Skip to content

Extract filename from DisplayName and add extension if missing#35050

Merged
kubaflo merged 10 commits into
inflight/currentfrom
mattleibow-patch-5
Apr 22, 2026
Merged

Extract filename from DisplayName and add extension if missing#35050
kubaflo merged 10 commits into
inflight/currentfrom
mattleibow-patch-5

Conversation

@mattleibow

@mattleibow mattleibow commented Apr 20, 2026 •
edited
Loading

Copy link
Copy Markdown
Member

Note

Are you waiting for the changes in this PR to be merged?
It would be very helpful if you could test the resulting artifacts from this PR and let us know in a comment if this change resolves your issue. Thank you!

Description

When Android content providers return a DisplayName that contains directory components (e.g. /storage/emulated/0/DCIM/Camera/IMG_20240101.jpg from Samsung, or /sdcard/.transforms/synthetic/picker/0/.../photo.png from Google Drive), CacheContentFile would pass the full path as a filename to GetTemporaryFile. This caused Java.IO.File(parent, child) to either ignore the parent directory (absolute paths) or attempt to create non-existent nested directories (relative paths), leading to broken cached file paths.

Changes

  • Extract EnsureFileName helper to FileSystemUtils.shared.cs — normalizes a display name by stripping directory components, guarding against empty/whitespace/reserved directory names (., ..), and appending a known extension when the name has none
  • Add empty-string guard — if Path.GetFileName returns empty (e.g. trailing separator somefolder/), falls back to a generated GUID name instead of creating a dot-file
  • Add 17 unit tests covering: simple names, paths with directories, null/empty/whitespace display names, trailing separators, ./.. reserved names, extension handling

Tested scenarios

Verified on Android API 36 emulator using a custom ContentProvider that injects problematic DisplayName values through the full EnsurePhysicalPath → CacheContentFile → EnsureFileName pipeline:

DisplayName Result
/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg IMG_20240101.jpg ✅
/sdcard/.transforms/synthetic/.../photo.png photo.png ✅
Download/Subdir/document.pdf document.pdf ✅
somefolder/ <guid>.jpg ✅
"" (empty) <guid>.jpg ✅
" " (whitespace) <guid>.jpg ✅
screenshot_12345 (no ext) screenshot_12345.jpg ✅
normal_photo.jpg normal_photo.jpg ✅

kubaflo and others added 7 commits April 15, 2026 11:03
@kubaflo @github-actions
[CI] Extend gate to all test types and decouple from PR review (#34705)
5fef66d 
## Summary

Extends the CI PR review pipeline to support **all test types** (UI
tests, device tests, unit tests, XAML tests) and restructures the review
flow by decoupling the gate from the copilot agent.

### Before
- Gate only supported UI tests (`TestCases.HostApp` /
`TestCases.Shared.Tests`)
- PRs with device tests, unit tests, or XAML tests were **skipped** by
the gate
- Gate ran as Phase 2 inside the copilot agent (4-phase: Pre-Flight →
Gate → Try-Fix → Report)
- Gate results were duplicated across all phase outputs
- AI summary comment included session history merging (841 lines of
code)

### After
- Gate supports **all test types** with auto-detection
- Gate runs as a **standalone script step** before the copilot agent
- Gate posts its own **separate PR comment** (`<!-- AI Gate -->`)
- AI summary is **simplified** (170 lines, always overwrites, no session
history)
- PR review is now 3 phases: Pre-Flight → Try-Fix → Report

## New Scripts

| Script | Purpose |
|--------|---------|
| `Detect-TestsInDiff.ps1` | Analyzes PR files, classifies tests by type
(UITest, DeviceTest, UnitTest, XamlUnitTest), extracts method names from
diffs |
| `post-gate-comment.ps1` | Posts/updates gate result as separate PR
comment |
| `RunTests.ps1` | Unified test runner entry point for all test types |

## Test Detection

```bash
pwsh .github/scripts/shared/Detect-TestsInDiff.ps1 -PRNumber 25129
```

```
📱 [DeviceTest] EditorTests (PlaceholderHorizontalTextAlignment)
   Filter:  Category=Editor
🖥️ [UITest] Issue10987
   Filter:  Issue10987
```

## New Review Flow

```
Step 0: Branch setup
Step 1: Gate (verify-tests-fail.ps1 — direct script, no copilot agent)
         → Posts <!-- AI Gate --> comment immediately
Step 2: PR Review (copilot agent — 3 phases: Pre-Flight, Try-Fix, Report)
         → Gate result passed in prompt
Step 3: Post AI Summary (<!-- AI Summary --> comment)
Step 4: Apply labels
```

## PR Comments (Two Separate Comments)

**Gate comment** (`<!-- AI Gate -->`):
```markdown
## 🚦 Gate — Test Verification
► Expand Full Gate — abc1234 · Fix editor alignment

### Gate Result: ✅ PASSED
| Step | Expected | Actual | Result |
| Without fix | FAIL | FAIL | ✅ |
| With fix | PASS | PASS | ✅ |
```

**AI Summary comment** (`<!-- AI Summary -->`):
Pre-Flight, Fix, Report sections only — no gate duplication.

## Key Changes

- **`verify-tests-fail.ps1`**: Auto-detects test type, routes to correct
runner (BuildAndRunHostApp, Run-DeviceTests, dotnet test), iterates over
all detected tests, `-Platform` mandatory
- **`Detect-TestsInDiff.ps1`**: Shared detection engine — reads
`[Category]` attributes for device test filtering, extracts method names
from PR diffs
- **`Review-PR.ps1`**: Gate as Step 1 (script), PR review as Step 2
(copilot), removed PR finalize step
- **`post-ai-summary-comment.ps1`**: Rewritten from 841 → 170 lines,
always overwrites
- **`pr-gate.md`**: Strict output template, no cross-phase duplication
rule
- **`pr-review/SKILL.md`**: 3 phases (removed Gate), no-duplication rule
- **`EstablishBrokenBaseline.ps1`**: Excludes
TestUtils/DeviceTests.Runners from fix file detection

## Verified

- Gate passed locally on Share device tests: without fix=FAIL ✅, with
fix=PASS ✅
- Detection tested on PRs: #25129, #34615, #34598, #31056
- Comments posted to 8 PRs from CI build artifacts

---------

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
@PureWeen @github-actions
Improve evaluate-pr-tests workflow: slash_command + workflow_dispatch…
eb0b82f 
…, security hardening (#34678)

## Description

Overhauls the `copilot-evaluate-tests` gh-aw workflow — switches to
on-demand triggers only (`/evaluate-tests` slash command + manual
`workflow_dispatch`), adds security hardening, and improves error
handling. No auto-runs on PR create/update.

### What Changed

**Triggers (on-demand only — no auto-runs)**
- Add `slash_command: evaluate-tests` — comment `/evaluate-tests` on a
PR to trigger
- Keep `workflow_dispatch` — manual trigger from Actions tab with PR
number input
- Disable `pull_request_target` — no auto-evaluation on PR create/update
- Add `bots: ["copilot-swe-agent[bot]"]` — Copilot-authored PRs can be
evaluated
- Add `labels: ["pr-review", "testing"]` — workflow runs are labeled

**Gate step (fast-fail for invalid requests)**
- Check PR is OPEN before evaluating (rejects closed/merged PRs with
clear message)
- Check for test source files in diff before spinning up agent
- Fall back to REST API for PRs with 300+ files (where `gh pr diff`
returns HTTP 406)
- All API errors surfaced with clear messages — no silent masking
- `exit 1` stops the workflow immediately — no wasted agent compute

**Access gating (`Checkout-GhAwPr.ps1`)**
- Reject fork PRs (`isCrossRepository` check)
- Verify PR author has write access (admin/write/maintain roles)
- Fix `ConvertFrom-Json` ordering — check exit code before JSON parsing
- Make infrastructure restore fatal on failure (was soft warning)
- Remove pre-delete pattern — `git checkout` overwrites in-place

**Workflow improvements**
- `hide-older-comments: true` — previous evaluations auto-collapse
- `report-as-issue: false` for noop — no issue created when nothing to
evaluate
- Timeout bumped 15 → 20 minutes
- Dry-run mode via `suppress_output` input (workflow_dispatch only)
- `Gather-TestContext.ps1` now receives `-PrNumber` parameter

**Security documentation (`gh-aw-workflows.instructions.md`)**
- Add "Before You Build" anti-patterns table — prefer built-in gh-aw
features
- Add Security Boundaries section with defense layers table
- Add Rules for gh-aw Workflow Authors (DO/DON'T list)
- Document `COPILOT_TOKEN` exposure and mitigations
- Add `slash_command` to fork behavior table
- Update `Checkout-GhAwPr.ps1` description to match current behavior

### Trigger Behavior

| Trigger | When it fires | Who can trigger |
|---------|---------------|-----------------|
| `/evaluate-tests` comment | Comment on a PR | Write-access
collaborators + copilot-swe-agent[bot] |
| `workflow_dispatch` | Actions tab → "Run workflow" → enter PR number |
Write-access collaborators |
| ~~`pull_request_target`~~ | ~~Auto on PR create/update~~ |
~~Disabled~~ |

### Security Model

Based on [GitHub Security Lab
guidance](https://securitylab.github.com/resources/github-actions-preventing-pwn-requests/):

- PR contents treated as **passive data** (read/analyze, never built or
executed)
- Agent runs in **sandboxed container** with `GITHUB_TOKEN` and `gh` CLI
scrubbed
- Write operations in **separate `safe_outputs` job** (not the agent)
- Agent output limited to `max: 1` comment via safe-outputs
- `Checkout-GhAwPr.ps1` rejects fork PRs and verifies write access
before checkout
- Infrastructure restore is fatal on failure — prevents running with
untrusted infra

### Validation

| Test | PR | Result |
|------|-----|--------|
| Open PR with tests | #34983 | ✅ Full success (gate → checkout → agent
→ comment) |
| No-test PR | #34876 | ✅ Gate fast-fail ("no test source files") |
| Merged PR | #34932 | ✅ Gate fast-fail ("MERGED — skipping") |

### Known Limitations

- Fork PRs via `/evaluate-tests` can supply modified `.github/skills/` —
accepted residual risk (agent sandboxed, output bounded). Tracked as
[gh-aw#18481](github/gh-aw#18481)
- `exit 1` in gate step shows ❌ in GitHub checks for no-test/closed PRs
— intentional (no built-in "skip" mechanism in gh-aw steps)
- `pull_request_target` commented out — can be re-enabled later for
auto-evaluation

---------

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
@PureWeen
@kubaflo @github-actions
Add automated milestone management with Versions.props detection (#34686
3c06583 
)

<!-- Please let the below note in for people that find this PR -->
> [!NOTE]
> Are you waiting for the changes in this PR to be merged?
> It would be very helpful if you could [test the resulting
artifacts](https://github.com/dotnet/maui/wiki/Testing-PR-Builds) from
this PR and let us know in a comment if this change resolves your issue.
Thank you!

## Summary

Adds automated milestone management for PRs and issues. A single
workflow handles both automatic milestoning on PR merge and manual
tag-based reconciliation after releases ship.

## Problem

When PRs merge, they often get milestoned for the wrong service release
or not milestoned at all. The actual release a PR ships in depends on
which Candidate PR carries the commits and when the SR branch is cut.
This creates milestone drift that makes release queries and notes
inaccurate.

## Solution

### Files

| File | Purpose |
|------|---------|
| `.github/scripts/Fix-MilestoneDrift.ps1` | Core engine — version
detection, milestone mapping, correction logic |
| `.github/scripts/Fix-MilestoneDrift.Tests.ps1` | 77 Pester unit tests
for all pure functions |
| `.github/workflows/fix-milestone-drift.yml` | Single workflow with
dual triggers (auto on merge + manual dispatch) |

---

## How It Works

### Version Detection

The source of truth is `eng/Versions.props`. For single-PR mode, the
script reads `MajorVersion` and `PatchVersion` **at the PR's merge
commit SHA** to determine what version the branch was building when the
PR merged.

### Milestone Mapping

| Patch | Example | Milestone | Rule |
|-------|---------|-----------|------|
| `0` | `10.0.0` | `.NET 10.0 GA` | Exact zero |
| `1-9` | `10.0.5` | `.NET 10.0 SR1` | Early patches |
| `N0` | `10.0.60` | `.NET 10 SR6` | patch / 10 |
| `N0+x` | `10.0.41` | `.NET 10 SR4.1` | Sub-patch = distinct milestone
|

Sub-patches are distinct — `.NET 10 SR4.1` != `.NET 10 SR4`. Script
warns and skips gracefully if the milestone doesn't exist on GitHub yet.

