Understanding Hot Reload Limits logo

Understanding Hot Reload Limits

Community
skydoves
understanding-hot-reload-limits

Use this skill to teach Claude exactly which Kotlin and Compose changes hot-reload under Compose HotSwan and which trigger a full incremental rebuild fallback. Root cause is Android Runtime (ART) class schema immutability; only method bodies are mutable at runtime, so any change to fields, signatures, constructors, interfaces, inline functions, or new resource ids forces a rebuild. Covers the supported-changes table, the rebuild-forcing list, the diff-then-batch workflow that keeps a hot-reload session inside the fast path, and the inline-function and new-resource-id pitfalls. Trigger when the user asks "why did this rebuild?", "why isn't this hot reloading?", wants to learn HotSwan's boundaries before adopting it, or when reviewing a refactor that risks pushing a hot-reload session into a full rebuild.

Overview

Publisherskydoves
Repositorycompose-performance-skills
Skill nameunderstanding-hot-reload-limits
Stars
515
Forks
19
Bundled files
Instructions only
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by skydoves on GitHub. Read the source before you install it.

Installation

Install the Understanding Hot Reload Limits AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/skydoves/compose-performance-skills/tree/main/hot-reload/understanding-hot-reload-limits
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/skydoves/compose-performance-skills.git /tmp/compose-performance-skills
mkdir -p .claude/skills
cp -r /tmp/compose-performance-skills/hot-reload/understanding-hot-reload-limits .claude/skills/understanding-hot-reload-limits
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Understanding Hot Reload Limits in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Understanding Hot Reload Limits on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Understanding Hot Reload Limits is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Understanding Hot Reload Limits: stay inside the fast path

Android Runtime (ART) enforces strict rules for class redefinition: a redefined class must have an identical schema (fields, method signatures, interfaces) as the previous version. Only method bodies are mutable at runtime. HotSwan automatically detects schema changes and falls back to a full incremental build, so failures are not silent. Knowing the boundary in advance is what keeps an iteration loop sub-second instead of multi-second.

When to use this skill

  • The developer asks "why did this trigger a full rebuild?", "why isn't this hot reloading?", or "what does HotSwan support?".
  • A PR review surfaces a refactor (parameter change, constructor change, interface extraction, new resource id) that would push a hot-reload session into a rebuild.
  • The developer is planning a session and wants to order edits so the slow ones happen at the end.
  • The developer reports an inline fun change that "didn't reload" and is debugging.

When NOT to use this skill

  • Setup or first-run troubleshooting. See ../setting-up-compose-hotswan/SKILL.md.
  • Pure state-preservation question (the reload happened but state was lost). See the state-preservation sibling skill.
  • Configuration of the AI iteration loop or MCP server. See the AI-loop sibling skill.

Prerequisites

  • HotSwan installed and verified per ../setting-up-compose-hotswan/SKILL.md (tool window status WATCHING, body-only edit reloads in under one second).
  • Familiarity with the Kotlin compiler's distinction between method-body changes and class-schema changes.

Boundary tables

Hot-reloadable (no rebuild)

ChangeNotes
Composable function bodytext, colors, modifiers, layout, control flow inside the function
Non-composable function bodyViewModel methods, mappers, utilities, repositories
Adding a new composablesame file or new file
Reordering composablesHotSwan 1.2.0+
Resource value changesstrings.xml value, colors.xml value, dimens.xml value
Extension functionsincluding suspend, including vararg
Adding data class propertiesAPI 30+ only
Numeric, string, float literal patchescompiled separately for fastest reload

Forces full rebuild (fallback)

ChangeReason
Adding or removing function parametersmethod signature changes
Constructor changesparameter list, default values, init blocks affecting fields
Interface or superclass changesclass hierarchy is part of the schema
Adding new resource idsnew R.string, R.drawable, R.id entries require generated R class regeneration
Inline functionsexpanded at every call site; no discrete unit to swap
Lambda count change inside a functioninternal lambda class renumbering
Removing a previously defined functionmethod table shrinks; schema changes
Adding data class properties below API 30constructor schema change without ART support

Workflow when a change does not hot-reload

  1. Watch the HotSwan tool window status at the moment of save. A schema change surfaces as a "falling back to incremental build" status. That is the explicit signal, not a silent failure.
  2. Read the change diff. Decide: is this a body-only change, or a schema change (signature, constructor, interface, field, new resource id, inline body, lambda count)?
  3. If a refactor must change a signature, batch it. First make the body-only changes that get the screen to a working visual state and hot-reload them. Then, in a separate save, change the signature and accept the rebuild as the last step of the session.
  4. For inline functions, drop inline for the duration of iteration. Refactor the function out of inline while iterating, hot-reload as needed, and restore inline before commit. If iteration is rare or the perf characteristic must be preserved, accept the rebuild.
  5. For new resource ids, batch all id additions first, accept one rebuild, then iterate on values. Editing <string name="title">Updated Title</string> (existing id) hot-reloads; adding <string name="brand_new_id">...</string> does not.

