Skip to content

chore: merge main into version7 - #5920

Merged
jamescrosswell merged 19 commits into
version7from
ai/v7-release-merge-c71d07
Nov 12, 2026
Merged

jamescrosswell merged 19 commits into
version7from
ai/v7-release-merge-c71d07

Conversation

@jamescrosswell

Copy link
Copy Markdown
Collaborator

Caution

Don't merge this with the PR button. The repo only allows squash merges, which would flatten this into a single-parent commit and lose the merge. This PR exists to run CI. Once it's green, version7 gets fast-forwarded to this branch's merge commit by a direct push.

Merges main (up to and including the 6.13.0 release) into version7 with a real merge commit, as part of the 7.0.0 release in #5897.

Conflict resolutions:

  • CHANGELOG.md: git auto-merged it but put 6.13.0 in the middle of the rc.3 section. It's rebuilt with the 7.0.0-rc.* sections above 6.13.0. Those get consolidated into Unreleased just before the release.
  • API snapshots: added main's DataCollection API and dropped ApplyDefaultTags, which version7 removed. DotNet11_0 (version7 only) was regenerated. Net4_8 can't regenerate on macOS, so it was resolved by hand; CI is the check for that one.
  • SentryLoggingOptionsSetupTests.cs: kept version7's test. Since feat(mel)!: the Microsoft.Extensions.Logging integration no longer initializes the SDK #5595, SentryLoggingOptions no longer derives from SentryOptions, so main's DataCollection assertions don't apply there. BindableDataCollectionOptionsTests covers that binding.
  • Directory.Build.props: kept 7.0.0-rc.3, and the dev VersionSuffix that craft cleared when it cut 6.13.0.
  • SystemWebRequestEventProcessor.cs: conflicting using lines only.

#skip-changelog

🤖 Generated with Claude Code

jamescrosswell and others added 19 commits November 6, 2026 11:28
…ltPii (#5853)

* feat: add experimental DataCollection options, successor to SendDefaultPii

Adds the option shape from the DataCollection spec as a no-op: nothing reads
it yet. The resolver and SendDefaultPii bridge follow in #5426.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* feat: bind DataCollection options from configuration

Mirrors DataCollectionOptions in BindableSentryOptions so the option can be
set from appsettings.json, following the nested-bindable pattern used by the
platform options. KeyValueCollectionBehavior has no public constructor, so it
binds through a Mode/Terms surrogate.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* ref: make SentryOptions.DataCollection nullable instead of exposing Enable()

Assigning the option is now the opt-in signal, matching sentry-go, sentry-php,
sentry-python and sentry-javascript. Removes Enable() and the internal
bookkeeping it needed to distinguish an unset option from a default one.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* fix: KeyValueCollectionBehavior.Terms is never null

An explicit null for a params array is legal, so DenyList(null) and
AllowList(null) produced an instance whose Terms was null, which would throw
in any consumer that iterates it. Coalesce in the constructor so both
factories are covered.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* docs: note that key-value categories bind from Mode/Terms, not a scalar

The bool shorthand code has for cookies, headers and query params has no
configuration equivalent, so a scalar is dropped by the binder. Document the
supported shape on the public type and lock the behavior with a test.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* fix: make DataCollection bind correctly on all configuration paths

Swaps the IncomingResponse/OutgoingResponse doc comments, which described the
wrong direction. Gives KeyValueCollectionBehavior a public constructor and
settable properties so the reflection binder can construct it: on
netstandard2.0 the options are bound directly rather than through
BindableSentryOptions, and any key-value category threw there. Binds
HttpBodies from an array so the spec's empty-array opt-out works instead of
silently collecting every body. Adds coverage for the four ApplyTo lines no
test reached.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* fix: document that the HttpBodies empty-array opt-out needs ["None"] in config

A JSON empty array emits a valueless key with no children, which the binder
surfaces as null, so "[]" cannot be told apart from the option being absent
and did not disable body collection. The previous test asserted against a
hand-built list rather than the binder, so it did not catch this.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* fix: null element in HttpBodies configuration crashed startup

The source-generated binder keeps null array elements, so a null entry in
HttpBodies reached Split and threw. Enum.TryParse already handles comma lists,
spacing and casing for a flags enum and rejects null, so the manual splitting
is gone. Sentry.Tests binds with the reflection binder, which drops nulls, so
the crash only shows up through SentryLoggingOptionsSetup, which now covers
DataCollection keys.

