ILSpy

9745 Commits

54 Branches

94 Tags

91 MiB

Tree: fa464968e1

Commit Graph

Author	SHA1	Message	Date
Siegfried Pammer	fa464968e1	Reorder the tree rows when the assembly list is sorted Sorting reorders the assembly list in place and reports it as a Move, which nothing downstream could act on: the tree node's handler had cases for Add, Remove and Reset only, and neither the child collection nor the flattener had a move at all. The rows therefore kept their pre-sort order until something else forced a rebuild. Moving the node rather than removing and re-inserting it keeps its identity, so an expanded subtree stays expanded and its row is not rebuilt. A run longer than one row has to be reported the way the consumer reads it: Avalonia's VirtualizingStackPanel applies a ranged move by removing OldItems.Count rows and re-inserting them at NewStartingIndex - (Count - 1), so a run reported by its final start index lands short by its own length. For a single row - a collapsed node, and every move a sort makes - the two readings coincide. Assisted-by: Claude:claude-opus-5[1m]:Claude Code	2 days ago
Siegfried Pammer	228041085b	Fix #3290 : marshal EnsureLazyChildren onto the tree's owning thread The crash is a NullReferenceException in GetNodeByVisibleIndex, reached when a background decompile realizes a node's children while the UI thread is indexing the flattener. Eight ILSpyTreeNode.Decompile overrides call EnsureLazyChildren from that task; two wrapped it in Dispatcher.UIThread.Invoke, six did not, and one of the two lost its wrapper in the Avalonia port with no test noticing for a release cycle. A rule every call site has to remember is a rule that gets broken again, so EnsureLazyChildren marshals itself instead: SetOwner already named the owning thread, and now also carries the host's way onto it. A call already on the owner runs inline, so a blocking invoke cannot deadlock on itself and a nested load costs no further hop; an unowned tree is left unmarshalled, which keeps building a subtree on a worker and publishing it on the UI thread legal. The affinity check stays as the regression detector, but its fail-fast throw was worthless on its own: tree mutation happens inside callers that catch Exception, so the throw ended up rendered into the decompiled output and the run passed. The violation is now recorded before the throw, and an assembly-level NUnit test action fails the test that produced one - an assembly-level teardown failure is reported but leaves the exit code at zero. Assisted-by: Claude:claude-opus-5:Claude Code	2 weeks ago

Author

SHA1

Message

Date

Siegfried Pammer

fa464968e1

Reorder the tree rows when the assembly list is sorted

Sorting reorders the assembly list in place and reports it as a Move, which
nothing downstream could act on: the tree node's handler had cases for Add,
Remove and Reset only, and neither the child collection nor the flattener had a
move at all. The rows therefore kept their pre-sort order until something else
forced a rebuild. Moving the node rather than removing and re-inserting it keeps
its identity, so an expanded subtree stays expanded and its row is not rebuilt.

A run longer than one row has to be reported the way the consumer reads it:
Avalonia's VirtualizingStackPanel applies a ranged move by removing OldItems.Count
rows and re-inserting them at NewStartingIndex - (Count - 1), so a run reported by
its final start index lands short by its own length. For a single row - a
collapsed node, and every move a sort makes - the two readings coincide.

Assisted-by: Claude:claude-opus-5[1m]:Claude Code

2 days ago

Siegfried Pammer

228041085b

Fix #3290 : marshal EnsureLazyChildren onto the tree's owning thread

The crash is a NullReferenceException in GetNodeByVisibleIndex, reached when a
background decompile realizes a node's children while the UI thread is indexing
the flattener. Eight ILSpyTreeNode.Decompile overrides call EnsureLazyChildren
from that task; two wrapped it in Dispatcher.UIThread.Invoke, six did not, and
one of the two lost its wrapper in the Avalonia port with no test noticing for a
release cycle. A rule every call site has to remember is a rule that gets broken
again, so EnsureLazyChildren marshals itself instead: SetOwner already named the
owning thread, and now also carries the host's way onto it. A call already on the
owner runs inline, so a blocking invoke cannot deadlock on itself and a nested
load costs no further hop; an unowned tree is left unmarshalled, which keeps
building a subtree on a worker and publishing it on the UI thread legal.

The affinity check stays as the regression detector, but its fail-fast throw was
worthless on its own: tree mutation happens inside callers that catch Exception,
so the throw ended up rendered into the decompiled output and the run passed. The
violation is now recorded before the throw, and an assembly-level NUnit test
action fails the test that produced one - an assembly-level teardown failure is
reported but leaves the exit code at zero.

Assisted-by: Claude:claude-opus-5:Claude Code

2 weeks ago

2 Commits (fa464968e1bb7ab84939d31d4dcd73eff3f44588)