### Milestone Normalization

GitHub milestones have inconsistent naming (`.NET 10.0 SR4` vs `.NET 10
SR4`, `.NET 10.0 GA` vs `.NET 10 GA`). The script normalizes both SR and
GA forms as equivalent.

### Branch Ownership Detection

Reads `<MajorVersion>` from `eng/Versions.props` on `origin/main`. No
hardcoded version numbers — automatically handles version transitions
when `main` moves from .NET 10 to .NET 11.

### Merge-Up Protection

Checks each PR's `base.ref` to skip PRs from different .NET versions
(prevents merge-up commits from causing incorrect milestoning):

| `base.ref` | MainBranch=`main` | MainBranch=`net11.0` |
|------------|-------------------|---------------------|
| `main` | ✅ | ❌ |
| `inflight/*`, `darc/*` | ✅ | ❌ |
| `net11.0` | ❌ | ✅ |
| `release/10.*` | ✅ | ❌ |
| `release/11.*` | ❌ | ✅ |

### Linked Issue Detection

Scans PR title and body for:
- `fix/fixes/fixed/close/closes/closed/resolve/resolves/resolved #N`
- Same keywords + `https://github.com/dotnet/maui/issues/N`

Bare informational URLs are ignored. Results are deduplicated.

---

## Single Workflow, Two Triggers

### Auto: On PR Merge (`pull_request_target`)

Triggers on every PR merge to `main`, `net*.0`, `inflight/*`, or
`release/*`. Reads `Versions.props` at the merge commit, sets the
milestone on the PR and its linked issues. If the milestone doesn't
exist yet on GitHub, warns and skips gracefully (no red CI).

### Manual: Tag Reconciliation (`workflow_dispatch`)

| Input | Type | Default | Description |
|-------|------|---------|-------------|
| `pr_number` | string | _(empty)_ | Single PR to check |
| `tag` | string | _(empty)_ | Release tag to audit (e.g. `10.0.60`) |
| `apply` | boolean | `false` | Apply changes (default: dry-run) |
| `create_issue` | boolean | `false` | Create GitHub issue with report |

---

## Safety

| Rule | Purpose |
|------|---------|
| Dry-run by default (manual) | `workflow_dispatch` requires explicit
`apply` checkbox |
| Date cutoff (2026-01-01) | Never touches PRs merged before 2026 |
| Branch check | Skips merge-up PRs from different .NET versions |
| Deduplication | Same issue corrected once even if linked from multiple
PRs |
| Fixing-keyword URLs | Bare informational issue references ignored |
| Milestone must exist | Warns and skips gracefully if milestone not
found |
| Merge commit fetch | Fetches commit on demand for PRs merged to
inflight |
| Warning diagnostics | Silent failures in version detection log
warnings |
| GA tag support | Tag mode handles first release tag (no previous tag)
|
| Input sanitization | Workflow inputs flow through `env:` vars, not
inline interpolation |
| Error propagation | `Invoke-ApplyCorrections` throws on failure; CI
goes red |
| Zero-check guard | Tag mode throws when all PRs fail to fetch
(prevents false "all correct") |

---

## Known Limitations

- **Preview/RC milestones**: The script currently maps `PatchVersion=0`
to `.NET X.0 GA`. It does not read `PreReleaseVersionLabel` or
`PreReleaseVersionIteration` from `Versions.props`, so PRs merged to
`net11.0` during the preview phase will not be automatically milestoned
(the workflow gracefully skips with a warning). Preview milestone
support is planned as a follow-up.
- **GA tag rate limit**: Running `-Tag X.0.0` (the very first release
tag) in manual mode walks the full git history, which could exhaust
GitHub API rate limits on large repos.

---

## Test Suite

77 Pester unit tests covering all pure functions:

| Block | Tests | Coverage |
|-------|-------|----------|
| `ConvertTo-Milestone` | 17 | GA, SR1-SR10, sub-patches (SR4.1,
SR10.1), early patches, preview/invalid |
| `Test-MilestoneMatch` | 12 | Exact, normalization (SR + GA), sub-patch
distinction, null/empty |
| `Find-MatchingMilestone` | 5 | Direct, normalized, alternate format,
non-existent |
| `Find-PreviousTag` | 9 | Ordered traversal, boundaries, major-version
isolation |
| `Get-LinkedIssues` | 11 | Keywords (incl. close/resolve bare), URLs,
dedup, case |
| `Get-PatchVersion` | 4 | Valid, GA, large, invalid |
| `Test-IsReleaseTag` | 5 | Valid release, wrong major, preview, invalid
|
| `Test-PrBelongsToVersion` | 14 | main/net11.0 contexts, inflight,
darc, release, feature |

---

## Local Usage

```powershell
# Dry-run: check a single PR
./Fix-MilestoneDrift.ps1 -PrNumber 34620 -RepoPath . -Verbose

# Dry-run: audit all PRs in a tagged release
./Fix-MilestoneDrift.ps1 -Tag 10.0.50 -RepoPath . -Verbose

# Apply fixes
./Fix-MilestoneDrift.ps1 -PrNumber 34620 -Apply

# Full tag audit + apply + report issue
./Fix-MilestoneDrift.ps1 -Tag 10.0.50 -Apply -CreateIssue

# Run tests
Invoke-Pester ./Fix-MilestoneDrift.Tests.ps1 -Output Detailed
```

## Validated Results

- 77 Pester tests pass
- Single-PR: #34047 → `PatchVersion=50` at merge → `.NET 10 SR5` ✅
- Single-PR: #34620 (untagged) → `PatchVersion=60` at merge → `.NET 10
SR6` ✅
- Single-PR: #34553 (merged to inflight) → fetches commit, reads
`PatchVersion=60` ✅
- Tag mode: `10.0.50` → 78 PRs checked, 90 corrections, 0 duplicates, 0
false positives ✅
- 50-PR sample against 10.0.60 range → correctly identifies SR7→SR6
drift ✅
- .NET 11 preview PR #34252 → gracefully warns "milestone not found",
exits 0 ✅

---------

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Jakub Florkowski <kubaflo123@gmail.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
@PureWeen @github-actions
Add preview/RC milestone support with release branch detection (#34999)
f4fb086 
<!-- Please let the below note in for people that find this PR -->
> [!NOTE]
> Are you waiting for the changes in this PR to be merged?
> It would be very helpful if you could [test the resulting
artifacts](https://github.com/dotnet/maui/wiki/Testing-PR-Builds) from
this PR and let us know in a comment if this change resolves your issue.
Thank you!

## Summary

Follow-up to #34686. Adds preview/RC milestone support and release
branch detection to the milestone drift fixer.

## Problem

PRs merged to \`net11.0\` were not being milestoned because
\`Versions.props\` on that branch always has
\`PreReleaseVersionIteration=1\` regardless of which preview the PR
actually ships in. The iteration is only bumped on release branches.

## Solution

### Release Branch Detection (Primary)

New detection step checks release branches first using \`merge-base
--is-ancestor\`. For each PR, it finds the **earliest** release branch
containing the merge commit:

| Release Branch | Milestone |
|---|---|
| \`release/10.0.1xx\` | \`.NET 10.0 GA\` |
| \`release/10.0.1xx-sr5\` | \`.NET 10 SR5\` |
| \`release/11.0.1xx-preview1\` | \`.NET 11.0-preview1\` |
| \`release/11.0.1xx-preview3\` | \`.NET 11.0-preview3\` |
| \`release/12.0.1xx-rc1\` | \`.NET 12.0-rc1\` |

### Detection Order

1. **Explicit \`-Tag\`** — if provided
2. **Release branches** — \`merge-base --is-ancestor\` against
\`release/{Major}.0.1xx-*\` branches, earliest match wins
3. **Versions.props** at merge commit — fallback for PRs not yet on any
release branch
4. **Tag range search** — last resort

### Preview/RC Milestone Mapping

\`ConvertTo-Milestone\` now accepts optional pre-release label and
iteration:

| Input | Milestone |
|---|---|
| \`11.0.0 + preview + 3\` | \`.NET 11.0-preview3\` |
| \`12.0.0 + rc + 1\` | \`.NET 12.0-rc1\` |
| \`10.0.60\` (stable) | \`.NET 10 SR6\` (unchanged) |

## Validated Results

| PR | Base | Current Milestone | Script Result | Method |
|---|---|---|---|---|
| #33524 | net11.0 | .NET 11.0-preview1 | .NET 11.0-preview1 ✅ | Release
branch |
| #33233 | net11.0 | .NET 11.0-preview1 | .NET 11.0-preview1 ✅ | Release
branch |
| #30132 | net11.0 | .NET 11.0-preview3 | .NET 11.0-preview3 ✅ | Release
branch |
| #33834 | net11.0 | .NET 11.0-preview3 | .NET 11.0-preview3 ✅ | Release
branch |
| #34214 | net11.0 | .NET 11.0-preview2 | .NET 11.0-preview3 ✅ | Release
branch (drift caught!) |
| #34945 | net11.0 | .NET 11.0-preview4 | preview1 (fallback) |
Versions.props (no p4 branch yet) |
| #34620 | main | .NET 10 SR6 | .NET 10 SR6 ✅ | Release branch |
| #34047 | main | .NET 10 SR4.1 | .NET 10 SR5 ✅ | Release branch (drift
caught!) |

PR #34214 is a real drift example: milestoned preview2 by a human, but
actually on the preview3 release branch.

## Test Suite

88 Pester tests (11 new):
- 6 for \`ConvertTo-Milestone\` preview/RC mapping
- 5 for \`ConvertBranchToMilestone\` (GA, SR, preview, RC, non-release)

---------

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
@kubaflo @github-actions
Add daily PR review queue workflow with actionability detection (#34818)
b5f9079 
<!-- Please let the below note in for people that find this PR -->
> [!NOTE]
> Are you waiting for the changes in this PR to be merged?
> It would be very helpful if you could [test the resulting
artifacts](https://github.com/dotnet/maui/wiki/Testing-PR-Builds) from
this PR and let us know in a comment if this change resolves your issue.
Thank you!

## Description

Adds a daily GitHub Actions workflow that generates a prioritized,
actionable PR review queue as a GitHub issue -- with intelligent
turn-state detection that tells you **whose turn it is** on each PR.

### What it does
- Runs weekdays at 8:00 UTC (or on-demand via `workflow_dispatch`)
- Queries open PRs across `dotnet/maui` and `dotnet/docs-maui`
- Determines actionability for each PR using a layered heuristic:
1. **Bot detection** -- filters out maestro, dependabot,
copilot-reviewer, etc.
  2. **Blocked states** -- do-not-merge, stale labels
3. **Project board status** -- reads the MAUI SDK Ongoing board (Ready
To Review, In Progress, Changes Requested, Todo, Approved)
4. **Review decision** -- APPROVED, CHANGES_REQUESTED with
author-response detection
5. **"Who spoke last"** -- compares author activity (issue comments +
inline review-thread replies) against reviewer activity timestamps
- Creates a GitHub issue organized by actionability: FTE-actionable PRs
first, waiting-on-author collapsed, bot PRs collapsed
- Auto-closes previous queue issues to avoid clutter
- Filters out `stale` and `do-not-merge` PRs

### Implementation
Uses a **plain GitHub Action** (not gh-aw) since the task is
deterministic formatting -- no AI reasoning needed:
- ~30s runtime (vs ~2min with LLM)
- Zero LLM cost
- No `COPILOT_GITHUB_TOKEN` dependency -- just built-in `GITHUB_TOKEN`
- Deterministic output every run
- No prompt injection surface

### Changes
- **New**: `.github/workflows/pr-review-queue.yml` -- scheduled workflow
with `schedule`, `workflow_dispatch`, and `pull_request` (validation
only) triggers
- **Modified**: `query-reviewable-prs.ps1` -- adds:
  - `-OutputFormat markdown` option with actionability-grouped output
- `Get-ProjectBoardStatuses` -- paginates full project board for all
status columns
  - `Get-TurnState` -- 5-layer decision tree for "whose turn is it"
  - `Test-IsBot` -- bot account detection
  - Comment + review-thread analysis for author activity detection
  - `#Requires -Version 7.0` for null-coalescing operator

### Example output
The daily issue shows:
- **Actionability Summary** table (FTE-actionable count, waiting on
author, backlog, bot, blocked)
- **P/0 Priority** PRs with turn state
- **Ready to Merge** (approved PRs)
- **Needs FTE Review** (milestoned first, then by age, with turn-state
detail)
- **Waiting on Author** (collapsed `<details>` section)
- **Bot / Automated** (collapsed)
- **Queue Health** stats

---------

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
@PureWeen @github-actions
Cache Gradle downloads between CI runs to avoid transient failures (#…
2e13713 
…34804)

## Problem

Android binding builds invoke Gradle to resolve Maven dependencies
(e.g., `androidx` packages). When Gradle downloads fail due to transient
DNS/network issues on CI agents, the build fails with:

```
gradlew exited with code 1
```

This is intermittent and often passes on retrigger, but wastes CI
capacity and blocks PRs. See #34800 for recent examples.