Also filters null and empty terms, which would otherwise match every key,
marks HttpBodyTypes and KeyValueCollectionMode experimental like the rest of
the new surface, and drops tests that only fail when a neighbour already does.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* test: expect the net48 HttpBodies binding divergence

On .NET Framework the logging setup binds onto the options directly rather
than through BindableSentryOptions, so indexed HttpBodies keys never reach
ParseHttpBodies and the category binds as None instead of the requested flags.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* fix: HttpBodies in configuration crashed .NET Framework startup

The logging setup bound configuration straight onto the options on
netstandard2.0, so an indexed key under HttpBodies made the reflection binder
try to construct Nullable<HttpBodyTypes> and throw. It now binds through
BindableSentryLoggingOptions on every target, as net6.0 and later already did,
which also removes the divergence where .NET Framework read HttpBodies as a
different value to .NET 8 and later.

The bindable parity test for SentryLoggingOptions has no skipped properties,
so the surrogate covers everything the direct bind did.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* test: restore the DataCollectionOptions tests an over-broad deletion removed

The script that dropped the low-value tests searched from the first [Fact] in
the file, so each deletion also took every test above it and left two empty
classes. Nothing covered the Terms filtering, the fresh Off instance or the
bool conversion; all three now fail under mutation again.

Drops BindableDataCollectionReflectionTests as redundant: Sentry.Tests does
not run the binding generator, so DataCollectionConfigurationTests already
exercises the surrogate with the reflection binder.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* ref: drop the DataCollection categories nothing in .NET reads yet

Mcp, Queues, FilePaths, StackFrameVariables and FrameContextLines had no
consumer anywhere in src/: there is no MCP instrumentation, nothing captures
queued job arguments, there is no file-system instrumentation, and nothing
populates SentryStackFrame.Vars, PreContext, PostContext or ContextLine.
The spec is still a candidate and has renamed or retyped options since June,
so a public property for something we do not instrument locks in today's
shape, while adding one later never breaks anyone.

DatabaseQueryData and GenAI stay - see #5433 and #5434.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* ref: opt into DataCollection through a callback instead of assignment

`SentryOptions.DataCollection` was a nullable property whose presence was the
opt-in signal, which made the natural way to write every category throw. Eleven
properties were nullable, so `options.DataCollection.UserInfo = false` and
`dataCollection.HttpHeaders.Request.Mode = ...` both NRE'd, and assigning a new
instance inside `UseSentry` silently discarded anything bound from
appsettings.json, since configuration binding runs first.

Replace it with `ConfigureDataCollection(Action<DataCollectionOptions>)`. Calling
it is the opt-in, so nothing is inferred from which fields were set; the options
it hands over start at the spec defaults and are non-null all the way down, so
every category can be set or mutated in place. Repeated calls build on the same
instance, and so does configuration, which still binds through the surrogate.

Constructors become internal: the only way to reach the options is the callback,
so a `new DataCollectionOptions()` that silently did nothing is no longer
expressible. This reduces the public surface while the API is experimental.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* ref: make KeyValueCollectionBehavior immutable

Terms only mean something next to their Mode, so in-place edits let code and
configuration combine into a behavior neither side wrote: config supplying
Terms while code sets Mode = AllowList yields AllowList[those terms], and code
appending a deny term to a config-set allow list widens it instead.

Dropping the public constructor and the Mode/Terms setters leaves Off(),
DenyList() and AllowList() as the only ways to build one, matching the spec's
Java example and its requirement that the type make mixing deny and allow terms
impossible. Each category is now replaced whole, from either side. Config keeps
treating Terms without Mode as a deny list, which is the spec default.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* docs: say the DataCollection options are not consumed yet, and address review nits

The XML docs described the destination in the present tense: that calling
ConfigureDataCollection replaces SendDefaultPii, which is then ignored. Nothing
reads DataCollection yet, so SendDefaultPii still governs collection entirely
and restricting a category has no effect. Both facts are now stated, and the
replacement is in the future tense, conditional on the integrations migrating.

