Nnethercote: /* Status updates for the week of 2018/12/17 */

2019-01-06T23:30:03Z

Status updates for the week of 2018/12/17

← Older revision		Revision as of 23:30, 6 January 2019
Line 1:		Line 1:
			= Status updates for the week of 2018/12/24 =

			== Nick ==

			Dec 17--21: Four day week due to PTO.

			rustc landings:
			* [https://github.com/rust-lang/rust/pull/56737 `TokenStream` improvements.]

	= Status updates for the week of 2018/12/17 =		= Status updates for the week of 2018/12/17 =

Nnethercote: /* Status updates for the week of 2018/12/10 */

2018-12-16T21:32:54Z

Status updates for the week of 2018/12/10

← Older revision		Revision as of 21:32, 16 December 2018
Line 1:		Line 1:
			= Status updates for the week of 2018/12/17 =

			== Nick ==

			Dec 10--14

			rustc landings:
			* [https://github.com/rust-lang/rust/pull/56269 Use a `SmallVec` within `_match::Matrix`.]
			* [https://github.com/rust-lang/rust/pull/56369 Remove `tokenstream::Delimited`.]
			* [https://github.com/rust-lang/rust/pull/56090 Overhaul `FileSearch` and `SearchPaths`.]
			* [https://github.com/rust-lang/rust/pull/56699 Use a `newtype_index!` within `Symbol`.]

	= Status updates for the week of 2018/12/10 =		= Status updates for the week of 2018/12/10 =

Nnethercote at 19:32, 12 December 2018

2018-12-12T19:32:22Z

← Older revision		Revision as of 19:32, 12 December 2018
Line 6:		Line 6:
	* Debugged bindgen a bit to try out the "packed" idea		* Debugged bindgen a bit to try out the "packed" idea
	* [https://github.com/rust-lang/rust/issues/56730 Regression] on Debian stable due to earlier patch; investigating		* [https://github.com/rust-lang/rust/issues/56730 Regression] on Debian stable due to earlier patch; investigating

	~~= Status updates for the week of 2018/12/03 =~~

	== Nick ==		== Nick ==

Tromey: status

2018-12-12T15:32:29Z

status

← Older revision		Revision as of 15:32, 12 December 2018
Line 1:		Line 1:
			= Status updates for the week of 2018/12/10 =

			== Tom ==

			* While at All Hands, [https://github.com/rust-lang/rust/pull/55010 template parameter patch] landed.
			* Debugged bindgen a bit to try out the "packed" idea
			* [https://github.com/rust-lang/rust/issues/56730 Regression] on Debian stable due to earlier patch; investigating

	= Status updates for the week of 2018/12/03 =		= Status updates for the week of 2018/12/03 =

Nnethercote: /* Status updates for the week of 2018/12/03 */

2018-12-11T02:43:11Z

Status updates for the week of 2018/12/03

← Older revision		Revision as of 02:43, 11 December 2018
Line 1:		Line 1:
			= Status updates for the week of 2018/12/03 =

			== Nick ==

			Dec 03--07

			* Orlando All Hands

	= Status updates for the week of 2018/12/03 =		= Status updates for the week of 2018/12/03 =

Nnethercote: Created page with "= Status updates for the week of 2018/12/03 = == Nick == Nov 26--30 rustc landings: * [https://github.com/servo/rust-smallvec/pull/133 (rust-smallvec) Add a new feature, `m..."

2018-12-11T02:41:29Z

Created page with "= Status updates for the week of 2018/12/03 = == Nick == Nov 26--30 rustc landings: * [https://github.com/servo/rust-smallvec/pull/133 (rust-smallvec) Add a new feature, `m..."

New page

= Status updates for the week of 2018/12/03 =

== Nick ==

Nov 26--30

rustc landings:
* [https://github.com/servo/rust-smallvec/pull/133 (rust-smallvec) Add a new feature, `may_dangle`.].
* [https://github.com/rust-lang/rust/pull/56336 Clean up and streamline the pretty-printer]
* [https://github.com/rust-lang/rust/pull/56268 Reuse the `P` in `InvocationCollector::fold_{,opt_}expr`.]

= Status updates for the week of 2018/11/26 =

== Nick ==

Nov 19--23

Firefox landings:
* {{Bug|1499907|Fix invalid JSON produced by dmd.py.}}

rustc landings:
* [https://github.com/rust-lang/rust/pull/55906 Clean up and streamline snapshot data structures.]

