docs(sidebar): mark completed issues and reorder them

1 files changed, 35 insertions, 0 deletions

diff --git a/docs/dev/pages/issues/_remove-with-dispatcher.md b/docs/dev/pages/issues/_remove-with-dispatcher.md
new file mode 100644
index 0000000..b0d972f
--- /dev/null
+++ b/docs/dev/pages/issues/_remove-with-dispatcher.md
@@ -0,0 +1,35 @@
+<h1 align="center">[T0] Remove with_dispatcher and with_dispatchers</h1>
+<p align="center">
+    Breaking: make <code>Dispatcher</code> registration compile-time collected in all modes
+</p>
+
+> [!NOTE]
+>
+> This is a **Breaking Change** planned for Mingling 0.5.0.
+
+## Background
+
+Mingling's commands must be registered through `with_dispatcher` in order to be usable when `dispatcher_tree` is disabled. This has always been a strange semantic: `chain`, `renderer`, `help`, `completion`, and `metadata` are all collected at compile time, so why is `Dispatcher` the exception?
+
+In fact, during usage of `dispatcher` from 0.1.0 to 0.4.0, no scenario has ever been encountered where **dynamic registration** was necessary. It is considered unnecessary.
+
+## Plan
+
+Make `Dispatcher` registration also compile-time collected in non-`dispatcher_tree` states starting from 0.5.0, and remove `with_dispatcher` / `with_dispatchers` (and the related `#[program_setup]` registration path if it becomes obsolete).
+
+## Tasks
+
+- [x] Design how dispatchers are collected at compile time when `dispatcher_tree` is disabled (consistent with how `chain` / `renderer` / `completion` / `metadata` are collected)
+- [x] Remove `with_dispatcher` and `with_dispatchers` from the `Program` API
+- [x] Update `gen_program!` and the macros so registration happens automatically
+- [x] Migrate examples, tests, and docs that call `with_dispatcher` / `with_dispatchers`
+- [x] Verify both `dispatcher_tree`-enabled and disabled modes behave identically
+
+## 🕘 Progress
+
+- [x] In Progress
+- [x] Complete
+
+<p align="center" style="font-size: 0.85em; color: gray;">
+    Written by @Weicao-CatilGrass
+</p>