Patterns

Pattern: function-body change (fast path) vs signature change (rebuild)

kotlin
// RIGHT (body-only change, hot-reloads)
@Composable
fun Greeting() {
    Text("Hello, World", color = Color.Blue)
}
kotlin
// WRONG for the fast path (signature change, forces rebuild)
@Composable
fun Greeting(name: String) {
    Text("Hello, $name")
}
// WRONG because: adding a parameter changes the method signature, which violates ART's class
// schema constraint. To stay inside the fast path, introduce the parameter with a default value
// in one save (still a rebuild; accept it once), then iterate on the body across subsequent saves.

Pattern: adding a new composable (fast path) vs changing a constructor (rebuild)

kotlin
// RIGHT (new composable in the same file, hot-reloads on HotSwan 1.2.0+)
@Composable
fun SecondaryAction() {
    Text("New")
}
kotlin
// WRONG for the fast path (constructor change)
data class User(val id: Long, val name: String, val age: Int)
// WRONG because: adding `age` extends the constructor. On API 30+ HotSwan supports adding
// data class properties; below API 30 the schema change forces a rebuild. When targeting older
// minSdk, batch property additions and accept a rebuild for each.

Pattern: inline function body change (rebuild)

kotlin
// WRONG for the fast path
inline fun <T> Modifier.observer(value: T, body: (T) -> Modifier): Modifier = body(value)
// WRONG because: inline functions are expanded at every call site at compile time. There is no
// discrete unit to swap, so editing the inline body forces a full rebuild of every call site.
// Drop `inline` for the duration of iteration, or accept the rebuild.
kotlin
// RIGHT for the fast path (non-inline variant during iteration)
fun <T> Modifier.observer(value: T, body: (T) -> Modifier): Modifier = body(value)

Pattern: resource value change (fast path) vs new resource id (rebuild)

xml
<!-- RIGHT: value change on an existing id, hot-reloads -->
<string name="title">Updated Title</string>
xml
<!-- WRONG for the fast path: new id, forces rebuild -->
<string name="brand_new_id">Hello</string>
<!-- WRONG because: adding a new string id grows the generated R class. R class regeneration
     is a schema change that ART cannot redefine in place. -->

Pattern: lambda count change (rebuild)

kotlin
// RIGHT (body change inside the existing lambda, hot-reloads)
@Composable
fun Row() {
    Button(onClick = { log("tapped") }) { Text("Tap") }
}
kotlin
// WRONG for the fast path (adds a second lambda, renumbers inner lambda classes)
@Composable
fun Row() {
    Button(onClick = { log("tapped") }) { Text("Tap") }
    Button(onClick = { log("second") }) { Text("Second") }
}
// WRONG because: a Kotlin function's compiled lambdas are anonymous classes named like
// Row$lambda$1, Row$lambda$2. Adding a lambda renumbers them and changes the class table,
// which is a schema change. The fix is to add the second composable as a separate top-level
// function (which is supported), then call it from Row once the structural change has rebuilt.

Mandatory rules

  • MUST consult the boundary tables before suggesting a change to a composable that the developer is iterating on. Suggesting a parameter add when the developer wanted speed wastes the iteration loop.
  • MUST NOT claim a change will hot-reload without checking it against the supported-changes table above.
  • MUST NOT disable HotSwan to "force a clean rebuild". Use Gradle's :app:clean for a clean build; HotSwan auto-falls-back when needed and turning it off costs the next round-trip.
  • MUST treat new resource ids and inline-function body changes as rebuild-forcing on every Android API level. The data-class-property exception is API 30+ only.
  • PREFERRED: order an iteration session so body-only edits come first and signature, constructor, interface, or new-resource-id changes are batched at the end.
  • PREFERRED: when a refactor needs a parameter add, introduce it once with a default value (accept the rebuild), then iterate on the body across subsequent saves.

Verification

  • A body-only change to a composable hot-reloads on save with the tool window status remaining WATCHING.
  • Adding a parameter to a composable triggers the fallback status in the tool window.
  • Adding a new <string> resource triggers the fallback; editing an existing <string> value does not.
  • Editing the body of an inline function triggers the fallback; converting it to a non-inline function and editing the body hot-reloads.
  • Adding a property to a data class hot-reloads on a device running API 30+; on API 29 and below it triggers the fallback.

References

Frequently asked questions

What does the Understanding Hot Reload Limits AI skill do?

Use this skill to teach Claude exactly which Kotlin and Compose changes hot-reload under Compose HotSwan and which trigger a full incremental rebuild fallback. Root cause is Android Runtime (ART) class schema immutability; only method bodies are mutable at runtime, so any change to fields, signatures, constructors, interfaces, inline functions, or new resource ids forces a rebuild. Covers the supported-changes table, the rebuild-forcing list, the diff-then-batch workflow that keeps a hot-reload session inside the fast path, and the inline-function and new-resource-id pitfalls. Trigger when...

Why use Understanding Hot Reload Limits on TypingMind?

Because you install it once and use it with any model. Understanding Hot Reload Limits is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Understanding Hot Reload Limits in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/skydoves/compose-performance-skills/tree/main/hot-reload/understanding-hot-reload-limits. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Understanding Hot Reload Limits?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Understanding Hot Reload Limits?

As many as you like. As long as a model supports skills, you can use Understanding Hot Reload Limits with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Understanding Hot Reload Limits AI skill free?

Yes. It is published on GitHub by skydoves under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