## Fix

Add a shared `cache-gradle.yml` template using the AzDO
[`Cache@2`](https://learn.microsoft.com/azure/devops/pipelines/release/caching)
task to persist `~/.gradle/caches` between pipeline runs. When the cache
is warm, Gradle resolves dependencies locally instead of downloading
from Maven Central, eliminating transient network failures.

The cache key includes:
- `Agent.OS` — separate caches for macOS/Windows/Linux
- `gradle-wrapper.properties` — invalidates when Gradle version changes
- `build.gradle` files — invalidates when dependencies change

The template is called from `provision.yml` so it applies to all
pipelines (build, pack, device tests, UI tests).

## Prior art

Adapted from the same approach in dotnet/android:
dotnet/android@0a68102
(PR #10986)

Fixes #34800

---------

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
@PureWeen @github-actions
Fix milestone fallback to use main's Versions.props instead of merge …
42d0dbf 
…commit (#35023)

<!-- Please let the below note in for people that find this PR -->
> [!NOTE]
> Are you waiting for the changes in this PR to be merged?
> It would be very helpful if you could [test the resulting
artifacts](https://github.com/dotnet/maui/wiki/Testing-PR-Builds) from
this PR and let us know in a comment if this change resolves your issue.
Thank you!

## Summary

Fixes milestone fallback to use `Versions.props` from `origin/main`
instead of the merge commit when no release branch contains the PR.

## Problem

When a PR merges to `inflight/current`, the script was reading
`Versions.props` at the merge commit. This gives stale values — e.g., PR
#34228 merged when `PatchVersion=60` (SR6), but `main` has since moved
to `PatchVersion=70` (SR7). The PR will ship in SR7, not SR6.

## Fix

When no release branch contains the commit, read `Versions.props` from
the development branch HEAD (`origin/main` for .NET 10, `origin/net11.0`
for .NET 11) instead of the merge commit. This reflects where the PR is
actually heading.

Release branch detection still takes priority — PRs already on a release
branch get the correct milestone from that branch.

## Validated

| PR | Before | After | Method |
|---|---|---|---|
| #34228 (inflight, stale) | .NET 10 SR6 (wrong) | .NET 10 SR7 (correct)
| main's Versions.props |
| #34620 (on SR6 branch) | .NET 10 SR6 | .NET 10 SR6 (unchanged) |
Release branch |
| #30132 (on preview3 branch) | .NET 11.0-preview3 | .NET 11.0-preview3
(unchanged) | Release branch |

91 Pester tests pass.

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
@github-actions

github-actions Bot commented Apr 20, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

🚀 Dogfood this PR with:

⚠️ WARNING: Do not do this without first carefully reviewing the code of this PR to satisfy yourself it is safe.

curl -fsSL https://raw.githubusercontent.com/dotnet/maui/main/eng/scripts/get-maui-pr.sh | bash -s -- 35050

Or

  • Run remotely in PowerShell:
iex "& { $(irm https://raw.githubusercontent.com/dotnet/maui/main/eng/scripts/get-maui-pr.ps1) } 35050"

@mattleibow mattleibow changed the title Extract filename and add extension if missing Extract filename from DisplayName and add extension if missing Apr 20, 2026
@mattleibow mattleibow marked this pull request as ready for review April 20, 2026 22:52
Copilot AI review requested due to automatic review settings April 20, 2026 22:52
Copilot AI reviewed Apr 20, 2026
View reviewed changes

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This PR hardens Android CacheContentFile behavior when content providers return a DisplayName that includes directory components, by extracting a shared filename-sanitization helper and adding unit tests around the new behavior.

Changes:

  • Add FileSystemUtils.EnsureFileName helper to strip directory components from DisplayName and append a known extension when missing.
  • Update Android CacheContentFile to use EnsureFileName instead of using the raw DisplayName.
  • Add unit tests covering common and edge-case DisplayName inputs and extension handling.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File Description
src/Essentials/src/FileSystem/FileSystemUtils.shared.cs Introduces EnsureFileName helper for normalizing/sanitizing content provider display names.
src/Essentials/src/FileSystem/FileSystemUtils.android.cs Uses EnsureFileName when constructing the temp file name for cached content URIs.
src/Essentials/test/UnitTests/FileSystemUtils_Tests.cs Adds test coverage for EnsureFileName across path-like display names and extension scenarios.

Comment thread src/Essentials/src/FileSystem/FileSystemUtils.shared.cs
Comment thread src/Essentials/src/FileSystem/FileSystemUtils.shared.cs Outdated
@mattleibow mattleibow requested a review from Copilot April 20, 2026 23:11
Copilot AI reviewed Apr 20, 2026
View reviewed changes

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

Comment thread src/Essentials/test/UnitTests/FileSystemUtils_Tests.cs
@mattleibow mattleibow requested a review from Copilot April 20, 2026 23:37
@mattleibow
Extract filename from DisplayName and add extension if missing
3b2c17d 
- Extract EnsureFileName to FileSystemUtils.shared.cs for cross-platform testability
- Strip directory components from DisplayName — some Android content providers
  (Samsung, Google Drive, third-party file managers) return full paths
- Add empty/whitespace guard — falls back to GUID instead of creating dot-files
- Add extension when the filename has none and a known extension is available
- Add 15 unit tests covering all edge cases

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Copilot AI reviewed Apr 20, 2026
View reviewed changes

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated no new comments.

This was referenced Apr 21, 2026
Closed
Closed
@MauiBot

MauiBot commented Apr 21, 2026 •
edited
Loading

Copy link
Copy Markdown
Collaborator

🤖 AI Summary

👋 @mattleibow — new AI review results are available. Please review the latest session below.

📊 Review Session — 3b2c17d · Extract filename from DisplayName and add extension if missing · 2026-04-21 10:28 UTC
🚦 Gate — Test Before & After Fix

Gate Result: ✅ PASSED

Platform: ANDROID · Base: main · Merge base: eb0b82fe

Test Without Fix (expect FAIL) With Fix (expect PASS)
🧪 FileSystemUtils_Tests FileSystemUtils_Tests ✅ FAIL — 15s ✅ PASS — 8s
🔴 Without fix — 🧪 FileSystemUtils_Tests: FAIL ✅ · 15s 
  Determining projects to restore...
  Restored /home/vsts/work/1/s/src/Graphics/src/Graphics/Graphics.csproj (in 712 ms).
  Restored /home/vsts/work/1/s/src/Essentials/test/UnitTests/Essentials.UnitTests.csproj (in 2.54 sec).
  Restored /home/vsts/work/1/s/src/Essentials/src/Essentials.csproj (in 5.05 sec).
  ##vso[build.updatebuildnumber]10.0.70-ci+azdo.13894288
  Graphics -> /home/vsts/work/1/s/artifacts/bin/Graphics/Debug/net10.0/Microsoft.Maui.Graphics.dll
  ##vso[build.updatebuildnumber]10.0.70-ci+azdo.13894288
  Essentials -> /home/vsts/work/1/s/artifacts/bin/Essentials/Debug/net10.0/Microsoft.Maui.Essentials.dll
/home/vsts/work/1/s/src/Essentials/test/UnitTests/FileSystemUtils_Tests.cs(285,33): error CS0117: 'FileSystemUtils' does not contain a definition for 'EnsureFileName' [/home/vsts/work/1/s/src/Essentials/test/UnitTests/Essentials.UnitTests.csproj]
/home/vsts/work/1/s/src/Essentials/test/UnitTests/FileSystemUtils_Tests.cs(295,33): error CS0117: 'FileSystemUtils' does not contain a definition for 'EnsureFileName' [/home/vsts/work/1/s/src/Essentials/test/UnitTests/Essentials.UnitTests.csproj]
/home/vsts/work/1/s/src/Essentials/test/UnitTests/FileSystemUtils_Tests.cs(302,33): error CS0117: 'FileSystemUtils' does not contain a definition for 'EnsureFileName' [/home/vsts/work/1/s/src/Essentials/test/UnitTests/Essentials.UnitTests.csproj]
/home/vsts/work/1/s/src/Essentials/test/UnitTests/FileSystemUtils_Tests.cs(312,33): error CS0117: 'FileSystemUtils' does not contain a definition for 'EnsureFileName' [/home/vsts/work/1/s/src/Essentials/test/UnitTests/Essentials.UnitTests.csproj]
/home/vsts/work/1/s/src/Essentials/test/UnitTests/FileSystemUtils_Tests.cs(319,33): error CS0117: 'FileSystemUtils' does not contain a definition for 'EnsureFileName' [/home/vsts/work/1/s/src/Essentials/test/UnitTests/Essentials.UnitTests.csproj]
/home/vsts/work/1/s/src/Essentials/test/UnitTests/FileSystemUtils_Tests.cs(326,33): error CS0117: 'FileSystemUtils' does not contain a definition for 'EnsureFileName' [/home/vsts/work/1/s/src/Essentials/test/UnitTests/Essentials.UnitTests.csproj]
/home/vsts/work/1/s/src/Essentials/test/UnitTests/FileSystemUtils_Tests.cs(338,33): error CS0117: 'FileSystemUtils' does not contain a definition for 'EnsureFileName' [/home/vsts/work/1/s/src/Essentials/test/UnitTests/Essentials.UnitTests.csproj]
/home/vsts/work/1/s/src/Essentials/test/UnitTests/FileSystemUtils_Tests.cs(348,33): error CS0117: 'FileSystemUtils' does not contain a definition for 'EnsureFileName' [/home/vsts/work/1/s/src/Essentials/test/UnitTests/Essentials.UnitTests.csproj]
/home/vsts/work/1/s/src/Essentials/test/UnitTests/FileSystemUtils_Tests.cs(360,33): error CS0117: 'FileSystemUtils' does not contain a definition for 'EnsureFileName' [/home/vsts/work/1/s/src/Essentials/test/UnitTests/Essentials.UnitTests.csproj]
/home/vsts/work/1/s/src/Essentials/test/UnitTests/FileSystemUtils_Tests.cs(371,33): error CS0117: 'FileSystemUtils' does not contain a definition for 'EnsureFileName' [/home/vsts/work/1/s/src/Essentials/test/UnitTests/Essentials.UnitTests.csproj]
🟢 With fix — 🧪 FileSystemUtils_Tests: PASS ✅ · 8s 
  Determining projects to restore...
  All projects are up-to-date for restore.
  ##vso[build.updatebuildnumber]10.0.70-ci+azdo.13894288
  Graphics -> /home/vsts/work/1/s/artifacts/bin/Graphics/Debug/net10.0/Microsoft.Maui.Graphics.dll
  ##vso[build.updatebuildnumber]10.0.70-ci+azdo.13894288
  Essentials -> /home/vsts/work/1/s/artifacts/bin/Essentials/Debug/net10.0/Microsoft.Maui.Essentials.dll
  Essentials.UnitTests -> /home/vsts/work/1/s/artifacts/bin/Essentials.UnitTests/Debug/net10.0/Microsoft.Maui.Essentials.UnitTests.dll
Test run for /home/vsts/work/1/s/artifacts/bin/Essentials.UnitTests/Debug/net10.0/Microsoft.Maui.Essentials.UnitTests.dll (.NETCoreApp,Version=v10.0)
VSTest version 18.0.1 (x64)

