{folly,fizz,mvfst,wangle,fbthrift,fb303,edencommon,watchman}: 2024.11.18.00 -> 2024.12.09.00 #362234
Conversation
ofborg
bot
requested review from
kylesferrazza,
abbradar,
pierreis,
risicle and
emilazy
ofborg
bot
added
11.by: package-maintainer
techknowlogick
force-pushed
the
bump-fb-deps
branch
from
08f2077 to
4f9ea14
Compare
techknowlogick
changed the title
techknowlogick
force-pushed
the
bump-fb-deps
branch
from
4f9ea14 to
8a39057
Compare
techknowlogick
force-pushed
the
bump-fb-deps
branch
from
8a39057 to
5caf7ce
Compare
|
Marking as "ready-for-review" as it passes on darwin, now I'm just running it through linux to see what's up, and I'd hopefully like to get some feedback before I need to rebase and build this all again (if time doesn't permit for others that's ok, as this is just a standard update). |
this is from libcxx-19, part of llvm-19.
what we've done for other issues is report the compile error and fix it. it looks like a bad signed char -> chardiff --git a/folly/memory/test/UninitializedMemoryHacksTest.cpp b/folly/memory/test/UninitializedMemoryHacksTest.cpp
index 38e27c3..17424af 100644
--- a/folly/memory/test/UninitializedMemoryHacksTest.cpp
+++ b/folly/memory/test/UninitializedMemoryHacksTest.cpp
@@ -283,7 +283,7 @@ TEST(UninitializedMemoryHacks, simpleStringWChar) {
}
TEST(UninitializedMemoryHacks, simpleStringSChar) {
- testSimple<std::basic_string<signed char>>();
+ testSimple<std::basic_string<char>>();
}
TEST(UninitializedMemoryHacks, simpleVectorChar) {
@@ -307,7 +307,7 @@ TEST(UninitializedMemoryHacks, randomStringWChar) {
}
TEST(UninitializedMemoryHacks, randomStringSChar) {
- testRandom<std::basic_string<signed char>>();
+ testRandom<std::basic_string<char>>();
}
TEST(UninitializedMemoryHacks, randomVectorChar) {
@@ -323,5 +323,5 @@ TEST(UninitializedMemoryHacks, randomVectorInt) {
}
// We are deliberately putting this at the bottom to make sure it can follow use
-FOLLY_DECLARE_STRING_RESIZE_WITHOUT_INIT(signed char)
+//FOLLY_DECLARE_STRING_RESIZE_WITHOUT_INIT(char)
FOLLY_DECLARE_VECTOR_RESIZE_WITHOUT_INIT(int)or could just comment out the two erroneous tests and the macro at the bottom. or change it to a |
techknowlogick
force-pushed
the
bump-fb-deps
branch
3 times, most recently
from
cab403f to
0bab4b5
Compare
|
@paparodeo ah, thanks! That is much easier than what I was doing. I've updated the PR based on your feedback. |
wegank
added
the
12.approvals: 1
|
Edit: I missed the trailing line in the diff and it failed to apply, it's now been added back. All good to go again, ty all <3 |
techknowlogick
force-pushed
the
bump-fb-deps
branch
from
0bab4b5 to
8fdf659
Compare
|
FWIW the test |
|
Hey, I noticed that the following fail on Which is currently being worked on at: Perhaps we should merge the update to these packages to |
|
OK, I just tested the |
|
@doronbehar when you tested wangle against staging-next did you bring over the other libraries too? Fb's codebase is rather intertwined and pretty much require each library to always be on the same tag. I'm not too knowledgeable about the branching rules in the nixpkgs repo (other than "something that causes a large amount of rebuilds go into staging else use master", and also the release branches), so I'll defer to your opinion on if this should go into staging-next. I'm happy to do the work if needed though. |
That's what I imagined, and I didn't do that. I took only the updates to the packages I listed above to be failing on
I'm too not sure what to do here, hence I ask for @vcunat's advice on the idea of merging all of the updates to |
techknowlogick
force-pushed
the
bump-fb-deps
branch
from
8fdf659 to
748bb47
Compare
The thermostat in my place would agree with that 😆
Fair enough, I'll await their advice too. In either case, I've just rebased this on recent commits to staging, including the commits for skipping the timing tests that fail under load, so at least there is one less reason for folly to fail in hydra. |
@doronbehar: all these exact derivations did succeed on Hydra, so your premise seems to be wrong (at least partially). |
wegank
removed
the
12.approvals: 1
|
|
Hmm I wonder how come I observed those issues back then, but never mind - indeed after a Not exactly the ones I listed above (hashes are different), but the development environment I wanted to test that builds for me on I'll leave to you the decision of whether to target this to |
|
You posted .drv hashes first and output hashes now. Of course those differ from each other. |
|
@vcunat The fail on darwin is fixed by this PR (sign chars stuff mentioned above). Should I target |
|
We should probably have merged this one into |
|
@emilazy apologies. Now that I'm maintaining more packages that have a wider blast radius I should've familiarized myself with exactly what the branching protocols are sooner. I've just taken the time to read through some RFCs to understand them. |
|
No mistake on your end – was just re: @vcunat saying “not sure how important these libs are on darwin”, it’s ultimately a trade‐off of rebuilds vs. fixes :) FWIW the |
Folly Diff: facebook/folly@refs/tags/v2024.11.18.00...v2024.12.09.00
Fizz Diff: facebookincubator/fizz@refs/tags/v2024.11.18.00...v2024.12.09.00
Fizz Changelog: https://github.com/facebookincubator/fizz/releases/tag/v2024.12.09.00
MVFST Diff: facebook/mvfst@refs/tags/v2024.11.18.00...v2024.12.09.00
Wangle Diff: facebook/wangle@refs/tags/v2024.11.18.00...v2024.12.09.00
FBThrift Diff: facebook/fbthrift@refs/tags/v2024.11.18.00...v2024.12.09.00
FB303 Diff: facebook/fb303@refs/tags/v2024.11.18.00...v2024.12.09.00
EdenCommon Diff: facebookexperimental/edencommon@refs/tags/v2024.11.18.00...v2024.12.09.00
Watchman Diff: facebook/watchman@refs/tags/v2024.11.18.00...v2024.12.09.00
Things done
nix.conf? (See Nix manual)sandbox = relaxedsandbox = truenix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/)Add a 👍 reaction to pull requests you find important.