[WIP] deps: update V8 to 8.0 #32095
[WIP] deps: update V8 to 8.0 #32095
Conversation
Major V8 updates are usually API/ABI incompatible with previous versions. This commit adapts NODE_MODULE_VERSION for V8 8.0. Refs: https://github.com/nodejs/CTC/blob/master/meetings/2016-09-28.md
Original commit message:
[testrunner] delete ancient junit compatible format support
Testrunner has ancient support for JUnit compatible XML output.
This CL removes this old feature.
[email protected],[email protected],[email protected]
CC=[email protected]
Bug: v8:8728
Change-Id: I7e1beb011dbaec3aa1a27398a5c52abdd778eaf0
Reviewed-on: https://chromium-review.googlesource.com/c/1430065
Reviewed-by: Jakob Gruber <[email protected]>
Reviewed-by: Michael Starzinger <[email protected]>
Commit-Queue: Tamer Tas <[email protected]>
Cr-Commit-Position: refs/heads/master@{#59045}
Refs: v8/v8@bd019bd
PR-URL: nodejs#26685
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Michaël Zasso <[email protected]>
Reviewed-By: Refael Ackermann <[email protected]>
PR-URL: nodejs#26685 Reviewed-By: Anna Henningsen <[email protected]> Reviewed-By: Michaël Zasso <[email protected]> Reviewed-By: Refael Ackermann <[email protected]>
Patch V8 (compiler/js-heap-broker.cc) to remove the use of an optional property, which is a fairly new C++ feature, since that requires a newer XCode version than the minimum requirement in BUILDING.md and thus breaks CI. PR-URL: nodejs#29694 Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: Anna Henningsen <[email protected]> Reviewed-By: Gus Caplan <[email protected]> Reviewed-By: Jiawen Geng <[email protected]> Reviewed-By: Michaël Zasso <[email protected]> Reviewed-By: Tobias Nießen <[email protected]> Reviewed-By: Ujjwal Sharma <[email protected]>
Co-authored-by: Ujjwal Sharma <[email protected]>
These methods will be removed in V8 8.1, hence we need to stop overriding them.
This reverts commit 5981fb7.
The following metadata has changed in V8: - v8dbg_class_ConsString__first_offset__int - Use v8dbg_class_ConsString__first__String instead. - Refs: v8/v8@b38dfaf - v8dbg_class_ConsString__second_offset__int - Use v8dbg_class_ConsString__second__String instead. - Refs: v8/v8@b38dfaf - v8dbg_class_Map__constructor_or_backpointer__Object - This constant has not changed, but the postmortem metadata generation script required an update. - v8dbg_class_SlicedString__offset_offset__int - Use v8dbg_class_SlicedString__offset__SMI instead. - Refs: v8/v8@b38dfaf - v8dbg_class_ThinString__actual_offset__int - Use v8dbg_class_ThinString__actual__String instead. - Refs: v8/v8@b38dfaf
This commit updates v8abbr.h to use the updated metadata.
Bump minimum version of ICU needed to build node to 65. Refs: v8/v8@74bf96e
This commit replaces Symbol::Name() with Symbol::Description(). Fixes: nodejs#30916
V8 is about to increase the max TypedArray length to 2**32-1, which Node inherits as Buffer.kMaxLength. Some tests relied on values greater than the previous max length (2**31-1) to throw errors; this updates those tests for the new max length.
Serialization support of WasmModuleObjects will be removed in https://crrev.com/c/2013110, thus we need to stop testing for that.
The {SetExpectInlineWasm} method is deprecated and non-functional since
V8 v8.1.
Thus node should stop calling it, so that it can be fully removed in a
future v8 version.
This removes uses of the "IsWebAssemblyCompiledModule" method, which is deprecated in V8 v8.1 and will be removed in v8.2. We could replace it by "IsWasmModuleObject", but since it's unused in node anyway, I just remove the definition.
nodejs-github-bot
added
build
|
@targos not sure if you was working on this already. If you were, I'm happy to close this PR. I cherry-picked commits from There are some commits cherry-picked from Do we have any way to ensure previously backported V8 commits are preserved when upgrading? I noticed #31957 was overridden, and I assume other backports were as well. |
Original commit message:
Fix MSVC component build
Add the necessary V8_EXPORT_PRIVATE attributes and a few other minor
changes to make building DLLs with MSVC happy. (Note: Debug builds still
seem to be failing in Torque, but this fixes Release builds).
Bug: v8:8791
Change-Id: Ia4d5372fd1cb961e6268a2b5c089bcd17822f1e5
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1996157
Reviewed-by: Jakob Kummerow <[email protected]>
Reviewed-by: Ulan Degenbaev <[email protected]>
Reviewed-by: Tobias Tebbi <[email protected]>
Commit-Queue: Jakob Kummerow <[email protected]>
Cr-Commit-Position: refs/heads/master@{#65822}
Refs: v8/v8@e8e324a
|
@mmarchini TBH, I would like to skip 8.0 and upgrade directly to 8.1, but if you want to merge 8.0 first, I have no objections. This is how I usually prepare an upgrade:
git checkout master
git pull upstream master
git checkout -b v8-8.0
git node v8 major --branch=8.0-lkgr
Feel free to ask if you have any question. |
|
Thank you for sharing! Should we document the steps? I can send a PR.
I was thinking 8.0 could be backportable to v13, but with v14 one month away it might not be worth the effort (especially with that Windows build failure). |
|
The windows build failure looks a lot like we deal in nodejs/node-v8#126 (comment). cc @targos
Any known problem for v8 8.1 backport to v13 ? I am inclined upgrade to v8 8.1 directly too. |
Unfortunately we didn't fix it, it was fixed upstream and we would need to dig up the patch ourselves :/
I haven't checked, but does it make sense to backport to v13 at this point? v14 is coming out next month. |
|
@mmarchini The symbol seems related to zlib, Can you try import this by @targos |
|
Ok, so I opened #32116 to upgrade directly to 8.1. We might still be able to backport to v13, I'm not sure (someone would need to work on backwards-compatible patches, but that would happen for 8.0 as well). I'll close this since that pull request is in way better shape (commits were properly squashes, reworded, etc.). |
Chrome release date: Feb 04th, 2020.
There's one outstanding issue on 8.0 (https://github.com/nodejs/node-v8/issues/120), but it might not be worth blocking the upgrade because of that (we're unlikely to see that issue outside very specific, synthetic test cases).
Also, build might be broken on Windows. I'll look into that before marking as ready for review.