Starting test execution, please wait...
A total of 1 test files matched the specified pattern.
[xUnit.net 00:00:00.00] xUnit.net VSTest Adapter v2.8.2+699d445a1a (64-bit .NET 10.0.0)
[xUnit.net 00:00:00.10]   Discovering: Microsoft.Maui.Essentials.UnitTests
[xUnit.net 00:00:00.32]   Discovered:  Microsoft.Maui.Essentials.UnitTests
[xUnit.net 00:00:00.33]   Starting:    Microsoft.Maui.Essentials.UnitTests
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_DotDotSegment_ReturnsFalse(path: "../../file.txt") [4 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_DotDotSegment_ReturnsFalse(path: "sub/../../file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_DotDotSegment_ReturnsFalse(path: "sub\\..\\file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_DotDotSegment_ReturnsFalse(path: "..\\file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_DotDotSegment_ReturnsFalse(path: "sub/../file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_DotDotSegment_ReturnsFalse(path: "../file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_RootedPath_ReturnsFalse(path: "///file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_RootedPath_ReturnsFalse(path: "/file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_RootedPath_ReturnsFalse(path: "//file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_SimpleNameWithExtension_ReturnsUnchanged(displayName: "document.pdf", extension: ".pdf", expected: "document.pdf") [2 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_SimpleNameWithExtension_ReturnsUnchanged(displayName: "readme.txt", extension: ".doc", expected: "readme.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_SimpleNameWithExtension_ReturnsUnchanged(displayName: "photo.jpg", extension: null, expected: "photo.jpg") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_ValidRelative_ReturnsTrue(path: "file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_ValidRelative_ReturnsTrue(path: "sub/./file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_ValidRelative_ReturnsTrue(path: "sub/file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_ValidRelative_ReturnsTrue(path: "sub/deep/file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_ValidRelative_ReturnsTrue(path: "./file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_BackslashDotDot_ReturnsNull(relativePath: "..\\readme.txt") [2 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_BackslashDotDot_ReturnsNull(relativePath: "subfolder\\..\\..\\readme.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_BackslashDotDot_ReturnsNull(relativePath: "..\\..\\data\\config.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_RelativeRoot_RootedRelative_ReturnsNull(relativePath: "/etc/config.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_RelativeRoot_RootedRelative_ReturnsNull(relativePath: "//images/logo.png") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_EmptyOrWhitespace_GeneratesGuid(displayName: "") [3 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_EmptyOrWhitespace_GeneratesGuid(displayName: "   ") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_NullOrEmpty_ReturnsTrue(path: null) [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_NullOrEmpty_ReturnsTrue(path: "") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_AbsoluteRoot_DotDotAboveRoot_ReturnsNull(relativePath: "../file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_AbsoluteRoot_DotDotAboveRoot_ReturnsNull(relativePath: "sub/../../file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_AbsoluteRoot_DotDotAboveRoot_ReturnsNull(relativePath: "../../file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_RelativeRoot_NormalizesSlashes [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_WindowsDriveLetter_ReturnsFalse [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_WindowsDriveLetter_ReturnsNull [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_RelativeRoot_DotDotAboveRoot_ReturnsNull(relativePath: "../../file.txt") [2 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_RelativeRoot_DotDotAboveRoot_ReturnsNull(relativePath: "../file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_RelativeRoot_DotDotAboveRoot_ReturnsNull(relativePath: "sub/../../file.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_Null_GeneratesGuid [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_TrailingSeparator_GeneratesGuid [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_NetworkStylePath_ReturnsNull(relativePath: "///localhost/C$/data.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_NetworkStylePath_ReturnsNull(relativePath: "///127.0.0.1/share/data.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_DotOrDotDot_GeneratesGuid(displayName: "..") [9 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_DotOrDotDot_GeneratesGuid(displayName: ".") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_EmptyExtension_DoesNotAppend [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_PathWithDirectories_StripsToFilename(displayName: "/data/user/0/com.app/files/doc.pdf", extension: ".pdf", expected: "doc.pdf") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_PathWithDirectories_StripsToFilename(displayName: "some/relative/path/image.png", extension: null, expected: "image.png") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_PathWithDirectories_StripsToFilename(displayName: "/storage/emulated/0/DCIM/photo.jpg", extension: null, expected: "photo.jpg") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_DoubleDotInFilename_ReturnsTrue(path: "a..b/c..d") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_DoubleDotInFilename_ReturnsTrue(path: "image..png") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_DoubleDotInFilename_ReturnsTrue(path: "name...ext") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.IsValidRelativePath_DoubleDotInFilename_ReturnsTrue(path: "foo..bar.js") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_PathWithDirectoriesNoExtension_AppendsExtension [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.NormalizePath_ReplacesForwardAndBackSlashes [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_RelativeRoot_ValidRelative_ReturnsRelativePath [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_RelativeRoot_SubPath_PreservesRelativeStructure [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_AbsoluteRoot_EmptyRelative_ReturnsRoot [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_NoExtension_AppendsExtension(displayName: "screenshot_12345", extension: ".png", expected: "screenshot_12345.png") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_NoExtension_AppendsExtension(displayName: "photo", extension: "jpg", expected: "photo.jpg") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_NoExtension_AppendsExtension(displayName: "download", extension: ".pdf", expected: "download.pdf") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_AbsoluteRoot_WindowsCaseSensitivity [2 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_AbsoluteRoot_RootedRelative_ReturnsNull(relativePath: "///data/readme.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_AbsoluteRoot_RootedRelative_ReturnsNull(relativePath: "/etc/config.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_AbsoluteRoot_RootedRelative_ReturnsNull(relativePath: "//images/logo.png") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_DoubleDotInFilename_Succeeds(relativePath: "image..png") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_DoubleDotInFilename_Succeeds(relativePath: "foo..bar.js") [< 1 ms]
[xUnit.net 00:00:00.49]   Finished:    Microsoft.Maui.Essentials.UnitTests
  Passed Tests.FileSystemUtils_Tests.Combine_DoubleDotInFilename_Succeeds(relativePath: "a..b/c..d.txt") [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_AbsoluteRoot_SingleDot_ReturnsPathWithinRoot [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.EnsureFileName_NoExtensionAndNoKnownExtension_ReturnsAsIs [< 1 ms]
  Passed Tests.FileSystemUtils_Tests.Combine_AbsoluteRoot_ValidRelative_ReturnsFullPath [< 1 ms]

Test Run Successful.
Total tests: 67
     Passed: 67
 Total time: 1.2347 Seconds
📁 Fix files reverted (3 files)
  • eng/pipelines/common/provision.yml
  • src/Essentials/src/FileSystem/FileSystemUtils.android.cs
  • src/Essentials/src/FileSystem/FileSystemUtils.shared.cs

New files (not reverted):

  • eng/pipelines/common/cache-gradle.yml
🔍 Pre-Flight — Context & Validation

Issue: No linked issue (standalone fix)
PR: #35050 - Extract filename from DisplayName and add extension if missing
Platforms Affected: Android
Files Changed: 2 implementation, 1 test

Key Findings

  • Android content providers (Samsung Gallery, Google Drive) sometimes return a full filesystem path (e.g., /storage/emulated/0/DCIM/Camera/IMG_20240101.jpg) in MediaStore.IMediaColumns.DisplayName instead of a bare filename
  • CacheContentFile previously passed this raw value directly to GetTemporaryFile(root, fileName), which calls Java.IO.File(parent, child) — an absolute child path silently ignores parent; a relative child path attempts to create non-existent nested subdirectories; both result in broken cached file paths
  • Fix extracts a new EnsureFileName helper in FileSystemUtils.shared.cs that strips directory components via Path.GetFileName, guards against empty/whitespace/reserved names (., ..), and appends a known extension when missing
  • 17 new unit tests cover all documented edge cases; 67 total tests pass
  • Inline review comment from Copilot about ./.. guard was already addressed in the current PR

Code Review Summary

Verdict: LGTM
Confidence: high
Errors: 0 | Warnings: 0 | Suggestions: 2

Key code review findings:

  • 💡 EnsureFileName is in shared.cs but only called from Android — correct placement for testability, but a brief comment noting intended scope would help future readers
  • 💡 No test for Windows-backslash DisplayName input — behavior is safe (backslash is not a separator on Linux), but the platform assumption is implicit

Fix Candidates

# Source Approach Test Result Files Changed Notes
PR PR #35050 Extract EnsureFileName helper that strips directory components, guards reserved names, appends extension ✅ PASSED (Gate) FileSystemUtils.shared.cs, FileSystemUtils.android.cs 17 new unit tests
🔬 Code Review — Deep Analysis

Code Review — PR #35050

Independent Assessment

What this changes: Extracts a new EnsureFileName(string? displayName, string? extension) helper in FileSystemUtils.shared.cs. The helper strips directory components from an Android content-provider DisplayName via Path.GetFileName, guards against empty/whitespace/reserved names (., ..), and appends a known extension when the name has none. The Android CacheContentFile method is simplified to delegate to it. 17 unit tests are added covering the full input space.

Inferred motivation: Some Android content providers (e.g., Samsung Gallery, Google Drive) return a full filesystem path instead of a bare filename in MediaStore.IMediaColumns.DisplayName. Passing that full path directly to Java.IO.File(parent, child) has two broken behaviors: absolute child paths silently ignore parent, and relative child paths attempt to create non-existent nested subdirectories. Both result in broken cached file paths.

Reconciliation with PR Narrative

Author claims: Exactly matches — Samsung and Google Drive return full paths in DisplayName, causing Java.IO.File(parent, child) to misbehave. The PR adds a sanitizing helper and guards against empty, whitespace, ., and .. filenames.

Agreement: Full agreement. The root cause analysis is correct and the code change directly addresses both the absolute-path and the no-extension sub-cases.

Findings

💡 Suggestion — EnsureFileName is in shared.cs but only called from Android

EnsureFileName compiles on all platforms but today is only used from CacheContentFile (Android-only). Placing it in shared.cs is actually the right call — it keeps it out of Android bindings and makes it independently unit-testable — but a brief inline comment noting the intended scope would help future readers understand why a function named with Android-specific semantics lives in the shared file.

💡 Suggestion — No test for a Windows-backslash DisplayName

On Android, Path.GetFileName only treats / as a separator, so an input like "DCIM\\photo.jpg" would be returned unchanged (backslash is not a separator on Linux). A comment in the implementation noting that only forward-slash paths are expected from Android content providers would make this platform assumption explicit. The current behavior (leave it unchanged) is safe since Java.IO.File(parent, "DCIM\\photo.jpg") creates a filename with a literal backslash rather than a path traversal, but it's a latent gotcha.

Devil's Advocate

"Could removing ./.. accidentally drop legitimate filenames like .bashrc?" No — Path.GetFileName(".bashrc") = ".bashrc", which is not equal to "." or "..". Only the exact directory-sentinel values are guarded.

"Could two different URIs map to the same cached filename?" GetTemporaryFile already creates a per-call GUID subdirectory, so even if two URIs produce vacation.jpg, they land in different subdirectories. No collision risk.

"Is assigning Path.GetFileName(filename) back to filename safe under #nullable enable?" Yes. Path.GetFileName carries [return: NotNullIfNotNull("path")], so when the input is guaranteed non-null (which it is at that point), the compiler treats the return as non-null. No warning, no suppression.

"Could this break any pre-existing callers?" The refactoring is transparent: CacheContentFile produced identical results for the happy path (plain filename with extension already present), and the new helper produces the same output for that case. The only behavioral changes are in the bug paths.

"Is CI trustworthy here?" All CI jobs pass. The new unit tests run and pass.

Verdict: LGTM

Confidence: high

Summary: The fix is correct and surgical. EnsureFileName properly isolates the sanitization logic, all edge cases (full absolute paths, relative paths, trailing separators, reserved names, null/empty/whitespace) are handled and tested, and the behavioral change for the existing happy-path callers is zero. The two 💡 suggestions are documentation-only and non-blocking.

