Spelling#13166
Conversation
jsoref
left a comment
There was a problem hiding this comment.
Note that this PR was not generated using AI.
The tooling is my own (check-spelling). I do occasionally use Google Suggest when I'm too lazy to try to type the correct spelling of a word (I'm as bad of a typist as everyone else).
I'll rewrite the PR minus the one item I've flagged
| command: make -i benchmark-platforms BENCHMARKS_SUITE=fio BENCHMARKS_TARGETS=test/benchmarks/fs:fio_test BENCHMARKS_FILTER='BenchmarkFioRead' BENCHMARKS_OPTIONS=--test.benchtime=1000x | ||
| - <<: *benchmarks | ||
| label: ":floppy_disk: FIO benchmarks (randwrte :empty_nest:)" | ||
| label: ":floppy_disk: FIO benchmarks (randwrite :empty_nest:)" |
There was a problem hiding this comment.
This might break some tagging...
| - <<: *benchmarks | ||
| label: ":tensorflow: TensorFlow benchmarks" | ||
| command: make -i benchmark-platforms BENCHMARKS_SUITE=tensorflow BENCHMARKS_TARGETS=test/benchmarks/ml:tensorflow_test BENCHMARKS_FILTER=BenchmarkTensorflowDashboard | ||
| command: make -i benchmark-platforms BENCHMARKS_SUITE=tensorflow BENCHMARKS_TARGETS=test/benchmarks/ml:tensorflow_test BENCHMARKS_FILTER=BenchmarkTensorFlowDashboard |
There was a problem hiding this comment.
This might break some filtering...
| needed. The [format of this file is | ||
| documented](https://clang.llvm.org/docs/JSONCompilationDatabase.html), but for the |
There was a problem hiding this comment.
The relevant rule is "Do not use (click) here links", for more information, see:
- https://www.w3.org/QA/Tips/noClickHere
- https://webaim.org/techniques/hypertext/link_text
- https://granicus.com/blog/why-click-here-links-are-bad/
- https://heyoka.medium.com/dont-use-click-here-f32f445d1021
I can split this to a distinct PR if requested.
| performance-sensitive code. PyTorch contains one non-test use of | ||
| `cudaMallocManaged()`[^cite-pytorch-uvm], but it is not immediately clear what | ||
| circumstances cause the containing function to be invoked. Tensorflow does not | ||
| circumstances cause the containing function to be invoked. TensorFlow does not |
| RUN curl -fsSL https://download.docker.com/linux/ubuntu/gpg | \ | ||
| gpg --dearmor --yes --output /usr/share/keyrings/docker.gpg && \ | ||
| echo "deb [signed-by=/usr/share/keyrings/docker.gpg] https://download.docker.com/linux/ubuntu" \ | ||
| "$(lsb_release -cs)" \ | ||
| "stable" > /etc/apt/sources.list.d/docker.list && \ | ||
| apt-get update && \ | ||
| apt-get -y install docker-ce-cli |
There was a problem hiding this comment.
This addresses a deprecation warning. It's the correct thing to do (code tested locally against ubuntu:22.04 and ubuntu:24.04 as well as in CI).
| @@ -626,7 +626,7 @@ func testDockerRun(ctx context.Context, t *testing.T, d *dockerutil.Container, o | |||
| } | |||
| cmd = append(cmd, testAlpineImage, "sh", "-c", "apk add curl && apk info -d curl") | |||
|
|
|||
| expectedOutput := "URL retrival utility and library" | |||
| expectedOutput := "URL retrieval utility and library" | |||
There was a problem hiding this comment.
You'll eventually need this change whether or not you take this PR as I changed the upstream:
docker run -it alpine:edge sh -c "apk add curl >/dev/null 2>/dev/null && apk info -d curl"
curl-8.20.0-r1 description:
URL retrieval utility and libraryI'm not entirely sure about the release cycle, it might only apply to the version after 3.22. so it might need to be merged later.
| @@ -1138,7 +1138,7 @@ func payload(l Layer) (buffer.Buffer, error) { | |||
| } | |||
|
|
|||
| // layerChecksum calculates the checksum of the Layer header, including the | |||
| // peusdeochecksum of the layer before it and all the bytes after it. | |||
| // PseudoHeaderChecksum of the layer before it and all the bytes after it. | |||
There was a problem hiding this comment.
This had a lot of damage, so I picked the raw term...
| @@ -41,7 +41,7 @@ namespace gvisor { | |||
| namespace testing { | |||
| namespace { | |||
|
|
|||
| constexpr char kSIGALRMToMainThread[] = "--itimer_sigarlm_to_main_thread"; | |||
| constexpr char kSIGALRMToMainThread[] = "--itimer_sigalrm_to_main_thread"; | |||
There was a problem hiding this comment.
notable. I can't find another side for this...
| @@ -27,7 +27,7 @@ func interfaceFunctionCrossPkg() { | |||
| test1.InterfaceFunction(i) | |||
| } | |||
|
|
|||
| // +checkesacape:all | |||
| // +checkescape:all | |||
There was a problem hiding this comment.
