throwing this kind of error:

1632563439888    addons.weblistener    WARN    Exception showing install confirmation dialog: TypeError: Ci.nsISecurityUITelemetry is undefined (resource://gre/components/amWebInstallListener.js:184:1) JS Stack trace: checkAllDownloaded@amWebInstallListener.js:184:1 < Installer@amWebInstallListener.js:81:3 < onWebInstallRequested@amWebInstallListener.js:335:5 < installAddonsFromWebpage@AddonManager.jsm:2206:16 < installAddonsFromWebpage@AddonManager.jsm:3509:5 < buildNextInstall@addonManager.js:107:9 < installAddonsFromWebpage/buildNextInstall/<@addonManager.js:149:9 < safeCall@AddonManager.jsm:186:5 < getInstallForURL_safeCall@AddonManager.jsm:1920:11 < getInstallForURL/<@XPIProvider.jsm:3936:7 < createDownloadInstall/<@XPIProvider.jsm:6628:33 < process@Promise-backend.js:931:23 < walkerLoop@Promise-backend.js:812:7 < scheduleWalkerLoop/<@Promise-backend.js:746:11

...this is by dragging and dropping the xpi into the browser window.
 

Ben.

Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...