🔧 Fix — Analysis & Comparison

Fix Candidates

# Source Approach Test Result Files Changed Notes
1 try-fix (opus) LastIndexOfAny('/', '\\') to strip directory separators; handles backslash on Linux too ✅ PASS 3 files Adds backslash edge-case coverage vs PR
2 try-fix (sonnet) string.Split('/', '\\') take last segment; FallbackFileName helper ✅ PASS 2 files More verbose than PR; Split pattern consistent with file
3 try-fix (codex) NormalizePath() then Path.GetFileName() on normalized string ✅ PASS 2 files Leverages existing helper; also handles backslash
4 try-fix (gpt-4.1) Android Java.IO.File.getName() inline in CacheContentFile ❌ FAIL (infra) — Wrong test invocation; code approach not validated
PR PR #35050 Path.GetFileName() in shared EnsureFileName helper; GUID fallback for empty/./..; append extension ✅ PASSED (Gate) 2 files Simplest; idiomatic .NET; 17 tests

Cross-Pollination

Model Round New Ideas? Details
claude-opus-4.6 2 No NO NEW IDEAS
gpt-5.3-codex 2 Proposed Use ContentResolver OpenableColumns.DisplayName — not viable (code already uses ContentResolver; problem is the value returned, not how it's fetched)

Exhausted: Yes — all meaningful path-stripping approaches covered (Path.GetFileName, LastIndexOfAny, Split, NormalizePath+GetFileName)

Selected Fix: PR's fix — Path.GetFileName() in a shared EnsureFileName helper

Reason: Simplest and most idiomatic .NET implementation. All alternatives that passed are functionally equivalent variants of the same core algorithm. The PR's approach uses the BCL's canonical Path.GetFileName() which is already designed for this purpose, requires no additional logic, and is the most readable. The 3 alternatives that handle backslash (attempts 1, 2, 3) add marginal coverage for a case that doesn't occur on Android (content provider paths always use /). The 17 unit tests in the PR comprehensively cover the documented edge cases.

📋 Report — Final Recommendation

✅ Final Recommendation: APPROVE

Phase Status

Phase Status Notes
Pre-Flight ✅ COMPLETE No linked issue; Android-only fix; unit tests
Code Review LGTM (high) 0 errors, 0 warnings, 2 non-blocking suggestions
Gate ✅ PASSED android — 67/67 tests pass
Try-Fix ✅ COMPLETE 4 attempts, 3 passing (1 infra failure)
Report ✅ COMPLETE

Code Review Impact on Try-Fix

Code review found LGTM with high confidence and no errors or warnings — only two non-blocking 💡 suggestions. The backslash edge-case suggestion directly inspired Attempt 1 (LastIndexOfAny) and Attempt 3 (NormalizePath + Path.GetFileName), both of which passed. Neither added meaningful value over the PR's approach because Android content providers exclusively use forward-slash paths. The code review's LGTM verdict confirmed that no model needed to address correctness concerns, allowing all 4 models to focus on exploring alternative algorithms rather than fixing bugs.

Summary

PR #35050 fixes a real Android bug where content providers (Samsung, Google Drive) return a full filesystem path in MediaStore.IMediaColumns.DisplayName, causing Java.IO.File(parent, child) to either ignore the parent directory or attempt to create non-existent nested subdirectories. The fix is well-scoped, correctly targeted, comprehensively tested with 17 new unit tests, and all Gate + Try-Fix results confirm it works. Three independent alternative implementations also passed tests, all converging on the same core algorithm — confirming the fix's soundness.

Root Cause

CacheContentFile passed the raw DisplayName value directly to GetTemporaryFile(root, fileName). Since DisplayName can contain a full path (e.g., /storage/emulated/0/DCIM/photo.jpg), Java.IO.File(parent, child) received either an absolute child (ignoring parent) or a relative child with non-existent subdirectories. A secondary issue was that display names without extensions would not get the MIME-derived extension appended.

Fix Quality

The fix is clean and surgical: a new EnsureFileName helper in FileSystemUtils.shared.cs centralizes the sanitization logic (strip directory components via Path.GetFileName, GUID fallback for empty/whitespace/./.., append extension when absent). The Android CacheContentFile method delegates to it in 3 lines replacing 4. The helper is placed in the shared file for platform-agnostic unit testability, which is the correct architectural choice. The 17 unit tests cover all documented edge cases, and 3 independently implemented alternatives all passed using equivalent logic — providing strong validation of correctness.

@MauiBot MauiBot added s/agent-approved AI agent recommends approval - PR fix is correct and optimal s/agent-fix-pr-picked AI could not beat the PR fix - PR is the best among all candidates s/agent-reviewed PR was reviewed by AI agent workflow (full 4-phase review) labels Apr 21, 2026
@kubaflo kubaflo changed the base branch from main to inflight/current April 21, 2026 11:48
kubaflo and others added 2 commits April 21, 2026 13:50
@kubaflo
Merge branch 'inflight/current' into mattleibow-patch-5
67e88ee
@kubaflo
Remove extraneous eng/pipelines changes from diff
e980750 
Revert cache-gradle.yml and provision.yml to match inflight/current
base branch — these were carried over from main when the PR base was
changed.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
This was referenced Apr 22, 2026
Closed
Closed
@kubaflo kubaflo merged commit be797c0 into inflight/current Apr 22, 2026
26 of 36 checks passed
@kubaflo kubaflo deleted the mattleibow-patch-5 branch April 22, 2026 11:28
@github-actions github-actions Bot added this to the .NET 10 SR7 milestone Apr 22, 2026
PureWeen pushed a commit that referenced this pull request Apr 22, 2026
@mattleibow @kubaflo @PureWeen
Extract filename from DisplayName and add extension if missing (#35050)
21332ed 
<!-- Please let the below note in for people that find this PR -->
> [!NOTE]
> Are you waiting for the changes in this PR to be merged?
> It would be very helpful if you could [test the resulting
artifacts](https://github.com/dotnet/maui/wiki/Testing-PR-Builds) from
this PR and let us know in a comment if this change resolves your issue.
Thank you!

## Description

When Android content providers return a `DisplayName` that contains
directory components (e.g.
`/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` from Samsung, or
`/sdcard/.transforms/synthetic/picker/0/.../photo.png` from Google
Drive), `CacheContentFile` would pass the full path as a filename to
`GetTemporaryFile`. This caused `Java.IO.File(parent, child)` to either
ignore the parent directory (absolute paths) or attempt to create
non-existent nested directories (relative paths), leading to broken
cached file paths.

### Changes

- **Extract `EnsureFileName` helper** to `FileSystemUtils.shared.cs` —
normalizes a display name by stripping directory components, guarding
against empty/whitespace/reserved directory names (`.`, `..`), and
appending a known extension when the name has none
- **Add empty-string guard** — if `Path.GetFileName` returns empty (e.g.
trailing separator `somefolder/`), falls back to a generated GUID name
instead of creating a dot-file
- **Add 17 unit tests** covering: simple names, paths with directories,
null/empty/whitespace display names, trailing separators, `.`/`..`
reserved names, extension handling

### Tested scenarios

Verified on Android API 36 emulator using a custom `ContentProvider`
that injects problematic `DisplayName` values through the full
`EnsurePhysicalPath` → `CacheContentFile` → `EnsureFileName` pipeline:

| DisplayName | Result |
|---|---|
| `/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` |
`IMG_20240101.jpg` ✅ |
| `/sdcard/.transforms/synthetic/.../photo.png` | `photo.png` ✅ |
| `Download/Subdir/document.pdf` | `document.pdf` ✅ |
| `somefolder/` | `<guid>.jpg` ✅ |
| `""` (empty) | `<guid>.jpg` ✅ |
| `"   "` (whitespace) | `<guid>.jpg` ✅ |
| `screenshot_12345` (no ext) | `screenshot_12345.jpg` ✅ |
| `normal_photo.jpg` | `normal_photo.jpg` ✅ |

---------
Co-authored-by: Jakub Florkowski <kubaflo123@gmail.com>
PureWeen pushed a commit that referenced this pull request Apr 28, 2026
@mattleibow @kubaflo @PureWeen
Extract filename from DisplayName and add extension if missing (#35050)
c034df1 
<!-- Please let the below note in for people that find this PR -->
> [!NOTE]
> Are you waiting for the changes in this PR to be merged?
> It would be very helpful if you could [test the resulting
artifacts](https://github.com/dotnet/maui/wiki/Testing-PR-Builds) from
this PR and let us know in a comment if this change resolves your issue.
Thank you!

## Description

When Android content providers return a `DisplayName` that contains
directory components (e.g.
`/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` from Samsung, or
`/sdcard/.transforms/synthetic/picker/0/.../photo.png` from Google
Drive), `CacheContentFile` would pass the full path as a filename to
`GetTemporaryFile`. This caused `Java.IO.File(parent, child)` to either
ignore the parent directory (absolute paths) or attempt to create
non-existent nested directories (relative paths), leading to broken
cached file paths.

### Changes

- **Extract `EnsureFileName` helper** to `FileSystemUtils.shared.cs` —
normalizes a display name by stripping directory components, guarding
against empty/whitespace/reserved directory names (`.`, `..`), and
appending a known extension when the name has none
- **Add empty-string guard** — if `Path.GetFileName` returns empty (e.g.
trailing separator `somefolder/`), falls back to a generated GUID name
instead of creating a dot-file
- **Add 17 unit tests** covering: simple names, paths with directories,
null/empty/whitespace display names, trailing separators, `.`/`..`
reserved names, extension handling

### Tested scenarios

Verified on Android API 36 emulator using a custom `ContentProvider`
that injects problematic `DisplayName` values through the full
`EnsurePhysicalPath` → `CacheContentFile` → `EnsureFileName` pipeline:

| DisplayName | Result |
|---|---|
| `/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` |
`IMG_20240101.jpg` ✅ |
| `/sdcard/.transforms/synthetic/.../photo.png` | `photo.png` ✅ |
| `Download/Subdir/document.pdf` | `document.pdf` ✅ |
| `somefolder/` | `<guid>.jpg` ✅ |
| `""` (empty) | `<guid>.jpg` ✅ |
| `"   "` (whitespace) | `<guid>.jpg` ✅ |
| `screenshot_12345` (no ext) | `screenshot_12345.jpg` ✅ |
| `normal_photo.jpg` | `normal_photo.jpg` ✅ |

---------
Co-authored-by: Jakub Florkowski <kubaflo123@gmail.com>
PureWeen pushed a commit that referenced this pull request Apr 29, 2026
@mattleibow @kubaflo @github-actions
Extract filename from DisplayName and add extension if missing (#35050)
f784343 
<!-- Please let the below note in for people that find this PR -->
> [!NOTE]
> Are you waiting for the changes in this PR to be merged?
> It would be very helpful if you could [test the resulting
artifacts](https://github.com/dotnet/maui/wiki/Testing-PR-Builds) from
this PR and let us know in a comment if this change resolves your issue.
Thank you!

## Description

When Android content providers return a `DisplayName` that contains
directory components (e.g.
`/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` from Samsung, or
`/sdcard/.transforms/synthetic/picker/0/.../photo.png` from Google
Drive), `CacheContentFile` would pass the full path as a filename to
`GetTemporaryFile`. This caused `Java.IO.File(parent, child)` to either
ignore the parent directory (absolute paths) or attempt to create
non-existent nested directories (relative paths), leading to broken
cached file paths.

### Changes

- **Extract `EnsureFileName` helper** to `FileSystemUtils.shared.cs` —
normalizes a display name by stripping directory components, guarding
against empty/whitespace/reserved directory names (`.`, `..`), and
appending a known extension when the name has none
- **Add empty-string guard** — if `Path.GetFileName` returns empty (e.g.
trailing separator `somefolder/`), falls back to a generated GUID name
instead of creating a dot-file
- **Add 17 unit tests** covering: simple names, paths with directories,
null/empty/whitespace display names, trailing separators, `.`/`..`
reserved names, extension handling

### Tested scenarios

Verified on Android API 36 emulator using a custom `ContentProvider`
that injects problematic `DisplayName` values through the full
`EnsurePhysicalPath` → `CacheContentFile` → `EnsureFileName` pipeline:

| DisplayName | Result |
|---|---|
| `/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` |
`IMG_20240101.jpg` ✅ |
| `/sdcard/.transforms/synthetic/.../photo.png` | `photo.png` ✅ |
| `Download/Subdir/document.pdf` | `document.pdf` ✅ |
| `somefolder/` | `<guid>.jpg` ✅ |
| `""` (empty) | `<guid>.jpg` ✅ |
| `"   "` (whitespace) | `<guid>.jpg` ✅ |
| `screenshot_12345` (no ext) | `screenshot_12345.jpg` ✅ |
| `normal_photo.jpg` | `normal_photo.jpg` ✅ |

---------
Co-authored-by: Jakub Florkowski <kubaflo123@gmail.com>
github-actions Bot pushed a commit that referenced this pull request May 6, 2026
@mattleibow @kubaflo
Extract filename from DisplayName and add extension if missing (#35050)
22be650 
<!-- Please let the below note in for people that find this PR -->
> [!NOTE]
> Are you waiting for the changes in this PR to be merged?
> It would be very helpful if you could [test the resulting
artifacts](https://github.com/dotnet/maui/wiki/Testing-PR-Builds) from
this PR and let us know in a comment if this change resolves your issue.
Thank you!