Other:
* More progress on new DMD viewer.

== Gabriele ==

* I triaged November 17th's nightlies, filed {{bug|1508212}} - Crash in java.lang.NullPointerException: at android.view.ViewRootImpl.performDraw(ViewRootImpl.java), {{bug|1508216}} - Crash in java.lang.IllegalArgumentException: at org.mozilla.gecko.sync.Utils.throwIfNull(Utils.java), {{bug|1508221}} - Crash in _chkstk | CreateNotificationBlock and added two signatures to {{bug|1479795}} - Crash in igd10iumd64.dll with Intel driver 23.20.16.4973
* I closed a bunch of old bugs to clean up some of my components: {{bug|1091975}} - Android NDK r10c failure building breakpad: error: 'ElfELFSIZE_Addr' does not name a type, {{bug|1119569}} - We should assert/crash if CGROUPS is not enabled in the kernel, {{bug|1362203}} - Crash in AsyncShutdownTimeout | profile-before-change | Crash reporter: blocking on minidump analysis
* I prepared a patch to figure out what executable we're stuck on in {{bug|1386760}}
* I prepared a patch to block the Facebook Uploader plugin from loading on Mac in {{bug|1502977}}. I regressed this when I refactored the crash annotations
* I wrote a patch to diagnose the timing difference described in {{bug|1505944}} but the results were inconclusive
* I filed and investigated {{bug|1508215}} which is about graphics adapter names showing up in weird ways on Socorro
* I filed {{bug|1508265}} - Re-enable getting minidumps of hung content processes on shutdown, and {{bug|1509298}} - Add tests to ensure that FallbackEncoding picks up locale changes correctly as follow ups to work I did last week and a number of bugs regarding crash reporting issues: {{bug|1509032}} - The pingsender leaves behind zombie processes on Linux, {{bug|1508267}} - Crash reports that have not been submitted are not deleted on Android, {{bug|1509030}} - The X11 widget does not populate the graphics adapter annotations
* I landed a patch to fix {{bug|1500026}} - The FallbackEncoding instance is not listening to the "intl:requested-locales-changed" event

= Status updates for the week of 2018/11/19 =

== Nick ==

Nov 12--16

rustc landings:
* [https://github.com/rust-lang/rust/pull/55525 Make `MatcherPos::stack` a `SmallVec`.]
* [https://github.com/rust-lang/rust/pull/55604 Avoid the Box in `TyCtxt::associated_items`.]
* [https://github.com/rust-lang/rust/pull/55905 Change `Lit::short_name` to `Lit::literal_name`.]
* [https://github.com/rust-lang/rust/pull/55805 Move `static_assert!` into librustc_data_structures].
* [https://github.com/rust-lang-nursery/ena/pull/14 (ena) Remove `OpenSnapshot` and `CommittedSnapshot` markers.]
* [https://github.com/rust-lang/rust/pull/55556 Use `Mmap` to open the rmeta file.]
* [https://github.com/rust-lang/rust/pull/55778 Wrap some query results in `Lrc`.]

Other:
* Good progress on new DMD viewer.

= Status updates for the week of 2018/11/12 =

== Nick ==

Nov 05--09

Four day week due to a public holiday.

rustc landings:
* [https://github.com/rust-lang/rust/pull/55501 Make `process_obligations`' computation of `completed` optional.]
* [https://github.com/rust-lang/rust/pull/55777 Use `Lit` rather than `P<Lit>` in `ast::ExprKind`.]
* [https://github.com/rust-lang/rust/pull/55816 Use `SmallVec` to avoid allocations in `from_decimal_string`.]
* [https://github.com/rust-lang/rust/pull/55745 Convert `outlives_components`' return value to a `SmallVec` outparam.]

Other:
* [https://blog.mozilla.org/nnethercote/2018/11/06/how-to-speed-up-the-rust-compiler-in-2018-nll-edition/ Blogged] about my Rust NLL improvements.

= Status updates for the week of 2018/11/05 =

== Nick ==

Oct 29--Nov 02

rustc landings:
* [https://github.com/rust-lang/rust/pull/55384 Avoid unnecessary allocations in `float_lit` and `integer_lit`.]
* [https://github.com/rust-lang/rust/pull/55574 Use `SmallVec` within `MoveData`.]
* [https://github.com/rust-lang/rust/pull/55558 Tweak `MatcherPos::matches`.]

Other
* Finished first draft of DHAT overhaul. I've been using it on rustc, it's nice!