Also from review: cover a key-value category in the config-merge test, so the
suite notices if ToBehavior stops returning null for an untouched surrogate and
starts resetting categories config never mentioned; drop the per-category
independence test, which only meant something while behaviors were mutable;
correct a stale StackFrameVariables mention; remove a stray blank line.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
…5 updates (#5887)

Bumps the github-actions group with 5 updates in the / directory:

| Package | From | To |
| --- | --- | --- |
| [getsentry/craft/.github/workflows/changelog-preview.yml](https://github.com/getsentry/craft) | `3.3.0` | `3.9.0` |
| [github/codeql-action/init](https://github.com/github/codeql-action) | `4.39.1` | `4.39.2` |
| [github/codeql-action/analyze](https://github.com/github/codeql-action) | `4.39.1` | `4.39.2` |
| [gradle/actions/setup-gradle](https://github.com/gradle/actions) | `6.4.0` | `6.4.1` |
| [actions/setup-dotnet](https://github.com/actions/setup-dotnet) | `6.0.0` | `6.1.0` |

Bumps the github-actions group with 1 update in the /.github/actions/environment directory: [actions/setup-dotnet](https://github.com/actions/setup-dotnet).


Updates `getsentry/craft/.github/workflows/changelog-preview.yml` from 3.3.0 to 3.9.0
- [Release notes](https://github.com/getsentry/craft/releases)
- [Changelog](https://github.com/getsentry/craft/blob/master/CHANGELOG.md)
- [Commits](getsentry/craft@4b763f5...2e29b08)

Updates `github/codeql-action/init` from 4.39.1 to 4.39.2
- [Release notes](https://github.com/github/codeql-action/releases)
- [Changelog](https://github.com/github/codeql-action/blob/main/CHANGELOG.md)
- [Commits](github/codeql-action@9b32540...1fc58d0)

Updates `github/codeql-action/analyze` from 4.39.1 to 4.39.2
- [Release notes](https://github.com/github/codeql-action/releases)
- [Changelog](https://github.com/github/codeql-action/blob/main/CHANGELOG.md)
- [Commits](github/codeql-action@9b32540...1fc58d0)

Updates `gradle/actions/setup-gradle` from 6.4.0 to 6.4.1
- [Release notes](https://github.com/gradle/actions/releases)
- [Commits](gradle/actions@3f5f9ad...8da46b5)

Updates `actions/setup-dotnet` from 6.0.0 to 6.1.0
- [Release notes](https://github.com/actions/setup-dotnet/releases)
- [Commits](actions/setup-dotnet@a98b568...0f4a354)

Updates `actions/setup-dotnet` from 6.0.0 to 6.1.0
- [Release notes](https://github.com/actions/setup-dotnet/releases)
- [Commits](actions/setup-dotnet@a98b568...0f4a354)

---
updated-dependencies:
- dependency-name: getsentry/craft/.github/workflows/changelog-preview.yml
  dependency-version: 3.9.0
  dependency-type: direct:production
  update-type: version-update:semver-minor
  dependency-group: github-actions
- dependency-name: github/codeql-action/init
  dependency-version: 4.39.2
  dependency-type: direct:production
  update-type: version-update:semver-patch
  dependency-group: github-actions
- dependency-name: github/codeql-action/analyze
  dependency-version: 4.39.2
  dependency-type: direct:production
  update-type: version-update:semver-patch
  dependency-group: github-actions
- dependency-name: gradle/actions/setup-gradle
  dependency-version: 6.4.1
  dependency-type: direct:production
  update-type: version-update:semver-patch
  dependency-group: github-actions
- dependency-name: actions/setup-dotnet
  dependency-version: 6.1.0
  dependency-type: direct:production
  update-type: version-update:semver-minor
  dependency-group: github-actions
- dependency-name: actions/setup-dotnet
  dependency-version: 6.1.0
  dependency-type: direct:production
  update-type: version-update:semver-minor
  dependency-group: github-actions
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Skip hunks an earlier clean Warden run on the PR already reviewed for: find-bugs, code-review, gha-security-review.

Co-authored-by: sentry-junior[bot] <264270552+sentry-junior[bot]@users.noreply.github.com>
Co-authored-by: Jeffrey Hung <jeffrey.hung@sentry.io>
…rently (#5878)

CachingTransport.ProcessCacheAsync waited on _processingSignal inside its
try block, so a cancelled wait still released the signal in the finally.
That freed the signal while another pass held it, letting the next flush
process the cache at the same time.