This occurred a couple of times. I'm assuming it's a typo. If it's intentionally wrong, I'd encourage something like // +checkescape_:all to hint that you really don't want checkescape...
| @@ -637,7 +637,7 @@ type callCommon interface { | |||
| Value() *ssa.Call | |||
| } | |||
|
|
|||
| // checkInstruction checks the legality the single instruction based on the | |||
| // checkInstruction checks the legality of the single instruction based on the | |||
There was a problem hiding this comment.
This isn't precisely a typo (there are a handful of other corrections that aren't), the heuristic looks for ARTICLE WORD ARTICLE and it's right often enough.
Signed-off-by: Josh Soref <[email protected]>
|
Please squash all your commits. Copybara does not have squash and merge support and we don't want to pollute the master branch. |
* basic level of memory safety through the use of a garbage collector (GC) * code implementing bufferv2 * comprehensive list of all supported CPU features * file system isolation it implements * format of this file is documented * full description of the header * full documentation * full example * full set of allowed syscalls * installation guide * more discussion * pkg/sentry/seccheck/points/common.proto * schema for all syscall trace points Signed-off-by: Josh Soref <[email protected]>
* ; as a result, * ; otherwise * ; otherwise, * 32-bit * a * accept * accommodate * accommodated * acquired * active * added * added to a * adding * address * addresses * again * all * allocated * allocating * allowed * altogether * an * and * and whose * announcements * another * any * at * at least * at most * atomic * atomically * attribute * attributes * balancing * bandwidth * bazelbuild * because * benchmark * big-endian * bit * bound * bound path * calculated * cannot * capabilities * case-insensitive * cgroup * change * checkescape * checks * checksummer * children's * chmoding * cleanup * cluster * coarse-grained * coerce * command * communication * communications * compatibility * concatenates * concurrently * congestion * connect * connection * constant * container * control * cookies * correctness * could * cross * deallocation * defaults * defines * delayed * dellink * dependent * depends * description * destroy * destroyed * detection * determining * directly invoke * directory * disallows * doesn't * dropped * dupack threshold * each * emitter * empty * emulation * encoder * endianness * endpoint * environment * ephemeral * equals * equivalent * errors * eventually * exceeds * exec * execve * exhausted * exists * explicit * explicitly * failure * fall back * feature * feature set * finite-buckets * flexibility * flood * following * for * for container * forcibly * foreground * forward * forwarding * fragments * fstat * futex * gathers * gc'd * get the addresses to which the socket is bound * github * granularity * handlers * have * hold * id * implemented * implements * increasing * incremented * indeterminately * indicates * inet * inherit * initial * initiate * inotify * input * instantiate * instantiated * instruction * intended * interfaces * internal * interpreter * interruptible * into * intrinsics * invalid * is behind * is equal to * is not supposed to * is still * isn't * its * javascript * less * limited * link * link-local * listender * little-endian * lookups * lowercase * maintenance * marshal * marshallable * marshalled * minimum * mismatch * mistakenly * mmapped * mnemonic * mountpoint * msg * multiple * must * name * neighbors * neither * neither checks nor adjusts * neither deprecate * netstack * new packet * nonexistence * nonexistent * not * occasionally * of * of the * of time * on all * on getting * on-fault * only * opening * operations * option * original * otherwise, * out-of-order * output * overridden * overrunning * overwhelmingly * packet * packets * parameters * params * passing-by * pointers * precede * precedence * precondition * preemption * preference * prefix * prematurely * prevention * process * processed * processing * profiling * protocol * provides * pseudoheader * pseudoheaderchecksum * queueing * randwrite * reachability * reachable * reason * reassembly * receive * received * recomputation * reconciled * red hat * regeneration * registers * rehashes * release * remained * request * response * responsible * results * retransmission * retransmitted * retrieves * rfc * routing * running * sanity * scenario * seccomp * section * semaphore * semaphores * sentry * sequence number * serialization * set up * short while * shuts down * sigalrm * sigalrms * solicitation * specified * ss threshold * stack * stateified * stateify * stdout * stopped * struct * stubbed * superfluous * supported * synopsis * syscall * system * tensorflow * than * that * that should contain the shim * the * the cgroups to which the container is initialised * the function receiver itself in the array * the side of a connection to which an endpoint belongs * then * timed out * times out * timestamp * tl;dr * to * to use * to which endpoint queue a particular segment belongs * to which the endpoint is connected * to which the socket is bound * toctou * topmost * transfer * transfer rate * two * unassign * uncommitted * unexpected * uninitialized * unknown * unmarshal * unnecessary * unprivileged * unrecognized * unsolicited * unspecified * unsupported * untrusted * upfront * uppercase * urpc * users * validate * vectorized * verify * vfio userspace * virtual * w.r.t. * when * whether or not * which the address is associated * without * wordpress * workaround * world * writable Signed-off-by: Josh Soref <[email protected]>
|
This is everything minus 4d7b2df as that should probably wait until it migrates from edge to the main form. Are you ok with including these three changes in a single commit? Or should I split them into distinct PRs? |
We're starting to use gVisor and I ran across a typo somewhere... I can't guarantee this is comprehensive, but it's a pretty good pass.
I'll annotate some of these changes...