= Status updates for the week of 2018/10/29 =

== Nick ==

Oct 22--26

rustc landings
* [https://github.com/rust-lang/rust/pull/55383 Use `SmallVec` for the queue in `coerce_unsized`.]
* [https://github.com/rust-lang/rust/pull/55346 Shrink `Statement`.]
* [https://github.com/rust-lang/rust/pull/55167 Add a "cheap" mode for `compute_missing_ctors`.]

Valgrind landings
* [https://bugs.kde.org/show_bug.cgi?id=399301 Use inlined frames in Massif XTree output.]

Other
* Lots more DHAT overhaul progress. (The above rustc improvements were found with it.)

= Status updates for the week of 2018/10/22 =

== Nick ==

Oct 15--19

Firefox landings:
* {{Bug|1499906|Sundry aboutMemory.js code improvements.}}

Other:
* Lots more progress on a new data viewer for DHAT. I started using it in earnest on rustc.

= Status updates for the week of 2018/10/15 =

== Gabriele ==

* I triaged October 6th's nightlies and filed {{bug|1497143}} - Crash in java.lang.IllegalStateException: at android.support.v4.app.FragmentManagerImpl.saveAllState(FragmentManager.java), {{bug|1497145}} - Crash in java.lang.OutOfMemoryError: at android.database.CursorWindow.nativeGetString(Native Method), {{bug|1497148}} - Crash in java.lang.IllegalArgumentException: at org.mozilla.gecko.GeckoProfile.<init>(GeckoProfile.java), {{bug|1497149}} with two distinct signatures - Crash in OOM | large | NS_ABORT_OOM | nsTArray_base<T>::EnsureCapacity<T> | mozilla::dom::HTMLMediaElement::DispatchAsyncEvent
* I updated and landed the fix for {{bug|1390547}}, now multi-line MozCrashReason annotations are properly handled both in Rust and C++
* I wrote a fix for {{bug|1497692}} which lowers the priority of the minidump-analyzer tool after a crash. The goal is to let the user go back to what it was doing ASAP
* I prepared a patch for {{bug|1498609}} which will tell us which of our users are using ECC memory and which are not. I also prepared a data review request since this is new data we didn't collect before
* I've spent some time gathering information on how to implement crash pings for GeckoView in {{bug|1483331}}. I've got a plan almost ready for that
* I've further worked on {{bug|1435683}}, integrating the new prototype service with the XPCOM machinery and finishing the script that generates the list of topics

== Nick ==

Oct 08--12

Firefox landings:
* {{Bug|1497734|Rename nsStaticAtom subclasses.}} A clean-up.
* {{Bug|1460180|Add a comment about RDF atoms.}}

rustc-perf landings:
* [https://github.com/rust-lang-nursery/rustc-perf/pull/293 Clarify the required positions of the `bench_local` options.] A minor docs improvement.

Valgrind landings:
* [https://bugs.kde.org/show_bug.cgi?id=399322 Improve callgrind_annotate output.]

Other:
* [https://blog.mozilla.org/nnethercote/2018/10/12/slimmer-and-simpler-static-atoms/ Blogged] about static atoms.
* Lots of progress on a new data viewer for DHAT.

= Status updates for the week of 2018/10/08 =

== Gabriele ==

* I've triaged September 29th's nightlies and filed {{bug|1495589}} - Crash in js::GCMarker::traverse<T> and {{bug|1495449}} - Crash in mozilla::recordreplay::child::ReportFatalError | mozilla::recordreplay::Stream::CheckInput
* I prepared a patch for {{bug|1390547}} which should fix rust panic reports with multi-line messages
* I uplifted the patch for {{bug|1463048}} which fixes a long-standing shutdown crash caused by asynchronous minidump generation
* I closed {{bug|1485260}} because it wasn't a regression, the change in the bug just made our tools finally see all the symbols present in Linux PGO builds
* I fixed {{bug|1494203}}, an issue with highlighting a page title in Fennec's awesomebar which I had introduced while fixing {{bug|1127855}}
* I uplifted both {{bug|1127855}} and {{bug|1494203}} to beta
* I've finally started working on {{bug|1435683}} having laid out a basic IDL for the new observer service and possibly found a partial solution for dealing with artifact builds

== Nick ==

Oct 01--05

Firefox landings:
* {{Bug|1494175|Streamline nsIWebBrowser{,Chrome}.idl.}} Dead code removal.
* {{Bug|1489047|Remove DOMString from XPIDL.}} A simplification.
* {{Bug|1494244|Remove dead code in and around nsFrameLoader methods.}}
* {{Bug|1451169|Improve static atom handling in various places.}}
* {{Bug|1449883|Improve static atom handling in dom/svg/.}}
* {{Bug|1449787|Make static atom pointers `constexpr`.}} Saves 18KB per process.

rustc-perf landings:
* [https://github.com/rust-lang-nursery/rustc-perf/pull/288 Change a `cp` command]. A step towards making it work on Mac.