## Description

When Android content providers return a `DisplayName` that contains
directory components (e.g.
`/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` from Samsung, or
`/sdcard/.transforms/synthetic/picker/0/.../photo.png` from Google
Drive), `CacheContentFile` would pass the full path as a filename to
`GetTemporaryFile`. This caused `Java.IO.File(parent, child)` to either
ignore the parent directory (absolute paths) or attempt to create
non-existent nested directories (relative paths), leading to broken
cached file paths.

### Changes

- **Extract `EnsureFileName` helper** to `FileSystemUtils.shared.cs` —
normalizes a display name by stripping directory components, guarding
against empty/whitespace/reserved directory names (`.`, `..`), and
appending a known extension when the name has none
- **Add empty-string guard** — if `Path.GetFileName` returns empty (e.g.
trailing separator `somefolder/`), falls back to a generated GUID name
instead of creating a dot-file
- **Add 17 unit tests** covering: simple names, paths with directories,
null/empty/whitespace display names, trailing separators, `.`/`..`
reserved names, extension handling

### Tested scenarios

Verified on Android API 36 emulator using a custom `ContentProvider`
that injects problematic `DisplayName` values through the full
`EnsurePhysicalPath` → `CacheContentFile` → `EnsureFileName` pipeline:

| DisplayName | Result |
|---|---|
| `/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` |
`IMG_20240101.jpg` ✅ |
| `/sdcard/.transforms/synthetic/.../photo.png` | `photo.png` ✅ |
| `Download/Subdir/document.pdf` | `document.pdf` ✅ |
| `somefolder/` | `<guid>.jpg` ✅ |
| `""` (empty) | `<guid>.jpg` ✅ |
| `"   "` (whitespace) | `<guid>.jpg` ✅ |
| `screenshot_12345` (no ext) | `screenshot_12345.jpg` ✅ |
| `normal_photo.jpg` | `normal_photo.jpg` ✅ |

---------
Co-authored-by: Jakub Florkowski <kubaflo123@gmail.com>
@kubaflo kubaflo added the s/agent-gate-passed AI verified tests catch the bug (fail without fix, pass with fix) label May 20, 2026
github-actions Bot pushed a commit that referenced this pull request May 25, 2026
@mattleibow @kubaflo @sheiksyedm
Extract filename from DisplayName and add extension if missing (#35050)
c28135b 
<!-- Please let the below note in for people that find this PR -->
> [!NOTE]
> Are you waiting for the changes in this PR to be merged?
> It would be very helpful if you could [test the resulting
artifacts](https://github.com/dotnet/maui/wiki/Testing-PR-Builds) from
this PR and let us know in a comment if this change resolves your issue.
Thank you!

## Description

When Android content providers return a `DisplayName` that contains
directory components (e.g.
`/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` from Samsung, or
`/sdcard/.transforms/synthetic/picker/0/.../photo.png` from Google
Drive), `CacheContentFile` would pass the full path as a filename to
`GetTemporaryFile`. This caused `Java.IO.File(parent, child)` to either
ignore the parent directory (absolute paths) or attempt to create
non-existent nested directories (relative paths), leading to broken
cached file paths.

### Changes

- **Extract `EnsureFileName` helper** to `FileSystemUtils.shared.cs` —
normalizes a display name by stripping directory components, guarding
against empty/whitespace/reserved directory names (`.`, `..`), and
appending a known extension when the name has none
- **Add empty-string guard** — if `Path.GetFileName` returns empty (e.g.
trailing separator `somefolder/`), falls back to a generated GUID name
instead of creating a dot-file
- **Add 17 unit tests** covering: simple names, paths with directories,
null/empty/whitespace display names, trailing separators, `.`/`..`
reserved names, extension handling

### Tested scenarios

Verified on Android API 36 emulator using a custom `ContentProvider`
that injects problematic `DisplayName` values through the full
`EnsurePhysicalPath` → `CacheContentFile` → `EnsureFileName` pipeline:

| DisplayName | Result |
|---|---|
| `/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` |
`IMG_20240101.jpg` ✅ |
| `/sdcard/.transforms/synthetic/.../photo.png` | `photo.png` ✅ |
| `Download/Subdir/document.pdf` | `document.pdf` ✅ |
| `somefolder/` | `<guid>.jpg` ✅ |
| `""` (empty) | `<guid>.jpg` ✅ |
| `"   "` (whitespace) | `<guid>.jpg` ✅ |
| `screenshot_12345` (no ext) | `screenshot_12345.jpg` ✅ |
| `normal_photo.jpg` | `normal_photo.jpg` ✅ |

---------
Co-authored-by: Jakub Florkowski <kubaflo123@gmail.com>
devanathan-vaithiyanathan pushed a commit to devanathan-vaithiyanathan/maui that referenced this pull request Jun 1, 2026
@mattleibow @kubaflo @devanathan-vaithiyanathan
Extract filename from DisplayName and add extension if missing (dotne…
21725d4 
…t#35050)

<!-- Please let the below note in for people that find this PR -->
> [!NOTE]
> Are you waiting for the changes in this PR to be merged?
> It would be very helpful if you could [test the resulting
artifacts](https://github.com/dotnet/maui/wiki/Testing-PR-Builds) from
this PR and let us know in a comment if this change resolves your issue.
Thank you!

## Description

When Android content providers return a `DisplayName` that contains
directory components (e.g.
`/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` from Samsung, or
`/sdcard/.transforms/synthetic/picker/0/.../photo.png` from Google
Drive), `CacheContentFile` would pass the full path as a filename to
`GetTemporaryFile`. This caused `Java.IO.File(parent, child)` to either
ignore the parent directory (absolute paths) or attempt to create
non-existent nested directories (relative paths), leading to broken
cached file paths.

### Changes

- **Extract `EnsureFileName` helper** to `FileSystemUtils.shared.cs` —
normalizes a display name by stripping directory components, guarding
against empty/whitespace/reserved directory names (`.`, `..`), and
appending a known extension when the name has none
- **Add empty-string guard** — if `Path.GetFileName` returns empty (e.g.
trailing separator `somefolder/`), falls back to a generated GUID name
instead of creating a dot-file
- **Add 17 unit tests** covering: simple names, paths with directories,
null/empty/whitespace display names, trailing separators, `.`/`..`
reserved names, extension handling

### Tested scenarios

Verified on Android API 36 emulator using a custom `ContentProvider`
that injects problematic `DisplayName` values through the full
`EnsurePhysicalPath` → `CacheContentFile` → `EnsureFileName` pipeline:

| DisplayName | Result |
|---|---|
| `/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` |
`IMG_20240101.jpg` ✅ |
| `/sdcard/.transforms/synthetic/.../photo.png` | `photo.png` ✅ |
| `Download/Subdir/document.pdf` | `document.pdf` ✅ |
| `somefolder/` | `<guid>.jpg` ✅ |
| `""` (empty) | `<guid>.jpg` ✅ |
| `"   "` (whitespace) | `<guid>.jpg` ✅ |
| `screenshot_12345` (no ext) | `screenshot_12345.jpg` ✅ |
| `normal_photo.jpg` | `normal_photo.jpg` ✅ |

---------
Co-authored-by: Jakub Florkowski <kubaflo123@gmail.com>
@PureWeen PureWeen mentioned this pull request Jun 2, 2026
Merged
PureWeen added a commit that referenced this pull request Jun 2, 2026
@PureWeen
May 4th, Candidate (#35234)
d386445 
## What's Coming

.NET MAUI inflight/candidate introduces significant improvements across
all platforms with focus on quality, performance, and developer
experience. This release includes 85 commits with various improvements,
bug fixes, and enhancements.


## Button
- [Android, iOS] Button: Fix VisualState properties not restored when
leaving custom state by @BagavathiPerumal in
#33346
  <details>
  <summary>🔧 Fixes</summary>

- [Button VisualStates do not
work](#19690)
  </details>

## CollectionView
- Fix CollectionView grid spacing updates for first row and column by
@KarthikRajaKalaimani in #34527
  <details>
  <summary>🔧 Fixes</summary>

- [[MAUI] I2_Vertical grid for horizontal Item Spacing and Vertical Item
Spacing - horizontally updating the spacing only applies to the second
column](#34257)
  </details>

- CarouselView: Fix cascading PositionChanged/CurrentItemChanged events
on collection update by @praveenkumarkarunanithi in
#31275
  <details>
  <summary>🔧 Fixes</summary>

- [[Windows] CurrentItemChangedEventArgs and PositionChangedEventArgs
Not Working Properly in
CarouselView](#29529)
  </details>

- [Windows] Fixed ItemSpacing doesn't work in Carousel View by
@SubhikshaSf4851 in #30014
  <details>
  <summary>🔧 Fixes</summary>

- [ItemSpacing on CarouselView is not applied on
Windows.](#29772)
  </details>

- Fix CollectionView not scrolling to top on iOS status bar tap by
@jfversluis in #34687
  <details>
  <summary>🔧 Fixes</summary>

- [[iOS] UICollectionView ScrollToTop does not
work](#19866)
  </details>

- [iOS] Fixed CollectionView Scroll Jitter for TextType HTML Labels by
@SubhikshaSf4851 in #34383
  <details>
  <summary>🔧 Fixes</summary>

- [CollectionView scrolling is jittery when ItemTemplate contains Label
with TextType="Html" in .NET
10](#33065)
  </details>

- Fix CollectionView Header is not visible when ItemsSource is not set
and an EmptyView is set in iOS, Mac platform by @KarthikRajaKalaimani in
#34989
  <details>
  <summary>🔧 Fixes</summary>

- [CollectionView Header is not visible when ItemsSource is not set and
EmptyView is set in iOS, Mac
platform](#34897)
  </details>

- [Android] Fix CollectionView EmptyView not displayed correctly by
@KarthikRajaKalaimani in #34956
  <details>
  <summary>🔧 Fixes</summary>

- [[Android] CollectionView - EmptyView not displayed
correctly](#34861)
  </details>

- [iOS] Fix CollectionView ScrollOffset not resetting when ItemsSource
changes by @SyedAbdulAzeemSF4852 in
#34488
  <details>
  <summary>🔧 Fixes</summary>

- [[IOS] CollectionView ScrollOffset does not reset when the ItemSource
is changed in iOS.](#26366)
- [Re-enable Issue7993 test on iOS/Catalyst - CollectionView scroll
position not reset when updating
ItemsSource](#33500)
  </details>

- [Revert] [iOS] Fixed CollectionView Scroll Jitter for TextType HTML
Labels by @SubhikshaSf4851 in #35341

## Core Lifecycle
- [Android] Fix NRE in ContainerView when Android Context is null during
lifecycle transition by @rmarinho in
#34901
  <details>
  <summary>🔧 Fixes</summary>

- [[Android] NullReferenceException in NavigationRootManager.Connect
when mapping Window
content](#34900)
  </details>

## DateTimePicker
- [Android] Fix for TimePicker Dialog doesn't update the layout when
rotating the device with dialog open by @HarishwaranVijayakumar in
#31910
  <details>
  <summary>🔧 Fixes</summary>

- [[Android] TimePicker Dialog doesn't update the layout when rotating
the device with dialog
open](#31658)
  </details>

- [Android, iOS] Fixed TimePicker FlowDirection Not Applied Across
Platforms by @Dhivya-SF4094 in #30369
  <details>
  <summary>🔧 Fixes</summary>

- [TimePicker FlowDirection Not Working on All
Platforms](#30192)
  </details>

- [Windows] Fixed TimePicker CharacterSpacing issue by @SubhikshaSf4851
in #30533
  <details>
  <summary>🔧 Fixes</summary>

- [[Windows] TimePicker CharacterSpacing Property Not Working on
Windows](#30199)
  </details>

- [MacCatalyst] Fix DatePicker Opened/Closed events not being raised by
@SubhikshaSf4851 in #34970
  <details>
  <summary>🔧 Fixes</summary>

- [[MacCatalyst] DatePicker Opened and Closed events are not raised on
Mac platform](#34848)
  </details>

## Dialogalert
- [Android] Fix AlertDialog, ActionSheet, and Prompt render with
Material 2 styles when Material 3 is enabled by @HarishwaranVijayakumar
in #35121
  <details>
  <summary>🔧 Fixes</summary>

