profile
viewpoint
If you are wondering where the data of this site comes from, please visit https://api.github.com/users/garbageslam/events. GitMemory does not store any data, but only uses NGINX to cache data for a period of time. The idea behind GitMemory is simply to give users a better reading experience.
Chris Beck garbageslam @mobilecoinofficial Colorado Art (in picture) by: https://kevintitzer.com/Sludge-From-NY-NJ

garbageslam/cegui-emscripten 18

A port of the CEGUI samples framework to WebGL / Emscripten (as a proof of concept)

aquileia/wesnoth 1

Free, turn-based strategy game with a high fantasy theme, featuring both single-player, and online/hotseat multiplayer combat.

garbageslam/asio-ssl-mutual-auth 1

Boost asio mutual authentication example.

garbageslam/cegui-mirror-tool 1

shell scripts to drive automatic mirror of bitbucket repo to github

garbageslam/clang 1

clone of clang (has some minor clang format patches)

garbageslam/empoxy 1

A port of libepoxy to emscripten

garbageslam/aHash 0

aHash is a non-cryptographic hashing algorithm that uses the AES hardware instruction

garbageslam/aligned 0

A newtype with alignment of at least `A` bytes

garbageslam/aligned-array 0

A rust crate for aligned values and aligned arrays

PullRequestReviewEvent
PullRequestReviewEvent
PullRequestReviewEvent
PullRequestReviewEvent
PullRequestReviewEvent

pull request commentrust-lang/cargo

Stabilize named profiles

The original RFC indicated that --all-targets without --profile would retain the old behavior where it would use different profiles for different targets. However, the implementation uses a single profile, to avoid confusion and to keep things simple.

As a user, this sounds great -- I had no idea that cargo did not use the same profile for the whole build, and it feels like a problematic behavior. For example in a benchmark there may a bench-specific compiler flags that are passed (and these would be set in the bench profile), but then if those flags don't apply to the dependencies then the profiler may not be able to work properly with all the code that is generated.

ehuss

comment created time in 3 days

PullRequestReviewEvent
PullRequestReviewEvent
PullRequestReviewEvent
PullRequestReviewEvent
PullRequestReviewEvent
PullRequestReviewEvent
PullRequestReviewEvent
PullRequestReviewEvent
PullRequestReviewEvent
PullRequestReviewEvent

delete branch garbageslam/mobilecoin

delete branch : simplify-fog-load-testing

delete time in 5 days

push eventmobilecoinfoundation/mobilecoin

Chris Beck

commit sha b2212db794f80ec1b1ee1f70f235b01012ea5718

simplify fog-load-testing implement, and fix it (#1032) * simplify fog-load-testing implement, and fix it The original goal of this commit was to remove nix and libc from being direct dependencies of fog-load-testing. When I first created this, I knew that fundamentally what I needed to do was make the load testing binary launch the binary under test, and then listen for SIGCHLD to see if it dies. This is important because otherwise the test continue even after that point. We also want to actually stop and reap the child when the test ends, or it is very annoying for the user to run the test -- often if you are changing config params, the server will fail on startup etc. Or if the watcher stuff is wrong, then the server won't make any progress. If the child continues running after the test ends and you have to manually find it and kill it that's just annoying. However, upreving nix recently turned out to be annoying, because of something where they didn't bump MSRV when everyone else did and created wierd constraints on their dependencies. I started looking at alternatives to nix / manually writing a signal handler, and started investigating if the rust std process module will actually do this. Sure enough it turns out that the rust std::process stuff does listen for SIGCHLD and does give you a way to check if the child died, and does give you a way to call "waitpid" effectively and reap the child like you are supposed to. https://github.com/rust-lang/rust/blob/cfff31bc833070a00578bd6178160aeed56f28ba/library/std/src/sys/unix/process/process_unix.rs#L187 https://github.com/rust-lang/rust/blob/cfff31bc833070a00578bd6178160aeed56f28ba/library/std/src/sys/unix/process/process_unix.rs#L594 So in this commit we ripped out all the manual SIGCHLD handler stuff, and we actually made a wrapper over `std::process:Child` that kills the child on drop if it isn't dead yet, which is an improvement over what previously existed, based on my testing. We also learned, in testing, that the fog load testing target is broken, because the test is not actually exercised in CI, and at some point we changed the arguments to the ingest server, around peers etc. Additionally we made it a requirement to activate the ingest server, and load testing wasn't doing that, because it wasn't under test. In this commit we also fixed that stuff. In the future, it would be nice to run the load testing target in CI or something, but we should probably also figure out how to harvest the data, which is a larger project. * fix clippy