= Status updates for the week of 2018/10/01 =

== Gabriele ==

* I've triaged September 22nd's nightlies and filed far too many bugs: {{bug|1493590}} - Crash in mozilla::layers::GPUVideoImage::GetAsSourceSurface, {{bug|1493591}} - Crash in mozilla::dom::WorkerPrivate::~WorkerPrivate, {{bug|1493593}} - Crash in avx::memset32, {{bug|1493607}} - Crash in java.lang.UnsatisfiedLinkError: dlopen failed: library "/data/data/org.mozilla.fennec_aurora/lib/libmozglue.so" not found at java.lang.Runtime.load0(Runtime.java), {{bug|1493611}} Crash in aom_lpf_vertical_4_neon, {{bug|1493630}} Crash in std::panicking::rust_panic_with_hook, {{bug|1493633}} - Crash in mozilla::gfx::PVRGPUChild::SendStartVRService, {{bug|1493635}} - Crash in libhwui.so@0x86de6, I've also added a signature to {{bug|1461267}}
* I've added a signature to {{bug|1127855}} and ended up investigating the bug, I ultimately traced it down to the fact that the lowercase version of a string might have a different number of characters than the original mixed-case - at least in some languages. I wrote a patch to fix it but...
* My fix introduced a new one, {{bug|1494203}}. I've already prepared a fix but it's still undergoing review.
* I've investigated the shutdown crash in {{bug|1488091}} and I have the feeling that it's related to how we write out preferences during shutdown. Unfortunately it's not easy to be certain and it might not even be our fault.
* While working on my observer refactoring I started getting an assertion when starting my nightly build, I ultimately traced it down to the preferences code using floating-point/string conversions in a locale-dependent way. I fixed the issue by making all the conversions locale-independent.
* I finally landed {{bug|1490240}} which removes all remaining warnings in the first-party crash reporting code and quite a few in mozglue too.
* I've reviewed a patch in {{bug|1493227}}, a workaround for a horrible issue with the Oculus Mobile runtime.

== Nick ==

A four day week due to a public holiday.

Firefox landings:
* {{Bug|1494515|Some static atom cleanups.}} Cleanup prep for making static atom pointer <code>constexpr</code>.

rustc landings:
* [https://github.com/rust-lang/rust/pull/54526 Shrink `StatementKind.] Reduced peak memory use by up to 4%.

rustc-perf landings:
* [https://github.com/rust-lang-nursery/rustc-perf/pull/286 Replace some erroneous mentions of `bench_local` with `profile`.] A docs fix.

= Status updates for the week of 2018/09/24 =

== Nick ==

rustc landings:
* [https://github.com/rust-lang/rust/pull/54286 Merge `bitvec.rs` and `indexed_set.rs`]. A big clean-up, and prerequisite for...
* [https://github.com/rust-lang/rust/pull/54318 Use `HybridBitSet` in `SparseBitMatrix`]. Reduced NLL peak memory by 14--45% on four rustc-perf benchmarks, and 96% on one external crate!
* [https://github.com/rust-lang/rust/pull/54211 Split `Liveness::users` into three]. Reduces non-NLL peak memory by 20% on keccak.
* [https://github.com/rust-lang/rust/pull/54420#event-1861676094 Compress `Liveness` data some more]. Reduces instruction counts on NLL benchmarks by up to 16%, and max-rss by up to 38%.
* [https://github.com/rust-lang/rust/pull/54370 Improve handling of type bounds in `bit_set.rs`]. A code clean-up.