- [[Android] AlertDialog, ActionSheet, and Prompt render with Material 2
styles when Material 3 is
enabled](#35119)
  </details>

## Docs
- docs: Add UITesting-Guide, ReleasePlanning, and ReleaseProcess to
docs/README.md index by @PureWeen in
#35195

- docs: Fix hardcoded path and add library overview in Essentials.AI
README by @PureWeen in #35194

- docs: Update branch reference from net10.0 to net11.0 in
DEVELOPMENT.md by @PureWeen in #35193

## Drawing
- Fix Path Rendering Issue Inside StackLayout When Margin Is Set by
@Shalini-Ashokan in #28071
  <details>
  <summary>🔧 Fixes</summary>

- [Path does not render if it has
Margin](#13801)
  </details>

- Fixed FlowDirection property not working on Drawable control and
GraphicsView by @Dhivya-SF4094 in
#34557
  <details>
  <summary>🔧 Fixes</summary>

- [[Android, Windows, iOS, macOS] FlowDirection property not working on
BoxView Control](#34402)
  </details>

- [iOS & Mac] Fix image tile misalignment in GraphicsView ImagePaint by
@SubhikshaSf4851 in #34935
  <details>
  <summary>🔧 Fixes</summary>

- [[iOS] Image resized with ResizeMode.Fit is not rendered correctly in
GraphicsView](#34755)
  </details>

- Fix Shadow does not honour Styles by @KarthikRajaKalaimani in
#35081
  <details>
  <summary>🔧 Fixes</summary>

- [Shadow does not honour
Styles](#19560)
  </details>

## Entry
- [iOS/macCatalyst] Fix Entry and Editor BackgroundColor reset when set
to null by @Shalini-Ashokan in #34741
  <details>
  <summary>🔧 Fixes</summary>

- [[iOS, Maccatalyst] Entry & Editor BackgroundColor not reset to
Null](#34611)
  </details>

- [Windows] Fix password Entry crash when setting text on empty field by
@praveenkumarkarunanithi in #33891
  <details>
  <summary>🔧 Fixes</summary>

- [[WinUI] Password Obfuscation causes unhandled
crash](#33334)
  </details>

## Essentials
- [Essentials] Use mean sea level altitude on Android API 34+ by
@KitKeen in #35097
  <details>
  <summary>🔧 Fixes</summary>

- [Add support for MslAltitudeMeters in Essentials Geolocation on
Android](#27554)
  </details>

## Flyout
- Fixed Flyout Not Displayed on Android When FlyoutWidth Is Set Only for
Desktop via OnIdiom by @NanthiniMahalingam in
#29028
  <details>
  <summary>🔧 Fixes</summary>

- [[Android] FlyoutWidth with OnIdiom shows no
flyout](#13243)
  </details>

- Revert "[Windows] Fix Flyout/Locked mode header collapse regression
causing UI test failures on candidate branch" by @kubaflo in
#35339

- Revert "Revert "[Windows] Fix Flyout/Locked mode header collapse
regression causing UI test failures on candidate branch"" by @kubaflo in
#35342

## Flyoutpage
- Fix [Android] Title of FlyOutPage is not updating anymore after
showing a NonFlyOutPage by @KarthikRajaKalaimani in
#34839
  <details>
  <summary>🔧 Fixes</summary>

- [[Android] Title of FlyOutPage is not updating anymore after showing a
NonFlyOutPage](#33615)
  </details>

## Label
- [iOS] Fix span Tap gesture on wrapped Label lines in iOS 26+ by
@SubhikshaSf4851 in #34640
  <details>
  <summary>🔧 Fixes</summary>

- [[iOS]Span TapGestureRecognizer does not work on the second line of
the span, if the span is wrapped to the next
line](#34504)
  </details>

## Layout
- Fixed Stacklayout is not rendered when clip is applied and StackLayout
placed child to the Border control in iOS/ Mac platform by
@KarthikRajaKalaimani in #33330
  <details>
  <summary>🔧 Fixes</summary>

- [[Mac/iOS] StackLayout fails to render content while applying Clip,
and the layout is placed inside a Border with Background in .NET
MAUI](#33241)
  </details>

## Map
- Fix Changing Location on a Pin does nothing by @NirmalKumarYuvaraj in
#30201
  <details>
  <summary>🔧 Fixes</summary>

- [[Maps] [Regression from Xamarin.Forms.Maps] Changing Location on a
Pin does nothing](#12916)
  </details>

## Mediapicker
- [iOS] Fix HEIC images picked via PickPhotosAsync not displayed by
@HarishwaranVijayakumar in #34954
  <details>
  <summary>🔧 Fixes</summary>

- [[iOS] [Regression] HEIC images picked via PickPhotosAsync not
displayed](#34953)
  </details>

- [Android] Fix MediaPicker.PickPhotosAsync UnauthorizedAccessException
on API 28 and below by @HarishwaranVijayakumar in
#34981
  <details>
  <summary>🔧 Fixes</summary>

- [MediaPicker.PickPhotos fails to modify image, tries to load original
source, fails to load source on Android
9.0](#34889)
  </details>

## Pages
- [iOS] Fix ContentPage with ToolbarItem Clicked event leaks when
presented as modal page by @devanathan-vaithiyanathan in
#35009
  <details>
  <summary>🔧 Fixes</summary>

- [ContentPage with ToolbarItem Clicked event leaks when presented as
modal page](#34892)
  </details>

## Platform
- [Android] Fix OnBackButtonPressed not invoked for Shell by
@Dhivya-SF4094 in #35150
  <details>
  <summary>🔧 Fixes</summary>

- [On Screen Back Button Does Not Fire OnBackButtonPressed in
Android](#9095)
  </details>

## RadioButton
- Fix RadioButtonGroup not working with ContentView by @Dhivya-SF4094 in
#34781
  <details>
  <summary>🔧 Fixes</summary>

- [RadioButtonGroup not working with
ContentView](#34759)
  </details>

- [Windows] Fix for RadioButton BorderColor and BorderWidth not updated
at runtime by @SyedAbdulAzeemSF4852 in
#28335
  <details>
  <summary>🔧 Fixes</summary>

- [RadioButton Border color not working for focused visual
state](#15806)
  </details>

- [iOS] Fix RadioButton BackgroundColor bleeding outside CornerRadius by
@SyedAbdulAzeemSF4852 in #34844
  <details>
  <summary>🔧 Fixes</summary>

- [[iOS] RadioButton BackgroundColor bleeds outside
CornerRadius](#34842)
  </details>

## SafeArea
- [iOS] Fix stale bottom safe area after changing SafeAreaEdges with
keyboard open by @praveenkumarkarunanithi in
#35083
  <details>
  <summary>🔧 Fixes</summary>

- [[iOS] ContentPage bottom has white space after changing SafeAreaEdges
while keyboard is open](#34846)
  </details>

## ScrollView
- [Windows] Fix Preserve ScrollView offsets when Orientation changes to
Neither by @SubhikshaSf4851 in #34827
  <details>
  <summary>🔧 Fixes</summary>

- [[Windows] ScrollView offsets do not preserve when Orientation changes
to Neither](#34671)
  </details>

## Searchbar
- [iOS] Fix SearchBar unexpected left margin in iPad windowed mode on 26
Version by @SubhikshaSf4851 in #34704
  <details>
  <summary>🔧 Fixes</summary>

- [in iPad windowed mode SearchBar adds left margin equivaltent to
SafeAreaInsets when placed inside
grid](#34551)
  </details>

## Shell
- [Windows] Fix for Shell.FlyoutBehavior="Flyout" forces the title
height space above the tab bar even if the page title is empty by
@BagavathiPerumal in #30382
  <details>
  <summary>🔧 Fixes</summary>

- [(Windows) Shell.FlyoutBehavior="Flyout" forces the title height space
above the tab bar even if the page title is
empty](#30254)
  </details>

- Fix Shell flyout items scrolling behind FlyoutHeader on iOS by @Qythyx
in #34936
  <details>
  <summary>🔧 Fixes</summary>

- [Shell flyout items scroll behind FlyoutHeader on
iOS](#34925)
  </details>

- [iOS, Mac] Fix Shell.CurrentState.Location stale in OnNavigated after
GoToAsync by @Vignesh-SF3580 in
#34880
  <details>
  <summary>🔧 Fixes</summary>

- [Shell.OnNavigated not called for route
navigation](#34662)
  </details>

- [iOS26]Fix
BackButtonBehavior_IsEnabled_False_BackButtonDoesNotNavigate UITest
fails by @devanathan-vaithiyanathan in
#34890
  <details>
  <summary>🔧 Fixes</summary>

- [[iOS 26] BackButtonBehavior_IsEnabled_False_BackButtonDoesNotNavigate
test fails with
TimeoutException](#34771)
  </details>

- [iOS] Fix Shell page memory leak when using TitleView with x:Name by
@Shalini-Ashokan in #35082
  <details>
  <summary>🔧 Fixes</summary>

- [[iOS] Title view memory
leak](#34975)
  </details>

- [Material 3] Fix Material 2 color flash in AppBar when switching tabs
for the first time by @Dhivya-SF4094 in
#35117
  <details>
  <summary>🔧 Fixes</summary>

- [Material 3: AppBar briefly displays Material 2 colors when switching
tabs for the first time](#35116)
  </details>

- [Android] Fix Shell/TabbedPage "More" BottomSheet uses hard-coded M2
colors when Material3 is enabled by @HarishwaranVijayakumar in
#35129
  <details>
  <summary>🔧 Fixes</summary>

- [[Android] Shell/TabbedPage "More" BottomSheet uses hard-coded M2
colors when Material3 is
enabled](#35127)
  </details>

- [Android] Shell: Fix top-tab unselected text visibility in Material 3
light theme by @SyedAbdulAzeemSF4852 in
#35128
  <details>
  <summary>🔧 Fixes</summary>

- [[Android] Shell top-tab unselected text appears too faint in Material
3 light theme](#35125)
  </details>

- Fix Shell.Items.Clear() memory leak by disconnecting child handlers on
removal (#34898) by @Shalini-Ashokan in
#35031
  <details>
  <summary>🔧 Fixes</summary>

- [Shell.Items.Clear() does not disconnect handlers
correctly](#34898)
  </details>

- [iOS&Mac] Fix Shell SearchHandler Query update on Initial load by
@SubhikshaSf4851 in #35008
  <details>
  <summary>🔧 Fixes</summary>

- [[iOS&Mac] Shell SearchHandler Query not shown in search bar on
initial load](#35005)
  </details>

## SwipeView
- [iOS,MacCatalyst] Fix for SwipeView.Open() throwing an
ArgumentException on the second programmatic call by @BagavathiPerumal
in #34982
  <details>
  <summary>🔧 Fixes</summary>

- [[net 11.0][iOS,MacCatalyst] SwipeView.Open() throws ArgumentException
on second programmatic
call](#34917)
  </details>

- [Android/iOS] Fix SwipeItem visibility change causing double command
execution in Execute mode by @praveenkumarkarunanithi in
#35087
  <details>
  <summary>🔧 Fixes</summary>

- [Changing visibility on an SwipeItem causes multiple items to be
executed](#7580)
  </details>

## Switch
- [iOS] Fix Switch ThumbColor reset on iOS 26+ theme changes. by
@Shalini-Ashokan in #33953
  <details>
  <summary>🔧 Fixes</summary>

- [Switch ThumbColor not Initialized Using VisualStateManager on iOS
Device](#33783)
- [I9-On macOS 26.2, the "Animate scroll" button is white by default on
iOS and Maccatalyst
platforms.](#33767)
  </details>

## TabbedPage
- [Windows] TabbedPage: Refresh layout when NavigationView size changes
by @BagavathiPerumal in #26217
  <details>
  <summary>🔧 Fixes</summary>

- [TabbedPage - ScrollView not allowing scrolling when it
should](#26103)
- [TabbedPage App on resize hides page bottom
content](#11402)
- [Grid overflows child ContentPage of parent TabbedPage on initial load
and when resizing on
Windows](#20028)
  </details>

- [Android] Material 3 Fixed BottomNavigationView overflowing in Tabbed
page by @NirmalKumarYuvaraj in #35064
  <details>
  <summary>🔧 Fixes</summary>

- [[Android] Material3 - TabbedPage bottom tabs overflowing the
contents](#35063)
  </details>