view details

push time in 5 days

PR merged mobilecoinfoundation/mobilecoin

Reviewers
simplify fog-load-testing implement, and fix it

The original goal of this commit was to remove nix and libc from being direct dependencies of fog-load-testing. When I first created this, I knew that fundamentally what I needed to do was make the load testing binary launch the binary under test, and then listen for SIGCHLD to see if it dies. This is important because otherwise the test continue even after that point. We also want to actually stop and reap the child when the test ends, or it is very annoying for the user to run the test -- often if you are changing config params, the server will fail on startup etc. Or if the watcher stuff is wrong, then the server won't make any progress. If the child continues running after the test ends and you have to manually find it and kill it that's just annoying.

However, upreving nix recently turned out to be annoying, because of something where they didn't bump MSRV when everyone else did and created wierd constraints on their dependencies.

I started looking at alternatives to nix / manually writing a signal handler, and started investigating if the rust std process module will actually do this.

Sure enough it turns out that the rust std::process stuff does listen for SIGCHLD and does give you a way to check if the child died, and does give you a way to call "waitpid" effectively and reap the child like you are supposed to.

https://github.com/rust-lang/rust/blob/cfff31bc833070a00578bd6178160aeed56f28ba/library/std/src/sys/unix/process/process_unix.rs#L187

https://github.com/rust-lang/rust/blob/cfff31bc833070a00578bd6178160aeed56f28ba/library/std/src/sys/unix/process/process_unix.rs#L594

So in this commit we ripped out all the manual SIGCHLD handler stuff, and we actually made a wrapper over std::process:Child that kills the child on drop if it isn't dead yet, which is an improvement over what previously existed, based on my testing.

We also learned, in testing, that the fog load testing target is broken, because the test is not actually exercised in CI, and at some point we changed the arguments to the ingest server, around peers etc. Additionally we made it a requirement to activate the ingest server, and load testing wasn't doing that, because it wasn't under test.

In this commit we also fixed that stuff.

In the future, it would be nice to run the load testing target in CI or something, but we should probably also figure out how to harvest the data, which is a larger project.

+101 -125

1 comment

5 changed files

garbageslam

pr closed time in 5 days

delete branch garbageslam/mobilecoin

delete branch : remove_unused_dependency

delete time in 5 days

push eventmobilecoinfoundation/mobilecoin

Chris Beck

commit sha 240ef5480838c111ee586eb3bc412f0c494c5bf2

remove unused "time" dependency (#1058)

view details

push time in 5 days

PullRequestReviewEvent

pull request commentmobilecoinfoundation/mobilecoin

Bump rand_chacha from 0.3.0 to 0.3.1

it seems like these are the changes, and they should be fine: https://github.com/rust-random/rand/pull/1135

dependabot[bot]

comment created time in 5 days

push eventgarbageslam/mobilecoin

Chris Beck

commit sha 81dd0c66e03893916baa5fa230c3a2385109ae7f

fix clippy

view details

push time in 5 days

push eventgarbageslam/mobilecoin

Chris Beck

commit sha c9febd3795300a8b09d9d445732b6394a963a13d

remove unused "time" dependency

view details

push time in 5 days

PullRequestReviewEvent
PullRequestReviewEvent