== Gabriele ==

* I triaged September 15th's nightlies and filed {{Bug|1491761}} - Crash in js::CloneRegExpObject
* I reviewed {{bug|1388134}} which lazifies reading application directories in the crash manager
* I've finished and landed the patch for {{bug|1463048}} which strips out asynchronous minidump generation as it proved unstable and hard to fix
* I closed four bugs that had become invalid or were fixed by the removal of the disk space wather: {{bug|1328321}}, {{bug|1206074}}, {{bug|876850}} and {{bug|863596}}
* I've landed {{bug|1490240}} which removes all the remaining warnings from the crash reporter code (and some from mozglue too) when building under Windows with clang-cl, unfortunately the patch was backed out for breaking some mozglue code
* I've filed and fixed {{bug|1492815}} which ensures that Fennec and GeckoView do not send the 'StackTraces' field when submitting a crash report. This field is large and redundant as it's being used only for generating crash pings
* I've investigated some intermittent failures in {{bug|1376773}} which tests crash detection in marionette but couldn't fix the issue yet
* I landed a fix for {{bug|1470223}} which will improve the quality of the stack-traces printed out in the logs for try runs
* I've prepared a fix for {{bug|1493365}} which was a gtest failure triggered by {{bug|1470223}}. It was a hard one to figure out because it was indirectly caused by the increase in memory consumption caused by the latter

= Status updates for the week of 2018/09/17 =

== Nick ==

'''Firefox landings'''
* {{Bug|1489744|Fix a bounds violation crash in the prefs parser}}. The crash only started showing up in ESR60, oddly enough.
* {{Bug|1490115|Handle unaccompanied low surrogate pairs in the prefs parser}}. Different to bug 1489744, but found while investigating it.
* {{Bug|1490158|Streamline nsISHistoryListener}}.
* {{Bug|1490524|Streamline nsISHistory}}.
* {{Bug|1490858|Streamline nsISHEntry}}.
* {{Bug|1490828|Handle mPersist in the nsSHEntry copy constructor}}. Unblocked landing of automated UBSan checking.
* {{Bug|1489381|Remove nsISettingsService.idl}}.
* {{Bug|1489010|Remove XPIDL support for `Null` and `Undefined` value-taking properties on `DOMString` parameters and attributes}}.

'''rustc landings'''
* [https://github.com/rust-lang/rust/pull/54213 De-overlap the lifetimes of `flow_inits` and `flow_{un,ever_}inits`]. Reduced NLL peak memory by 27% for keccak and 8% for inflate.
* [https://github.com/rust-lang/rust/pull/54177 Remove bitslice.rs]. A clean-up that will facilitate subsequent perf improvements.

'''Other'''
* Booked Orlando flights.
* Did my first review as a newly-minted rustc reviewer.

== Gabriele ==
* I’ve reviewed the patch for {{Bug|1430041}} which adds support for the socket process to the crash reporting code

== Tom ==
* Pinged nrc about rustup review
* Down the rabbit hole of debugging a failing i586 musl build via travis
* Went to a “rust office hours” to give a gdb lesson; took notes, may write a short tutorial

= Status updates for the week of 2018/09/10 =

== Tom ==
* Another round on [https://github.com/rust-lang/rust/pull/54004 DWARF enum patch]; r+ but now crashing on x86 musl
* Fix [https://sourceware.org/bugzilla/show_bug.cgi?id=23626 gdb crash] in corner case with empty enum; landed
* Figured out why [https://github.com/rust-lang/rust/issues/54126 rustup-installed lldb crashes] on macOS; sent PR

== Gabriele ==
* I’ve ripped out quite a bit of [https://bugzilla.mozilla.org/show_bug.cgi?id=1488401 old unused code]
* I’ve cleared all the compiler warnings in the [https://bugzilla.mozilla.org/show_bug.cgi?id=1490240 crash reporter code]
* I’ve been trying to improve our [https://bugzilla.mozilla.org/show_bug.cgi?id=1470223 stack traces fixup script] but I haven’t figured out how to unit-test it yet
* I’ve taught mach how to automatically install the latest JDK when [https://bugzilla.mozilla.org/show_bug.cgi?id=1444881 bootstrapping Fennec on Gentoo]

Low Level Tools/Status2018Q4 - Revision history

Nnethercote: /* Status updates for the week of 2018/12/17 */

Nnethercote: /* Status updates for the week of 2018/12/10 */

Nnethercote at 19:32, 12 December 2018

Tromey: status

Nnethercote: /* Status updates for the week of 2018/12/03 */

Nnethercote: Created page with "= Status updates for the week of 2018/12/03 = == Nick == Nov 26--30 rustc landings: * [https://github.com/servo/rust-smallvec/pull/133 (rust-smallvec) Add a new feature, `m..."