- [Windows] Fix for Multiple Tabs Being Selected in WinUI TabbedPage by
@SyedAbdulAzeemSF4852 in #33312
  <details>
  <summary>🔧 Fixes</summary>

- [WinUI TabbedPage can have multiple tabs
selected](#31799)
  </details>

## Theming
- [iOS] Fix StaticResource Hot Reload crash on iOS by @StephaneDelcroix
in #35020
  <details>
  <summary>🔧 Fixes</summary>

- [The maui app quit and no errors in error list after editing
ResourceDictionary XAML file on iOS Simulator with MAUI SR6
10.0.60](#35018)
  </details>

## Toolbar
- [Windows] Fix for CS1061 build error caused by missing
HasMenuBarContent property in MauiToolbar by @BagavathiPerumal in
#35040

## Tooling
- Fix VisualStateGroups duplicate name crash with implicit styles
(#34716) by @StephaneDelcroix in
#34719
  <details>
  <summary>🔧 Fixes</summary>

- [SourceGen: VisualStateManager.VisualStateGroups causes 'Names must be
unique' at startup](#34716)
  </details>

## WebView
- Refactor the HybridWebView and properly support complex parameters by
@mattleibow in #32491

- [Android] Fix WebView scrolling inside ScrollView by @Shalini-Ashokan
in #33133
  <details>
  <summary>🔧 Fixes</summary>

- [[Android] WebView's content does not scroll when placed inside a
ScrollView](#32971)
  </details>


<details>
<summary>🔧 Infrastructure (1)</summary>

- [Windows] Fix Narrator announcing ContentView children twice when
Description is set by @praveenkumarkarunanithi in
#33979
  <details>
  <summary>🔧 Fixes</summary>

- [[Windows] SemanticProperties.Description announced twice when set on
focusable container cell (Label
inside)](#33373)
  </details>

</details>

<details>
<summary>🧪 Testing (14)</summary>

- [Testing] SafeArea Feature Matrix Test Cases for ContentPage by
@TamilarasanSF4853 in #34877
- [Windows] Fix CollectionView ScrollTo related test cases failed in CI
by @HarishwaranVijayakumar in #34907
  <details>
  <summary>🔧 Fixes</summary>

- [[Testing][Windows]CollectionView ScrollTo related test cases failed
in CI](#34772)
  </details>
- [Testing] Fixed Build error on inflight/ candidate PR 35234 by
@HarishKumarSF4517 in #35241
- Fix CI for
ValidateKeyboardRuntime_SwitchContainerToSoftInput_WhileKeyboardOpen
test failure in May 4th Candidate by @devanathan-vaithiyanathan in
#35307
- [Windows] Fix Flyout/Locked mode header collapse regression causing UI
test failures on candidate branch by @BagavathiPerumal in
#35312
- [iOS/macCatalyst] [Candidate Fix] Editor shadow and theme regression
caused by BackgroundColor reset on initial handler connection by
@Shalini-Ashokan in #35343
- [Testing] Fixed UI test image failure in PR 35234 - [30/03/2026]
Candidate - 1 by @HarishKumarSF4517 in
#35325
- [iOS] Fix ShellFeatureMatrix test failures on candidate branch by
@Vignesh-SF3580 in #35346
- [Windows] Fix Issue29529VerifyPreviousPositionOnInsert test failure on
candidate branch by @praveenkumarkarunanithi in
#35398
- [Android] [Candidate Fix] Shell: Fix handler disconnect timing to
preserve WebView navigation and memory leak fix by @Shalini-Ashokan in
#35417
- [Testing]Revert 'Fix Preserve ScrollView offsets when Orientation
changes to Neither' by @TamilarasanSF4853 in
#35412
- [Windows] Fix VerifyAllIndicatorDotsShowShadowsWhenIndicatorSize test
failure on candidate branch by @praveenkumarkarunanithi in
#35458
- [Testing] Fixed test failure in PR 35234 - [05/08/2026] Candidate by
@TamilarasanSF4853 in #35362
- [Testing] Fixed test failure in PR 35234 - [05/04/2026] Candidate - 3
by @TamilarasanSF4853 in #35639

</details>

<details>
<summary>📦 Other (6)</summary>

- [UIKit] Avoid useless measure invalidation propagation cycles by
@albyrock87 in #33459
- BindableObject property access micro-optimizations by @albyrock87 in
#33584
- Extract filename from DisplayName and add extension if missing by
@mattleibow in #35050
- [core] Add keyed-DI screenshot extensibility for 3rd-party platform
backends by @Redth in #35096
  <details>
  <summary>🔧 Fixes</summary>

- [`ViewExtensions.CaptureAsync(IView)` and `IPlatformScreenshot` need
extensibility for third-party platform
backends](#34266)
  </details>
- Fix MainThread throwing on custom platform backends by @Redth in
#35070
  <details>
  <summary>🔧 Fixes</summary>

- [`MainThread.BeginInvokeOnMainThread` throws on custom platform
backends - Common UI-thread marshaling pattern crashes; `Dispatcher`
works but isn't the documented/recommended
path](#34101)
  </details>
- Tests: Add 11 missing UnitConverters unit tests by @PureWeen in
#35191

</details>

<details>
<summary>📝 Issue References</summary>

Fixes #7580, Fixes #9095, Fixes #11402, Fixes #12916, Fixes #13243,
Fixes #13801, Fixes #15806, Fixes #19560, Fixes #19690, Fixes #19866,
Fixes #20028, Fixes #26103, Fixes #26366, Fixes #27554, Fixes #29529,
Fixes #29772, Fixes #30192, Fixes #30199, Fixes #30254, Fixes #31658,
Fixes #31799, Fixes #32971, Fixes #33065, Fixes #33241, Fixes #33334,
Fixes #33373, Fixes #33500, Fixes #33615, Fixes #33767, Fixes #33783,
Fixes #34101, Fixes #34257, Fixes #34266, Fixes #34402, Fixes #34504,
Fixes #34551, Fixes #34611, Fixes #34662, Fixes #34671, Fixes #34716,
Fixes #34755, Fixes #34759, Fixes #34771, Fixes #34772, Fixes #34842,
Fixes #34846, Fixes #34848, Fixes #34861, Fixes #34889, Fixes #34892,
Fixes #34897, Fixes #34898, Fixes #34900, Fixes #34917, Fixes #34925,
Fixes #34953, Fixes #34975, Fixes #35005, Fixes #35018, Fixes #35063,
Fixes #35116, Fixes #35119, Fixes #35125, Fixes #35127

</details>

**Full Changelog**:
main...inflight/candidate
@PureWeen PureWeen modified the milestones: .NET 10 SR7, .NET 10 SR8 Jun 11, 2026
Shalini-Ashokan pushed a commit to Shalini-Ashokan/maui that referenced this pull request Jun 15, 2026
@mattleibow @kubaflo @Shalini-Ashokan
Extract filename from DisplayName and add extension if missing (dotne…
20ab7ee 
…t#35050)

<!-- Please let the below note in for people that find this PR -->
> [!NOTE]
> Are you waiting for the changes in this PR to be merged?
> It would be very helpful if you could [test the resulting
artifacts](https://github.com/dotnet/maui/wiki/Testing-PR-Builds) from
this PR and let us know in a comment if this change resolves your issue.
Thank you!

## Description

When Android content providers return a `DisplayName` that contains
directory components (e.g.
`/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` from Samsung, or
`/sdcard/.transforms/synthetic/picker/0/.../photo.png` from Google
Drive), `CacheContentFile` would pass the full path as a filename to
`GetTemporaryFile`. This caused `Java.IO.File(parent, child)` to either
ignore the parent directory (absolute paths) or attempt to create
non-existent nested directories (relative paths), leading to broken
cached file paths.

### Changes

- **Extract `EnsureFileName` helper** to `FileSystemUtils.shared.cs` —
normalizes a display name by stripping directory components, guarding
against empty/whitespace/reserved directory names (`.`, `..`), and
appending a known extension when the name has none
- **Add empty-string guard** — if `Path.GetFileName` returns empty (e.g.
trailing separator `somefolder/`), falls back to a generated GUID name
instead of creating a dot-file
- **Add 17 unit tests** covering: simple names, paths with directories,
null/empty/whitespace display names, trailing separators, `.`/`..`
reserved names, extension handling

### Tested scenarios

Verified on Android API 36 emulator using a custom `ContentProvider`
that injects problematic `DisplayName` values through the full
`EnsurePhysicalPath` → `CacheContentFile` → `EnsureFileName` pipeline:

| DisplayName | Result |
|---|---|
| `/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` |
`IMG_20240101.jpg` ✅ |
| `/sdcard/.transforms/synthetic/.../photo.png` | `photo.png` ✅ |
| `Download/Subdir/document.pdf` | `document.pdf` ✅ |
| `somefolder/` | `<guid>.jpg` ✅ |
| `""` (empty) | `<guid>.jpg` ✅ |
| `"   "` (whitespace) | `<guid>.jpg` ✅ |
| `screenshot_12345` (no ext) | `screenshot_12345.jpg` ✅ |
| `normal_photo.jpg` | `normal_photo.jpg` ✅ |

---------
Co-authored-by: Jakub Florkowski <kubaflo123@gmail.com>
Dhivya-SF4094 pushed a commit to Dhivya-SF4094/maui that referenced this pull request Jun 15, 2026
@mattleibow @kubaflo @Dhivya-SF4094
Extract filename from DisplayName and add extension if missing (dotne…
0412d6c 
…t#35050)

<!-- Please let the below note in for people that find this PR -->
> [!NOTE]
> Are you waiting for the changes in this PR to be merged?
> It would be very helpful if you could [test the resulting
artifacts](https://github.com/dotnet/maui/wiki/Testing-PR-Builds) from
this PR and let us know in a comment if this change resolves your issue.
Thank you!

## Description

When Android content providers return a `DisplayName` that contains
directory components (e.g.
`/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` from Samsung, or
`/sdcard/.transforms/synthetic/picker/0/.../photo.png` from Google
Drive), `CacheContentFile` would pass the full path as a filename to
`GetTemporaryFile`. This caused `Java.IO.File(parent, child)` to either
ignore the parent directory (absolute paths) or attempt to create
non-existent nested directories (relative paths), leading to broken
cached file paths.

### Changes

- **Extract `EnsureFileName` helper** to `FileSystemUtils.shared.cs` —
normalizes a display name by stripping directory components, guarding
against empty/whitespace/reserved directory names (`.`, `..`), and
appending a known extension when the name has none
- **Add empty-string guard** — if `Path.GetFileName` returns empty (e.g.
trailing separator `somefolder/`), falls back to a generated GUID name
instead of creating a dot-file
- **Add 17 unit tests** covering: simple names, paths with directories,
null/empty/whitespace display names, trailing separators, `.`/`..`
reserved names, extension handling

### Tested scenarios

Verified on Android API 36 emulator using a custom `ContentProvider`
that injects problematic `DisplayName` values through the full
`EnsurePhysicalPath` → `CacheContentFile` → `EnsureFileName` pipeline:

| DisplayName | Result |
|---|---|
| `/storage/emulated/0/DCIM/Camera/IMG_20240101.jpg` |
`IMG_20240101.jpg` ✅ |
| `/sdcard/.transforms/synthetic/.../photo.png` | `photo.png` ✅ |
| `Download/Subdir/document.pdf` | `document.pdf` ✅ |
| `somefolder/` | `<guid>.jpg` ✅ |
| `""` (empty) | `<guid>.jpg` ✅ |
| `"   "` (whitespace) | `<guid>.jpg` ✅ |
| `screenshot_12345` (no ext) | `screenshot_12345.jpg` ✅ |
| `normal_photo.jpg` | `normal_photo.jpg` ✅ |

---------
Co-authored-by: Jakub Florkowski <kubaflo123@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

Copilot code review Copilot Copilot left review comments

@kubaflo kubaflo kubaflo approved these changes

Assignees

No one assigned

Labels

s/agent-approved AI agent recommends approval - PR fix is correct and optimal s/agent-fix-pr-picked AI could not beat the PR fix - PR is the best among all candidates s/agent-gate-passed AI verified tests catch the bug (fail without fix, pass with fix) s/agent-reviewed PR was reviewed by AI agent workflow (full 4-phase review)

Projects

None yet

Milestone

.NET 10 SR8

Development

Successfully merging this pull request may close these issues.

5 participants

@mattleibow @MauiBot @kubaflo @PureWeen