Acquire the signal before the try, so only the holder releases it.

Fixes #5875

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
#5863)

* fix: pause and resume sessions when mobile apps move to the background

Closes #5842

Co-Authored-By: Claude <noreply@anthropic.com>

* fix: address review of session lifecycle tracking

Delay the Android background-launch pause past a cold start, resume on
DidBecomeActive on iOS, use the Lock type, and test the platform trackers.

Co-Authored-By: Claude <noreply@anthropic.com>

* test: isolate iOS lifecycle tracker tests on a private notification center

Co-Authored-By: Claude <noreply@anthropic.com>

* fix: start mobile sessions when the app comes to the foreground

Background launches (a push, a broadcast, a job, a background fetch) no
longer start a session, matching sentry-java and sentry-cocoa. This
replaces the Android 5-second and iOS at-init background checks.

The previous run's session is still recovered at init, split out of
Hub.StartSession, and its file is deleted once recovered. Otherwise a
background launch that never starts a session would send it again on
the next launch, as crashed if that background run crashed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: start the Android session at init when ProcessLifecycleOwner is not attached

Apps that remove androidx.startup's InitializationProvider never attach
ProcessLifecycleOwner, so ON_START never fires and no session would start.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test: drop session start test already covered by HubTests

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix: start the Android session at init only when the process is in the foreground

Matches SentryAndroid.init in sentry-java: a foreground launch has a session
from init, so a crash before ON_START counts as a crashed session, and a
background launch never starts one, even when ProcessLifecycleOwner is not
attached.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: James Crosswell <jamescrosswell@users.noreply.github.com>
…5877)

An exception escaping a managed callback passed to native configureScope: raised
Runtime.MarshalManagedException. Its handler captured the exception, and that capture read
the replay ID through the same callback, so under memory pressure it recursed until the
stack overflowed and the original report was lost.

CocoaScope.Configure now wraps every native ConfigureScope call. While it runs, the handler
hands the exception back to Configure instead of capturing it: UnwindNativeCode on Mono,
ThrowObjectiveCException on CoreCLR. RuntimeAdapter.IsMono, which was always false on .NET,
now detects Mono, and the handler no longer throws if its own capture fails.

Fixes #5867

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
#5440)

* ref: Add internal sensitive-data denylist and key-value filter utility

First step of the DataCollection migration (#5420, #5421). Internal only,
nothing wired up yet. Mirrors filterKeyValueData/filtering-snippets from
the JS reference implementation.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* ref: Align sensitive-data filter with the spec and the Java SDK

Ignore empty and null terms when matching: an empty term would otherwise
filter every value in deny-list mode and allow every value through in
allow-list mode, which becomes reachable once terms are bound from
configuration.

Trim the denylist to the spec's canonical 17 terms. Treating cookie and
set-cookie as substring terms over-filtered every other category; the
exact header-name rule belongs at the header call site (#5422).

Decode keys before matching so an encoded key such as %74oken is
recognised, while callers still store the raw key and value.

Return an ordered list rather than a dictionary, so repeated headers such
as Set-Cookie are preserved instead of silently collapsed.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* feat: add experimental DataCollection options, successor to SendDefaultPii

Adds the option shape from the DataCollection spec as a no-op: nothing reads
it yet. The resolver and SendDefaultPii bridge follow in #5426.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* ref: Remove unreachable branches from the sensitive-data filter

Uri.UnescapeDataString never throws on malformed input, it returns the
input unchanged, so the UriFormatException handler could not run.

Off mode now returns [Filtered] from FilterValue instead of being
asserted away, which drops the unreachable default switch arm and makes
the mode's result explicit.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* feat: bind DataCollection options from configuration

Mirrors DataCollectionOptions in BindableSentryOptions so the option can be
set from appsettings.json, following the nested-bindable pattern used by the
platform options. KeyValueCollectionBehavior has no public constructor, so it
binds through a Mode/Terms surrogate.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* ref: make SentryOptions.DataCollection nullable instead of exposing Enable()

Assigning the option is now the opt-in signal, matching sentry-go, sentry-php,
sentry-python and sentry-javascript. Removes Enable() and the internal
bookkeeping it needed to distinguish an unset option from a default one.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* ref: Use the public KeyValueCollectionBehavior in the sensitive-data filter

The internal KeyValueFilterBehavior/KeyValueFilterMode pair was written
when this phase was expected to land before the public option shape. That
shape now exists with identical modes and semantics, so the internal pair
was duplication and is removed.

