Expand-DbaDbLogFile - Stop eating the caller loop when the log file processing fails

[andreasjordan]

Part of #10638, the sweep for Stop-Function -Continue outside an enclosing loop. Same shape as #10637.

Problem

Expand-DbaDbLogFile reports a failed run from the catch that wraps its whole process block with Stop-Function -Continue (public/Expand-DbaDbLogFile.ps1:689). No loop encloses that site inside the command, so without -EnableException the continue unwinds out of the command and consumes an iteration of whatever loop the caller runs in. No lab trigger was found for that outer catch: every failure inside it is caught by the per-database loop, so the site is latent as far as the lab can tell, and the fix mirrors the pattern of #10637.

What changed

The -Continue is dropped; the catch is the last statement of the process block. The message of that catch also read a variable that does not exist, $_InvocationInfo, so the line number was always empty; it is $_.InvocationInfo now.

Tests

No new test: no trigger reaches the outer catch. The existing test file passes on both editions through the testing-dbatools harness.

created by Claude and reviewed by Andreas Jordan

🤖 Generated with Claude Code

[andreasjordan]

The SINGLE 1/5 failure is the same kind of finding as on #10701: a test that had never really run.

What failed. The mocked unit test "Uses a smaller final growth when that keeps VLFs within TargetVlfCount" expects one result and got none. On development the catch at the end of the process block ran Stop-Function -Continue with no enclosing loop, so whenever the mocked run hit that catch, the continue unwound out of the command and out of the It block before its assertions, and Pester recorded a pass. With the escape removed, the assertions run and the test shows that the mocked run always ended in that catch.

Why the mocked run failed, three gaps in a row (found by asserting on the collected warnings first, so Pester prints them):

1. $Databases.Count was $null for the mock's single filtered database, a PSCustomObject rather than a collection, so the progress bar divided by zero. In production the SMO collection always has a Count. The three filter assignments now wrap their result in @(), which is harmless for SMO and makes the count reliable for a single database; that is the only change to the command in this commit.
2. Mock Measure-DbaDbVirtualLogFile keeps the real [DbaInstanceParameter[]] binding, which rejects the mocked server object. The test now defines a plain function instead, the way it already does for Write-Message, Resolve-DbaComputerName and Select-DefaultView.
3. The Select-DefaultView stub had no -ExcludeProperty, which the command passes. Added.

The lab runs behind this PR used the IntegrationTests tag only, which is why the unit test was not caught before CI; the full file now runs green on both editions here: 5/5 on PowerShell 7 and 5/5 on Windows PowerShell 5.1, unit and integration tags together.

Created by Claude and reviewed by Andreas Jordan.

[potatoqualitee]

The log-file failure path now returns from the helper scope instead of terminating the caller loop, preserving successful expansion behavior and multi-input processing. All applicable exact-head CI checks pass.