Update jemalloc to 5.0.1 #34
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Looks like CI is red? Can this also avoid a version bump? There seem to be a ton of changes in flight... |
|
@alexcrichton I did the version bump because that's what the last commit that updated the allocator did. I'll revert that. |
|
Did this update bring anything useful? Landing it in rust-lang/rust was abandoned rust-lang/rust#45163 (comment), and now using a new version of this crate in Servo fails to land for possibly the same reason: servo/servo#20640 (comment). |
|
@SimonSapin it uses Are those bugs reported in jemalloc upstream? Also, what's Android's system allocator? |
|
I don’t know if they’re really upstream bugs. Servo (and I think Rust as well?) uses a fairly old Android toolchain, presumably for compat with old Android versions. I haven’t found out for sure, but it’s possible that the The system allocator is not really relevant here. Even if it happens to be jemalloc, it’s probably a different version built with a different build system. |
|
@SimonSapin I've reported that upstream: jemalloc/jemalloc#1175 We'll see what happens. Jemalloc 5.1 will be released soon, so maybe a fix can land on upstream jemalloc for it. I am fine with reverting the jemalloc version here, but if @alexcrichton agrees I'd like to add here the same android build bots that we are using in libc/stdsimd to prevent this from happening again. Upgrading jemallocator shouldn't lead to failures 6 months later in neither Rust nor servo.. |
SimonSapin
referenced
this pull request
in servo/servo
Apr 16, 2018
CC https://github.com/alexcrichton/jemallocator/pull/40, rust-lang/rust#49669 The new version of jemallocator requires a more recent jemalloc https://github.com/alexcrichton/jemallocator/pull/34 which doesn’t build on our current Android toolchain jemalloc/jemalloc#1175. To avoid blocking on figuring that out, duplicate ~70 lines from jemallocator and use the older jemalloc-sys directly.
bors-servo
referenced
this pull request
in servo/servo
Apr 16, 2018
Fork the jemallocator crate, fix for nightly-2018-04-15 CC https://github.com/alexcrichton/jemallocator/pull/40, rust-lang/rust#49669 The new version of jemallocator requires a more recent jemalloc https://github.com/alexcrichton/jemallocator/pull/34 which doesn’t build on our current Android toolchain jemalloc/jemalloc#1175. To avoid blocking on figuring that out, duplicate ~70 lines from jemallocator and use the older jemalloc-sys directly. <!-- Reviewable:start --> --- This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/20641) <!-- Reviewable:end -->
|
I'd be tempted to either patch this locally and/or work around it somehow personally rather than be beholden to ancient versions due to variuos platforms |
|
Steps to reproduce (in this crate, I’ve not managed to get the same build error with jemalloc directly): In a temporary directory: wget https://dl.google.com/android/repository/android-ndk-r12b-linux-x86_64.zip
unzip -q android-ndk-r12b-linux-x86_64.zip
android-ndk-r12b/build/tools/make_standalone_toolchain.py --arch arm --api 19 --install-dir toolchainIn jemallocator: Full output below. Changing Full output: |
This was referenced Apr 16, 2018
|
@SimonSapin so I just started adding android CI to jemallocator and had a dejavu... In fact, before submitting this PR upgrading the jemalloc version, I submitted a PR that did exactly that: add CI for 4 Android targets, amongst many other targets, that would have caught this: https://github.com/alexcrichton/jemallocator/pull/31 In that PR I wrote:
To which @alexcrichton responded:
So the work for expanding the CI of jemallocator was already done. The problem is that nobody did the job of updating jemalloc in rustc, catching failures, and fixing things as necessary. @sfackler tried, but IMO it is much easier to catch and debug failures here in isolation than when running the full rust-lang/rust testsuite. I almost think that if the tests do not pass here, it is actually pointless to try to get the rust-lang/rust tests to pass on all relevant targets. Should I reopen that PR ? |
|
FWIW that PR uses |
Moggers
referenced
this pull request
in Moggers/servo
Jun 13, 2018
CC https://github.com/alexcrichton/jemallocator/pull/40, rust-lang/rust#49669 The new version of jemallocator requires a more recent jemalloc https://github.com/alexcrichton/jemallocator/pull/34 which doesn’t build on our current Android toolchain jemalloc/jemalloc#1175. To avoid blocking on figuring that out, duplicate ~70 lines from jemallocator and use the older jemalloc-sys directly.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Updates jemalloc to rust-lang/jemalloc branch: rust-2017-10-09