MatchesAny tolerates a null term list, because DenyList(null) leaves Terms
null and the SDK must not throw while capturing an event. It also iterates
by index, since IReadOnlyList allocates an enumerator where an array does
not and this runs per key per event.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* fix: KeyValueCollectionBehavior.Terms is never null

An explicit null for a params array is legal, so DenyList(null) and
AllowList(null) produced an instance whose Terms was null, which would throw
in any consumer that iterates it. Coalesce in the constructor so both
factories are covered.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* ref: Let the null-term guard carry the nullable deny-terms argument

additionalDenyTerms is nullable, so passing it straight to MatchesAny
replaces a second null check at the call site and keeps the guard
load-bearing now that KeyValueCollectionBehavior.Terms is never null.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* docs: note that key-value categories bind from Mode/Terms, not a scalar

The bool shorthand code has for cookies, headers and query params has no
configuration equivalent, so a scalar is dropped by the binder. Document the
supported shape on the public type and lock the behavior with a test.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* fix: make DataCollection bind correctly on all configuration paths

Swaps the IncomingResponse/OutgoingResponse doc comments, which described the
wrong direction. Gives KeyValueCollectionBehavior a public constructor and
settable properties so the reflection binder can construct it: on
netstandard2.0 the options are bound directly rather than through
BindableSentryOptions, and any key-value category threw there. Binds
HttpBodies from an array so the spec's empty-array opt-out works instead of
silently collecting every body. Adds coverage for the four ApplyTo lines no
test reached.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* fix: document that the HttpBodies empty-array opt-out needs ["None"] in config

A JSON empty array emits a valueless key with no children, which the binder
surfaces as null, so "[]" cannot be told apart from the option being absent
and did not disable body collection. The previous test asserted against a
hand-built list rather than the binder, so it did not catch this.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* test: cover the empty-term guard where it is now load-bearing

KeyValueCollectionBehavior sanitizes its own Terms as of #5853, so the
four behavior-level null/empty-term tests duplicated DataCollectionOptionsTests.
additionalDenyTerms is caller-supplied and unsanitized, so test the guard there.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* test: drop two empty-term tests that pin neither guard

Verified by mutation: with the empty-term guard removed from MatchesAny
both stayed green, while the additionalDenyTerms theories failed. The
behavior type strips the term before it reaches the guard.

Also document that Off requires the caller to drop the key, since
FilterValue returns [Filtered] there only to fail closed.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* fix: do not let a malformed escape throw out of DecodeKeyForMatching

Uri.UnescapeDataString returns the input unchanged for a malformed escape
on modern .NET but throws UriFormatException on .NET Framework. sentry-java
guards the same call in decodeQueryParamName. #5424 feeds untrusted query
keys through here, and the SDK must not throw while enriching an event.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* chore(warden): Remove find-bugs, already covered by code-review

Co-Authored-By: David Cramer <23610+dcramer@users.noreply.github.com>

* fix(warden): Use built-in code-review skill

The getsentry/skills code-review is a review guidelines checklist, not
Warden's bug-finding skill. Drop the remote so Warden resolves the
built-in.

Co-Authored-By: David Cramer <23610+dcramer@users.noreply.github.com>

---------

Co-authored-by: sentry-junior[bot] <264270552+sentry-junior[bot]@users.noreply.github.com>
Co-authored-by: David Cramer <23610+dcramer@users.noreply.github.com>
* ref: add the DataCollection resolver and SendDefaultPii bridge

Integrations will read their data collection policy from
DataCollectionResolver rather than from SendDefaultPii or DataCollection
directly. Resolution is lazy, since SentryOptions stays mutable after
Init, and falls back to SendDefaultPii while DataCollection is null.

No consumer migrated yet, so this is behaviour-neutral.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* ref: fix the IncomingRequestBody legacy fallback and tighten the resolver tests

Incoming request bodies have never been gated on SendDefaultPii:
RequestBodyExtractionDispatcher switches on RequestSize alone, and both
ASP.NET Core and AddAspNet call it unconditionally. That direction now
falls back to true, so #5430 cannot cost an app that set
MaxRequestBodySize its request bodies.

Tests configure one body direction at a time, and leave UserInfo at its
default where DatabaseQueryData is the subject, so that reading the wrong
category fails.

Docs say what is true today: until the integrations migrate, SendDefaultPii
still decides what they collect, and setting a category - not only
restricting one - has no effect yet.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

* ref: drop the IncomingRequestBody comment from DataCollectionResolver

The test and the PR description already carry the reasoning, and src/ aims
for no comments.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
…#5890)

Only Authorization and Proxy-Authorization were redacted, so headers such as X-Api-Key or a raw Cookie were sent verbatim. Apply the sensitive-terms denylist, plus an exact Cookie/Set-Cookie match, wherever headers are collected automatically. Header names are kept.

Fixes #5422

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
)

* fix(android): drop native events when BeforeSend conversion fails

* fix(android): separate segfault suppression from BeforeSend filtering

Co-Authored-By: Codex <noreply@openai.com>

* test(android): cover suppression with a disabled BeforeSend filter

---------

Co-authored-by: Codex <noreply@openai.com>
Its only caller, a DEBUG-only response id assert in HttpTransport, was
removed by #544 (Envelopes) in 2020.

Co-authored-by: sentry-junior[bot] <264270552+sentry-junior[bot]@users.noreply.github.com>
…5905)

Hangfire.MemoryStorage's monitoring API enumerates job state lists while the
server is writing them, which made WaitForJob throw "Collection was modified".
Hangfire.InMemory returns history newest-first, so read History[0].

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* ci: Watch the new upstream AssemblyStore reader path

dotnet/android#13380 moved the reader to tests/Utilities/AssemblyStore. The old path no longer changes, so the watcher would skip every run and never alert again.

Refs #5898

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* ref: Note why the Android assembly reader keeps LZ4

Upstream dropped LZ4 from .NET 11 builds (dotnet/android#11730) and from its own reader (dotnet/android#13380), but apps built with .NET 9 and .NET 10 still ship LZ4-compressed assemblies.

Refs #5898

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Brings in everything up to and including the 6.13.0 release.

Conflict resolutions:
- CHANGELOG.md: the 7.0.0-rc.* sections stay above 6.13.0 until they
  are consolidated just before the 7.0.0 release.
- Directory.Build.props: keep VersionPrefix 7.0.0-rc.3 and the dev
  VersionSuffix that craft cleared when it cut 6.13.0.
- SystemWebRequestEventProcessor.cs: version7 no longer needs
  Sentry.Reflection; main's SensitiveDataFilter needs DataCollection.
- SentryLoggingOptionsSetupTests.cs: keep version7's test. Since #5595
  SentryLoggingOptions no longer derives from SentryOptions, so main's
  DataCollection assertions don't apply; BindableDataCollectionOptionsTests
  covers that binding.
- API snapshots: add main's DataCollection API, drop ApplyDefaultTags,
  which version7 removed. DotNet11_0 regenerated.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@codecov

codecov Bot commented Nov 12, 2026 •

Copy link
Copy Markdown

Codecov Report

❌ Patch coverage is 99.30314% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 77.30%. Comparing base (25b505e) to head (521a9dd).
⚠️ Report is 4 commits behind head on version7.

Files with missing lines Patch % Lines
...try/Internal/DataCollection/SensitiveDataFilter.cs 97.22% 2 Missing ⚠️
Additional details and impacted files
@@             Coverage Diff              @@
##           version7    #5920      +/-   ##
============================================
+ Coverage     76.98%   77.30%   +0.32%     
============================================
  Files           538      543       +5     
  Lines         19714    19944     +230     
  Branches       3872     3921      +49     
============================================
+ Hits          15176    15418     +242     
+ Misses         3666     3661       -5     
+ Partials        872      865       -7     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

@jamescrosswell
jamescrosswell merged commit 521a9dd into version7 Nov 12, 2026
71 of 72 checks passed
@jamescrosswell
jamescrosswell deleted the ai/v7-release-merge-c71d07 branch November 12, 2026 01:59
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants