<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by golang学习记 on Medium]]></title>
        <description><![CDATA[Stories by golang学习记 on Medium]]></description>
        <link>https://medium.com/@2084267015?source=rss-844c6602f269------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*R18RSqVgIFu3rArpt2KUKw.png</url>
            <title>Stories by golang学习记 on Medium</title>
            <link>https://medium.com/@2084267015?source=rss-844c6602f269------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 14:25:38 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@2084267015/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Claude Code Just Announce: You Can Now Create Your Own Themes, Animated Backgrounds, and More]]></title>
            <link>https://medium.com/@2084267015/claude-code-just-announce-you-can-now-create-your-own-themes-animated-backgrounds-and-more-b0a98a2aae73?source=rss-844c6602f269------2</link>
            <guid isPermaLink="false">https://medium.com/p/b0a98a2aae73</guid>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[claude-code]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[golang学习记]]></dc:creator>
            <pubDate>Tue, 17 Nov 2026 13:44:56 GMT</pubDate>
            <atom:updated>2026-11-17T13:44:56.245Z</atom:updated>
            <content:encoded><![CDATA[<p>I’ve noticed something funny about developers since Ai era comes</p><p>Most people kill time at work by scrolling through TikTok or shopping online</p><p>Developers? We spend our free time customizing our tools</p><p>Changing terminal colors. Trying new fonts</p><p>Tweaking editor themes. Moving panels around until everything feels just right</p><p>Try our best to make our env better</p><p>And if you use Claude Code every day, you probably know the feeling deeply</p><p>You open your computer, stare at that familiar dark interface, and watch Claude generate code for hours.</p><p>After a full day of coding, your eyes are exhausted</p><p>And honestly, the black-and-white terminal look gets a little boring.</p><p>Well, Anthropic just gave us something interesting.</p><p><strong>Claude Code Desktop now supports custom themes. And it’s much more than changing a few colors.</strong></p><p>You can even add animated backgrounds to your workspace.</p><p>Yes, your AI coding environment can now have its own personality.</p><h3>1. Meet Claude Code Mods</h3><p>Previously, customizing Claude Code mostly meant tweaking terminal colors, fonts, and a few external tools.</p><p>Now Anthropic has introduced Mods.</p><p>The idea is simple: tell Claude what kind of interface you want, and let Claude build it for you.</p><p>For example, you could ask:</p><p>“Create a cyberpunk theme for me. I want a rainy Tokyo street in the background, neon lights, and an overall dark color palette</p><p>Claude can generate a Mod based on your description.</p><p>You can customize interface colors, add animated backgrounds, and change the overall look of your workspace</p><p>Think underwater scenes, sunsets, particle animations, and all sorts of visual effects.</p><p>This is no longer just about picking a theme from a dropdown menu.</p><p>It’s more like having Claude design its own home</p><h3>2. How to Create Your First Custom Theme</h3><p>Let’s get our hands dirty.</p><p>Claude Code introduces a new command:</p><pre>/custom-theme</pre><p>As the name suggests, this is where the customization begins</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Hv5DtlmUpoxSUuYi.png" /></figure><p>For example, one of the built-in themes is called honu-lagoon.</p><p>The name comes from the famous artificial lagoon at Ko Olina on Oahu, Hawaii</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*6Q_T3eXuJJAGk5Qp.png" /></figure><p>And the result is gorgeous.</p><p>Run the command and you get this beautiful blue ocean-inspired environment.</p><p>There’s even a sea turtle swimming around in the water.</p><p>I mean, come on. Who wouldn’t want to write code in an environment like this?</p><p>But the really interesting part is that you’re not limited to built-in themes.</p><p>You can simply describe what you want.</p><p>For example:</p><p>“Create a synthwave sunset theme for me.”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*6jeOaaZAb0Ft_fvk.png" /></figure><p>A few seconds later, Claude generates the theme and applies it</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*mz-mtGwN4Vv001Sq.png" /></figure><p>And here’s the result.</p><p>Pretty cool than I think</p><h3>3. More Themes to Explore</h3><p>There’s one slightly confusing thing about the built-in themes.</p><p>If you want to use the official theme examples, you first need to install the theme examples plugin.</p><p>Run this command:</p><pre>/plugin install theme-examples@claude-plugins-official</pre><p>Once installed, you can explore more ready-made themes.</p><p>Here are two of my favorites.</p><p><strong>Night Reef</strong></p><p>A dark underwater theme featuring glowing sea turtles and manta rays.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*icym8Y0jGHTcn2KH.png" /></figure><p><strong>Paper Garden</strong></p><p>A paper-cut garden style, with green leaves and coral-colored flowers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*dwgGmdI6dZEgjYCY.png" /></figure><p>Honestly, these look less like coding environments and more like little digital worlds</p><h3>4. Different Themes for Different Sessions</h3><p>Here’s another detail I really like for it</p><p>You don’t have to apply the same theme everywhere.</p><p>Claude Code lets you customize individual sessions.</p><p>Use:</p><pre>/custom-theme &lt;theme-name&gt; session</pre><p>This applies the selected theme to the current session.</p><p>If you leave out session,</p><p>the theme is applied across all sessions.</p><p>And if you get tired of the customization and want to return to the original look, just do it:</p><pre>/custom-theme off</pre><p>One command, and you’re back to the familiar environment.</p><p>I can already imagine people assigning different themes to different projects.</p><p>A cyberpunk theme for backend development.</p><p>A peaceful ocean theme for debugging.</p><p>And maybe a colorful garden theme for those days when production is on fire.</p><p>Whatever helps you survive another day of coding.</p><h3>5. Even Merging Code Gets a Celebration</h3><p>Anthropic also added a little visual treat for the merge experience.</p><p>When you merge code, Claude Code can show a fireworks animation</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*51rYWe2Swrl3sZMq.png" /></figure><p>It’s a small detail, but I love it so much</p><p>It means something done to me</p><p>After spending hours fixing bugs, reviewing changes, and fighting with Git, sometimes you just deserve a little celebration</p><p>Even if the fireworks are virtual.</p><h3>Final words</h3><p>Let’s be honest.</p><p>Custom themes probably won’t make me write code 30% faster</p><p>They won’t fix your bugs, improve your architecture, or magically make Claude smarter.</p><p>But I genuinely like what Anthropic is doing now</p><p>AI coding tools have spent the last few years competing on model intelligence, context windows, and coding capabilities.</p><p>Now they’re starting to compete on the experience itself.</p><p>And that’s interesting.</p><p>Previously, we used Claude Code to build software only</p><p>Now we can ask Claude Code to build and customize the environment we use to work with it</p><p>I think being happy through the process is more important than finishing something</p><p>still,thank you for reading</p><p>see you next time</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b0a98a2aae73" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[VS Code Has a Hidden Superpower: You Can Work on GitHub Repositories Without Cloning Them]]></title>
            <link>https://medium.com/@2084267015/vs-code-has-a-hidden-superpower-you-can-work-on-github-repositories-without-cloning-them-3e3299d3badf?source=rss-844c6602f269------2</link>
            <guid isPermaLink="false">https://medium.com/p/3e3299d3badf</guid>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[github]]></category>
            <category><![CDATA[productivity]]></category>
            <dc:creator><![CDATA[golang学习记]]></dc:creator>
            <pubDate>Mon, 16 Nov 2026 12:28:25 GMT</pubDate>
            <atom:updated>2026-11-16T12:28:25.924Z</atom:updated>
            <content:encoded><![CDATA[<p>There is a story from me</p><h3>Last Saturday afternoon, I went to a local VS Code meetup</h3><p>These meetups usually focus on practical things: coding tricks, useful extensions, and workflows that developers have picked up over the years In there own real work</p><p>But the last couple of years have changed things quite a bit</p><p>AI has brought so many new possibilities to everyday development. Even a regular Saturday afternoon can turn into a series of unexpected discoveries.</p><p>More than once, I found myself thinking:</p><p><strong>“Wait, VS Code can actually do that?”</strong></p><p>Some people say VS Code is falling behind in the AI era</p><p>I don’t really agree</p><p>I think VS Code is still one of the best editing experiences around.</p><p>Even Cursor has very strong abilities of Ai dev</p><p>And sometimes, its most interesting features have nothing to do with AI</p><p>One of the things shared at the meetup was a feature called <strong>Virtual Workspaces</strong></p><p>Before that, I dont even know that concept of vscode</p><p>I’d never really explored it before</p><p>And honestly, it surprised me very strongly</p><p>whats next is what I learn from that meetup</p><h3>1. What Is a Virtual Workspace?</h3><p>Think about how you normally start working on a new project when You go to a new work</p><p>You find a repository on GitHub,</p><p>clone it,</p><p>choose a directory,</p><p>and wait for the files to download</p><p>Now you have a local copy of the project</p><p>Let’s call this a <em>physical workspace</em>. The files actually exist on your computer. You can find them with ls, open them in any editor, and do pretty much whatever you want</p><p>sometimes these steps are so boring when I repeat many times</p><p>A Virtual Workspace turns this workflow upside down.</p><p><strong>You can open a GitHub repository directly in VS Code without downloading the project to your local disk</strong></p><p>That’s right. No traditional clone. No project directory to manage</p><p>It is so simple,I like it so much</p><p>VS Code uses a virtual file system to make files from the remote repository look and behave much like local files</p><p>You can still do many things like browse the project, search across files, edit code, review diffs, and commit changes.</p><p>The files aren’t sitting in a normal directory on your computer</p><p>Run ls in your usual project folder.</p><p>Nothing new will be there</p><p>It’s a surprisingly different way to work to me</p><h3>2. How to Get Started</h3><p>Getting started is also pretty simple</p><p>This is also why I love it</p><p>First, open the VS Code Extensions Marketplace and search for:</p><p>GitHub Repositories</p><p>Install the extension and restart VS Code if needed.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*izWMzj3zEP1v6QxO.png" /></figure><p>There are two easy ways to open a repository.</p><p><strong>Option 1: Use the Remote Indicator</strong></p><p>Click the green Remote indicator in the bottom-left corner of VS Code.</p><p><strong>Option 2: Use the Command Palette</strong></p><p>Press:</p><p>Cmd + Shift + P</p><p>Then search for:</p><p>Open Repository</p><p>Sign in to GitHub if prompted, enter a repository URL, and hit Enter</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/419/0*n1XGdVeRLRBdxeVj.png" /></figure><p>You can also open a repository through the alternative entry point shown below</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*miZGQyBiynjwYwuU.png" /></figure><p>A moment later, the file explorer fills up with files as usual</p><p>You’ll notice something different in the path bar: GitHub.</p><p>If you see that</p><p>Congratulations. You’re now working inside a Virtual Workspace</p><h3>3. What Can You Actually Do?</h3><p>Quite a lot, actually</p><p>Once the repository is open, you can:</p><ul><li>Browse files and folders.</li><li>Search across the entire repository.</li><li>Edit source files.</li><li>Review changes and diffs.</li><li>Create commits and send changes to GitHub.</li></ul><p>There are also a few details that make this workflow particularly interesting</p><p><strong>First, there’s no traditional Push step.</strong></p><p>When you commit your changes in this workflow, the changes are committed directly to the remote repository.</p><p>You don’t need to go through the usual local commit followed by git push process.</p><p><strong>Branch switching is also much easier than before</strong></p><p>You don’t have to manually run git pull just to refresh your local copy before switching branches. The extension handles the remote repository updates for you.</p><p>And here’s another nice touch.</p><p>Imagine you’ve modified a few files but haven’t committed them yet. Normally, switching branches might make you think about running git stash first</p><p>With this workflow, the extension takes care of preserving those changes when switching branches</p><p>We dont need to put our worries on lost things</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/622/0*6jqQZjDoouY_vnYA.png" /></figure><p>For quickly exploring a repository or making a small change, this feels surprisingly smooth.</p><h3>4. But There’s a thing to keep in mind</h3><p>Virtual Workspaces aren’t a replacement for traditional local development.</p><p>There are some important limitations.</p><p>For example, you can’t simply treat the workspace like a normal local project and expect everything to work.</p><p>Running terminal commands, debugging applications, installing dependencies, and executing local build processes are not supported in the same way.</p><p>If you need to run the project, you’ll still want a traditional local workspace or another suitable remote development environment</p><p>So, where does Virtual Workspace make the most sense?</p><p>I’d say:</p><ul><li>Quickly exploring an unfamiliar GitHub project.</li><li>Reviewing someone else’s code.</li><li>Making small changes to a repository.</li><li>Checking configuration files and documentation.</li><li>Fixing a typo without cloning an entire project.</li></ul><p>It’s not designed to replace your main development environment</p><p>It’s designed to remove unnecessary setup when you just need to work with some code.</p><h3>Final thingking</h3><p>I left that Saturday meetup with a few new ideas.</p><p>After fixing a couple of small issues, I committed my changes and closed my vscode</p><p>No new project directory. No local clone. Nothing extra sitting on my disk.</p><p>It felt so satisfying</p><p>I think this is one of the reasons I still enjoy using VS Code.</p><p>People often talk about how quickly AI tools are evolving. And that’s fair. But a great developer experience isn’t just about having the smartest AI assistant.</p><p>It’s also about removing friction from everyday work.</p><p>Virtual Workspaces are a good example.</p><p>If you alose like the world of No cloning and No setup</p><p>Just open it and explore</p><p>And honestly, that’s a pretty nice way to spend an afternoon</p><p>hope you will have a nice day with this feature</p><p>Last,If you like my story or words,you can give me a clap or follow that is a real joy for me</p><p>Thank you for reading</p><p>see you next time</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3e3299d3badf" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Many Years In the Waiting, Go Will Introduce Generic Collection API, So Beautiful]]></title>
            <link>https://medium.com/@2084267015/many-years-in-the-waiting-go-will-introduce-generic-collection-api-so-beautiful-fc1df2a2f5c9?source=rss-844c6602f269------2</link>
            <guid isPermaLink="false">https://medium.com/p/fc1df2a2f5c9</guid>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[golang]]></category>
            <category><![CDATA[programming]]></category>
            <dc:creator><![CDATA[golang学习记]]></dc:creator>
            <pubDate>Sun, 15 Nov 2026 08:32:19 GMT</pubDate>
            <atom:updated>2026-11-15T08:32:19.518Z</atom:updated>
            <content:encoded><![CDATA[<p>If you’ve been writing Go for a few years, you’ve probably experienced this like me</p><p>You need to remove duplicates from a slice</p><p>Or check whether two slices have anything in common.</p><p>What do you do usaully before?</p><p>You open Google and search:</p><p>“How to implement a Set in Go?”</p><p>Then you copy the same piece of code you’ve probably copied a hundred times before:</p><pre>set := make(map[string]struct{})</pre><pre>for _, item := range list {<br>    set[item] = struct{}{}<br>}</pre><p>Simple. Effective. Very Go.</p><p>And then you do it again.</p><p>These days, I don’t even bother opening Google. I just ask Codex, DeepSeek, or another AI tool to generate the same code for me</p><p>Apparently, I’ve managed to automate the process of writing boilerplate code that shouldn’t have existed in the first place.</p><p>The problem gets worse when I need more than deduplication</p><p>What if I need a union?</p><p>An intersection?</p><p>A difference between two collections?</p><p>Or maybe I want to iterate over a map in sorted order.</p><p>Suddenly, I am writing more loops, building helper functions, or pulling in another dependency like github.com/deckarep/golang-set.</p><p>I’ve done this more times than I’d like to admit.</p><p>But there’s some interesting news from the Go ecosystem</p><p>The experimental golang.org/x/exp/collection module now provides a collection of data structure packages, including set and hashed.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/859/1*RbhLYje937uArAoKhZnubA.png" /></figure><p>Let’s take a look at what they actually offer.</p><p>And then I will show you how I use it in my work</p><h3>1. The set Package: Finally, Real Set Operations</h3><p>This is probably the most immediately useful package</p><p>The set package provides a convenient wrapper around Go&#39;s familiar map[E]struct{}.</p><p>In other words, it doesn’t introduce some complicated new data structure. It simply gives us a much nicer API.</p><p>For example, imagine you have two user lists and want to find their common friends.</p><pre>package main</pre><pre>import (<br>    &quot;fmt&quot;</pre><pre>    &quot;golang.org/x/exp/collection/set&quot;<br>)</pre><pre>func main() {<br>    userA := set.Of(&quot;Alice&quot;, &quot;Bob&quot;, &quot;Charlie&quot;)<br>    userB := set.Of(&quot;Bob&quot;, &quot;Charlie&quot;, &quot;David&quot;)</pre><pre>    // Intersection: common users<br>    common := userA.Intersection(userB)<br>    fmt.Println(common)</pre><pre>    // Union: all unique users<br>    all := userA.Union(userB)<br>    fmt.Println(all)</pre><pre>    // Difference: users only in A<br>    onlyA := userA.Difference(userB)<br>    fmt.Println(onlyA)</pre><pre>    // Check whether two sets overlap<br>    if userA.Intersects(userB) {<br>        fmt.Println(&quot;They have common friends!&quot;)<br>    }<br>}</pre><p>Much cleaner, I feel when I see this first time</p><p>You also get methods like Equal(), Contains(), and SymmetricDifference()</p><p>One detail I particularly like is the distinction between methods with and without the With suffix.</p><p>For example:</p><pre>result := userA.Union(userB)</pre><p>Returns a new set.</p><p>But:</p><pre>userA.UnionWith(userB)</pre><p>Modifies the original set directly.</p><p>The same pattern applies to Intersection, Difference, and other operations.</p><p>I also like the set.Of() constructor.</p><p>Java developers might recognize the idea from List.of(), Set.of(), and Map.of().</p><p>I’m not claiming Go copied Java here, but I do appreciate this kind of API design. It’s short, readable, and tells you exactly what it does at the first glance</p><p>Sometimes good API design really is that simple.</p><h3>2. The hashed Package: When Built-in Maps Aren&#39;t Enough</h3><p>At this point, you might be wondering:</p><p>“If set is basically a map, why do we need another package called hashed?&quot;</p><p>Good question</p><p>what’a next is the reason</p><p>Go’s built-in maps require comparable keys</p><p>That means you can’t directly use a struct containing a slice as a map key.</p><p>For example:</p><pre>type User struct {<br>    ID      int<br>    Hobbies []string<br>}</pre><p>This type cannot be used as a key in a regular Go map because slices aren’t comparable.</p><p>But what if you want two User values to be considered equal whenever their IDs match?</p><p>That’s why hashed exists</p><p>It allows you to define your own hashing and equality rules through the maphash.Hasher interface.</p><p>I will give a simple example</p><pre>type User struct {<br>    ID      int<br>    Hobbies []string<br>}</pre><pre>type UserHasher struct{}</pre><pre>func (UserHasher) Hash(<br>    h *maphash.Hash,<br>    u User,<br>) {<br>    maphash.WriteComparable(h, u.ID)<br>}</pre><pre>func (UserHasher) Equal(a, b User) bool {<br>    return a.ID == b.ID<br>}</pre><p>Then create a map:</p><pre>m := hashed.NewMap[User, string](UserHasher{})</pre><pre>u1 := User{<br>    ID:      1,<br>    Hobbies: []string{&quot;coding&quot;},<br>}</pre><pre>m.Set(u1, &quot;Active&quot;)</pre><pre>u2 := User{<br>    ID:      1,<br>    Hobbies: []string{&quot;reading&quot;},<br>}</pre><pre>fmt.Println(m.Get(u2)) // Active</pre><p>Even though the two users have different hobbies, they’re considered equal because their IDs match.</p><p>That’s pretty powerful.</p><p>Of course, there’s an important rule: your hash function and equality function must agree. Equal values must produce the same hash.</p><p>And here’s something the official documentation explicitly warns about:</p><p><strong>If your key is already comparable and uses normal Go equality, just use the built-in map.</strong></p><p>Don’t use hashed just because it looks more sophisticated.</p><h3>3. What About Sorted Collections?</h3><p>This is another feature Go developers including me have wanted for a long time</p><p>Go maps don’t guarantee iteration order.</p><p>If you need sorted output, you normally have to collect the keys, sort them, and then iterate over the map.</p><p>For example:</p><pre>keys := make([]string, 0, len(m))</pre><pre>for key := range m {<br>    keys = append(keys, key)<br>}</pre><pre>sort.Strings(keys)</pre><pre>for _, key := range keys {<br>    fmt.Println(key, m[key])<br>}</pre><p>It works, but it’s not very elegant.</p><p>The collection ecosystem is exploring richer data structures for these kinds of use cases</p><p>One particularly interesting idea is a range-based view, often described as a Span: instead of copying an entire collection, you work with a selected range of its contents.</p><p>However, I couldn’t verify the previously mentioned sorted.NewMap() and From().To() APIs in the currently published golang.org/x/exp/collection documentation, so I wouldn&#39;t recommend copying those examples into production code yet.</p><p>It’s worth keeping an eye on this area, though</p><p>if this package lands, it will give go devs much efficience than before</p><h3>Final Thoughts</h3><p>I like what Go is doing here</p><p>The real improvement is much more practical: making common operations easier to express without forcing every developer to write the same helper functions again and again</p><p>it is so boring</p><p>Still, I’m happy to see this direction.</p><p>I’ve spent enough years writing map[string]struct{} and asking AI to generate another set implementation.</p><p>Maybe one day, when I need a simple set operation, I can finally just write:</p><pre>common := userA.Intersection(userB)</pre><p>And get back to building something interesting.</p><p>That’s all I really wanted sometimes</p><p>If you like my story,you can give me a clap or follow that will make me so happy</p><p>Thank you for reading</p><p>see you next time</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fc1df2a2f5c9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[VS Code Just Fixed One of My Biggest Problems With AI Coding: The Session Inbox]]></title>
            <link>https://medium.com/@2084267015/vs-code-just-fixed-one-of-my-biggest-problems-with-ai-coding-the-session-inbox-97f528c1dd55?source=rss-844c6602f269------2</link>
            <guid isPermaLink="false">https://medium.com/p/97f528c1dd55</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[golang学习记]]></dc:creator>
            <pubDate>Fri, 13 Nov 2026 14:11:00 GMT</pubDate>
            <atom:updated>2026-11-13T14:11:00.686Z</atom:updated>
            <content:encoded><![CDATA[<p>Hello Guys</p><p>I recently found something that might actually save my sanity.</p><p>I’ve been using AI coding tools more and more lately. Claude Code, Codex, and a few other agents are now part of my daily workflow deeply</p><p>something happy,at the same time something unhappy</p><p>And something strange has happened.</p><p>I spend more time talking to AI than writing code.</p><p>It used to be the other way around. Most of my time went into actually building things. Now, a huge chunk of my day is spent managing AI conversations.</p><p>At any given moment, I might have three Claude Code sessions running on different branches, while Codex quietly works on a bug in the background.</p><p>Sounds amazing, right?</p><p>One developer. An entire engineering team.</p><p>Well, not quite to me</p><h3>The Problem: I’m Managing AI Agents Instead of Writing Code</h3><p>Here’s what my typical day looks like.</p><p>One AI agent is stuck, waiting for my permission.</p><p>Another one has finished running tests and is ready for a review.</p><p>A third one has crashed, with a wall of red error messages.</p><p>And where am I?</p><p>what I do mostly is</p><p>Jumping between workspaces, tabs, and conversation lists, trying to figure out which agent needs me.</p><p>It’s like being a manager with dozens of remote employees working in different rooms</p><p>Finnaly I hate this feeling</p><p>Because there’s no central dashboard.</p><p>I have to walk into every room just to check what’s going on.</p><p>By the end of the day, I’ve barely written any code. My neck hurts, my eyes are tired, and somehow I’m still behind schedule.</p><p>The funny thing is that AI agents already have notifications.</p><p>But notifications alone don’t solve the problem.</p><p>When you’re running multiple agents across different projects, you don’t just need to know that something happened.</p><p>You need to know what actually needs your attention.</p><p>That’s when I discovered a new feature in VS Code:</p><p><strong>Session Inbox.</strong></p><p>And honestly,I know this is exactly what I need</p><h3>Meet Session Inbox: An Inbox for Your AI Agents</h3><p>The idea immediately reminded me of email</p><p>Think about it.</p><p>When you receive a few messages, it’s easy to handle them one by one.</p><p>But when hundreds of messages start coming in, you need an inbox. You need a way to sort things, prioritize them, and decide what deserves your attention.</p><p>VS Code is bringing that same idea to AI agent sessions.</p><p>Instead of checking every workspace individually, you get one unified view of what all your agents are doing.</p><p>You can quickly see which tasks need your attention, which agents are still working, and which ones have done and are ready for review</p><p>No more hunting through conversation lists.</p><p>Just one simple inbox</p><h3>How to Enable It</h3><p>There’s one small catch.</p><p>This is currently an experimental feature, and it’s disabled by default.</p><p>To enable it:</p><ol><li>Open VS Code Settings.</li><li>Search for sessions.list.inboxEnabled.</li><li>Turn it on.</li><li>Go to the Agents session list.</li><li>Change the grouping mode to <strong>Inbox Mode</strong>.</li></ol><p>You can also enable it directly from the settings interface shown below</p><p>This is also my take</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/0*65EYQF91MSJmmwT7.png" /></figure><p>Once enabled, your sessions are no longer just a long list of conversations.</p><p>They’re organized into three useful groups</p><h3>Three Groups That Actually Make Sense</h3><h3>1. Needs Your Attention</h3><p>This is the most important section.</p><p>It contains sessions that need you to do something.</p><p>For example:</p><ul><li>An agent is waiting for permission.</li><li>It has encountered a problem and needs your input.</li><li>It cannot continue without a decision from you.</li></ul><p>These sessions appear at the top.</p><p>That’s exactly where they should be.</p><p>You don’t have to remember many things</p><p>VS Code will bring it to you, this feel so good</p><h3>2. Still Running</h3><p>These are the agents that are still doing their jobs.</p><p>Nothing needs your attention yet.</p><p>You can leave them alone and get back to your own work</p><h3>3. Ready to Merge</h3><p>This is probably my favorite section.</p><p>These are sessions where the agent has finished its work and is ready for you to review.</p><p>Maybe it has implemented a feature. Maybe it has fixed a bug and passed the tests.</p><p>Either way, the next step is yours.</p><p>Review the changes, check the result, and decide whether to merge.</p><p>It’s a much cleaner way to handle completed tasks.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/279/0*7UkhORKrxV7Ly4J4.png" /></figure><h3>From “I Find Work” to “Work Find Me”</h3><p>This is the biggest change for me</p><p>Before Session Inbox,</p><p>I waste much time on managing these agents</p><p>Every time I did this, I broke my concentration.</p><p>Now, I don’t need to keep checking</p><p>If an agent is still working, I leave it alone.</p><p>If it’s finished, I’ll review it when I’m ready.</p><p>If it needs my help, it shows up in the inbox</p><p>last but not least, vscode is too kind</p><p>VS Code also lets you snooze a session and deal with it later.</p><p>That’s another surprisingly useful detail for me</p><p>Sometimes an agent needs a decision, but you’re in the middle of something else. You don’t necessarily want to deal with it right now.</p><p>Just snooze it and come back when you’re ready.</p><p>The whole workflow feels much more natural.</p><p><strong>I used to look for work. Now, the work comes to me.</strong></p><p>Everyone in the AI industry is obsessed with model intelligence, context windows, coding benchmarks, and generation speed.</p><p>And sure, those things have their own importance</p><p>But I think we’re starting to run into a different problem.</p><p>When AI agents become capable enough to handle multiple tasks at once, managing those agents becomes a challenge of its own.</p><p>A model can generate thousands of lines of code in seconds.</p><p>But if I have to manually check ten different windows to figure out what happened, That is another disaster</p><p><strong>The user interface for human-AI collaboration is becoming just as important as the AI itself.</strong></p><p>Anyway, enough talking.</p><p>My inbox just moved another agent into “Needs Your Attention”.</p><p>Looks like something needs their manager again.</p><p>I must back to work</p><p>see you next time</p><p>dont forget give me a clap or follow if you like my story</p><p>thank you for spending time reading</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=97f528c1dd55" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Zed Just Got Native macOS Tabs, and It’s a Bigger Deal Than You Think]]></title>
            <link>https://medium.com/@2084267015/zed-just-got-native-macos-tabs-and-its-a-bigger-deal-than-you-think-2640244de33b?source=rss-844c6602f269------2</link>
            <guid isPermaLink="false">https://medium.com/p/2640244de33b</guid>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[tech]]></category>
            <dc:creator><![CDATA[golang学习记]]></dc:creator>
            <pubDate>Tue, 10 Nov 2026 13:03:04 GMT</pubDate>
            <atom:updated>2026-11-10T13:03:04.687Z</atom:updated>
            <content:encoded><![CDATA[<p><em>For our daily drive many times,One tiny setting. A much nicer workflow</em></p><p>Today i will tell a story like this for you as a dev</p><h3>The Problem Nobody Talks About</h3><p>If you’re a developer working on a Mac,</p><p>As far as I know,many devs on vscode and zed team use mac computer as their main OS</p><p>you probably encounter this situation.</p><p>You’re working on one project. Then you need to check another repository. Maybe you want to compare two implementations, inspect a bug on a different branch, or quickly open a small side project. Or any other things like that</p><p>So you press Cmd + N for that thing</p><p>Another window appears.</p><p>Then another</p><p>And another</p><p>A few hours later, You dont konw how many windows you are opening</p><p>You try to find the window you opened earlier. Swipe up. Swipe sideways. Click around. Eventually, you find the right one</p><p>And sometimes you want close some unimporttant ones</p><p>Congratulations. You’ve spent more time managing windows than writing code</p><p>Either way，you will consume much time on it</p><p>But you may konw macOS has had a solution for this for years: native window tabs.</p><p>Safari has them. Finder has them.</p><p>Plenty of native Mac apps have them</p><p>But code editors? The experience has historically been a little less happy</p><p>But the good news comes</p><p>Now, Zed is bringing macOS native window tabs into its workflow.</p><p>And honestly, this is one of those small features that makes you wonder why you didn’t have it sooner.</p><h3>Wait, Isn’t This Just Another Tab Bar?</h3><p>Not quite.</p><p>There are two different kinds of tabs developers tend to mix up.</p><p>The first is editor tabs: the familiar row of files inside your IDE. main.rs, app.tsx, config.json, and so on.</p><p>The second is window tabs: a macOS-level feature that lets multiple application windows live inside one window, each represented by its own tab.</p><p>That’s what Zed is integrating.</p><p>Normally, opening each project in a separate window means juggling three windows</p><p>just like this at zed</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/386/0*epbGgyZMXR9oCl_2.png" /></figure><p>With native window tabs enabled, macOS can group those windows together</p><p>You get one window with multiple tabs, while each project keeps its own editor state, file tree, terminal, and workspace context</p><p>you can see many projects at the same time</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Hi8a4y2srtGrBX-Y.png" /></figure><p>It’s a surprisingly elegant arrangement that I never see at any other code editor</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*2wYrz51o3GaVeSzD.png" /></figure><p>a window is a tab,And then a projecy is a tab</p><h3>How to Enable It</h3><p>Here’s the fun part. You don’t need a complicated extension or some experimental plugin.</p><p>Zed exposes a dedicated configuration option:</p><p>use_system_window_tabs</p><p>Open Zed’s Command Palette with Cmd + Shift + P, search for zed: open settings, and add the following to your settings.json:</p><pre>{<br>  &quot;use_system_window_tabs&quot;: true<br>}</pre><p>That’s it on the Zed side</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/839/0*sQQ4JgHtV_3-FHsS.png" /></figure><p>The only downside is Zed hasn’t put it in the Settings UI yet</p><p>There is one more macOS setting to check</p><p>Open:</p><p><strong>System Settings → Desktop &amp; Dock → Windows → Prefer tabs</strong></p><p>Set it to <strong>Always</strong>.</p><p>macOS also offers options such as In Full Screen and Never. The exact behavior of window grouping follows this system preference.</p><p>Now open multiple Zed projects in separate windows.</p><p>macOS can group them into native tabs, giving you a much cleaner way to switch between workspaces.</p><p>One small detail worth remembering: this is different from Zed’s regular editor tab settings. You don’t need to change tab_bar.show or any of the file-tab configuration options to enable system window tabs.</p><p>In fact,I want talk more deeper</p><p>Here’s where things get interesting.</p><p>Zed isn’t built like a typical Electron editor.</p><p>It’s written in Rust, uses GPUI for its interface, and relies on GPU-accelerated rendering.</p><p>All is fresh</p><p>That makes macOS integration particularly interesting</p><p>At the same time,it is alse difficult for zed devs</p><p>Native window tabs aren’t simply another visual component.</p><p>They involve the operating system’s window management behavior, including how windows are grouped, activated, and presented to the user.</p><p>Instead of drawing a half-baked tab bar inside the application, Zed can delegate this behavior to macOS itself.</p><p>That’s a meaningful architectural choice</p><p>It means the operating system remains responsible for the window-level interaction, while Zed focuses on the actual development workspace.</p><p>Zed give us a good example of how to build a custom UI with native platform features</p><h3>A Small Feature That Makes a Big Difference</h3><p>I know what some VS Code users might be thinking.</p><p>“Wait, VS Code already supports native macOS tabs. What’s so special about this?”</p><p>Yes</p><p>VS Code has supported native window tabs through its own window.nativeTabs setting for quite some time.</p><p>Absolutely Zed isn’t inventing the concept</p><p>I think Zed has spent a lot of time rethinking what a modern code editor should feel like: startup performance, rendering, collaboration, project navigation</p><p>The interesting question isn’t whether Zed has one more checkbox than vsc</p><p>It’s whether the entire application feels like a natural part of the operating system.</p><p>Wrapped</p><p>Sometimes it’s just fewer windows that make uour life better and easy</p><p>last but not leat , if you like this story,you can give me a clap or follow that will make me feel so happy as a little writer</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2640244de33b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Go1.28 Is Getting maps.Grow: Now You Can Preallocate a Map You Didn't Create]]></title>
            <link>https://medium.com/@2084267015/go1-28-is-getting-maps-grow-now-you-can-preallocate-a-map-you-didnt-create-61051fdf08fa?source=rss-844c6602f269------2</link>
            <guid isPermaLink="false">https://medium.com/p/61051fdf08fa</guid>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[golang]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[golang学习记]]></dc:creator>
            <pubDate>Sun, 08 Nov 2026 14:11:12 GMT</pubDate>
            <atom:updated>2026-11-08T14:11:12.532Z</atom:updated>
            <content:encoded><![CDATA[<p>If you’re a Go developer, you’ve probably heard this advice a hundred times:</p><p><em>If you know how big your map is going to be, preallocate it with </em><em>make(map[K]V, size) to avoid unnecessary allocations and map growth.</em></p><p>It’s good advice. No argument there</p><p>Java developers will probably recognize the same idea from working with hash maps.</p><p>But there’s one rather inconvenient assumption hiding behind it:</p><p><strong>You have to be the person who creates the map.</strong></p><p>What if someone hands you an already initialized map</p><p>this situation is common when multi systems need the interaction</p><p>That’s exactly the problem Go’s new maps.Grow proposal is trying to solve.</p><p>Accepted in late October, this addition is expected to arrive with Go 1.28.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/617/1*TPL3YdIs4KLeLS_kCU3IjA.png" /></figure><h3>The Real Problem: A Map That Quietly Eats Your CPU</h3><p>Let’s make this concrete and simple to understand</p><p>Imagine you’re building a high-performance gateway service. One of its jobs is to copy HTTP response headers from an upstream server to the downstream client.</p><p>The code looks harmless enough:</p><pre>func copyHeaders(dst, src http.Header) {<br>    for k, vv := range src {<br>        for _, v := range vv {<br>            dst.Add(k, v)<br>        }<br>    }<br>}</pre><p>Here, dst is the header map we&#39;re writing to.</p><p>This map has already been created somewhere else in the request lifecycle. We didn’t initialize it, so we don’t get to decide its initial capacity.</p><p>Now imagine the upstream server sends back 30 different headers, while the destination map starts with room for only 10.</p><p>As we keep inserting map entries, Go may need to allocate more memory, grow the hash table, and move existing entries around</p><p>That will get worse and worse than you think at the beginning</p><p>One request doesn’t sound like a big deal. But multiply that by thousands or millions of requests, and those little allocations start adding up.</p><p>In one of our gateway services, this kind of map growth accounted for roughly 5% of CPU usage.</p><p>That’s not exactly the kind of work you want your CPU spending its time on.</p><p>Previously, one workaround was to replace the original map with a larger one. Some libraries, including Honeycomb’s SDK, have used variations of this approach: create a new map, copy the old entries over, and replace the reference.</p><p>It works, but it’s not particularly elegant.</p><h3>INtroduce maps.Grow</h3><p>This is where the new API comes in:</p><pre>func Grow[M ~map[K]V, K comparable, V any](m M, n int)</pre><p>The idea is refreshingly simple:</p><p><strong>Give an existing map enough room for at least </strong><strong>n additional entries before you start inserting them.</strong></p><p>No need to create a replacement map. No need to manually copy existing entries</p><p>We can simply change our previous example:</p><pre>func copyHeaders(dst, src http.Header) {<br>    // Make room for the incoming entries first.<br>    maps.Grow(dst, len(src))</pre><pre>    for k, vv := range src {<br>        for _, v := range vv {<br>            dst.Add(k, v)<br>        }<br>    }<br>}</pre><p>That’s it. That is So Simple Change</p><p>Instead of potentially triggering multiple growth operations inside the loop, we give the runtime a heads-up about how much space we’ll need.</p><p>One important detail: len(src) counts distinct header keys, not the total number of header values. That&#39;s exactly what matters for map capacity.</p><p>And yes, maps.Grow works with named map types too, thanks to its generic type constraint.</p><h3>Why Doesn’t maps.Grow Return a New Map?</h3><p>If you’ve used slices.Grow, you might be wondering:</p><p>Why do we write this?</p><p>And this is a classic interview question</p><pre>s = slices.Grow(s, n)</pre><p>But with maps, we can simply do this?</p><pre>maps.Grow(m, n)</pre><p>The difference comes down to how slices and maps are represented in Go.</p><p>A slice is a small descriptor containing three things:</p><ul><li>A pointer to its underlying array</li><li>Its current length</li><li>Its capacity</li></ul><p>When a slice grows beyond its available capacity, Go may allocate a new underlying array. The slice descriptor itself needs to be updated to point to that new array. That’s why slices.Grow returns a slice.</p><p>A map works differently. A map value refers to an underlying runtime hash table. Growing that table doesn’t require the caller to replace the map variable.</p><p>So maps.Grow can modify the existing map directly without returning anything.</p><p>One caveat worth mentioning: maps are still not safe for concurrent writes. Sharing a map between goroutines doesn’t magically become safe just because you’ve called Grow. You&#39;ll still need proper synchronization.</p><h3>There’s More: maps.Copy Gets Smarter Too</h3><p>The proposal doesn’t stop at introducing maps.Grow</p><p>Go team make one thing left better</p><p>It also improves maps.Copy</p><p>Previously, copying a map meant iterating through the source and inserting entries into the destination one by one.</p><p>With the new optimization, maps.Copy can grow the destination map before starting the copy, using the source map&#39;s size as a hint.</p><p>That means existing code like this:</p><pre>maps.Copy(dst, src)</pre><p>can benefit without requiring any changes at the call site.</p><p>A nice little performance improvement that doesn’t make your code any uglier. I’ll take that.</p><h3>A Couple of Things to pay attension to</h3><p>Before getting too excited, there are two details worth remembering.</p><p><strong>1. A nil map will panic.</strong></p><p>You can’t grow a nil map:</p><pre>var m map[string]int</pre><pre>maps.Grow(m, 10) // panic</pre><p>Just like you can’t insert entries into a nil map.</p><p>Initialize it first:</p><pre>m := make(map[string]int)<br>maps.Grow(m, 10)</pre><p><strong>2. Don’t confuse growth with an exact capacity guarantee.</strong></p><p>The n argument tells the runtime how much additional room you want. It doesn&#39;t expose the map&#39;s actual capacity or promise an exact memory allocation.</p><p>Think of maps.Grow as a way to give the runtime useful information ahead of time, not as a way to control every detail of its memory layout</p><h3>Wrapped</h3><p>For everyday CRUD applications, you might never notice this change.</p><p>But if you’re building API gateways, serialization libraries, middleware, or anything that repeatedly inserts data into maps created elsewhere, maps.Grow is a genuinely useful addition.</p><p>The best part is that it solves an awkward problem without introducing a complicated API.</p><p>I want to say that this design a great quality-of-life dev improvement</p><p>Sometimes the nicest language improvements aren’t flashy new features. They’re the ones that remove a workaround you’ve been putting up with for years.</p><p><strong>Go 1.28 will come soon</strong></p><p>I am looking forward to seesing Go get better and better</p><p>If you like it,you can give me a clap or follow,That will gain more power in creating something better</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=61051fdf08fa" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Chrome Finally Gets Vertical Tabs — And It’s About Time]]></title>
            <link>https://medium.com/@2084267015/chrome-finally-gets-vertical-tabs-and-its-about-time-f7e0bafe9442?source=rss-844c6602f269------2</link>
            <guid isPermaLink="false">https://medium.com/p/f7e0bafe9442</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[work]]></category>
            <dc:creator><![CDATA[golang学习记]]></dc:creator>
            <pubDate>Sat, 07 Nov 2026 13:31:37 GMT</pubDate>
            <atom:updated>2026-11-07T13:31:37.739Z</atom:updated>
            <content:encoded><![CDATA[<p><em>If you’re the kind of person who keeps 50 browser tabs open while debugging a production issue, this one’s for you.</em></p><p>As people who spend most of their working hours on the internet</p><p>maybe we spent much ineffecient time on finding a special tab, in fact I am the ineffecient one when I work for my life</p><p>Picture this.</p><p>You’re trying to track down a production bug. Your Chrome browser has GitHub open for the source code, Stack Overflow for a similar issue, your company’s internal API documentation, three different testing environments, a couple of competitor websites, and, of course, Weibo for a quick break.</p><p>Before you realize it, you’ve got moren than <strong>50 tabs open</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*2Gn8Wyb79T2dD4m2.png" /></figure><p>Now you need to get back to that GitHub page you opened two hours ago.</p><p>You move your mouse to the top of the screen and start playing a frustrating game of Where’s Waldo. The tabs have become so narrow that their titles are practically invisible. All you can see are tiny favicons lined up like passengers squeezing onto a rush-hour train.</p><p>You click around, hoping to find the right one.</p><p>Wrong tab.</p><p>You try again.</p><p>And somehow, you will close the page where you had just finished filling out an important form.</p><p>At this point, you start wondering whether your monitor needs to be twice as wide.</p><p>This is one of those browser design problems we’ve simply learned to live with: <strong>horizontal space is limited, but tabs keep piling up horizontally.</strong></p><p>Once you have more than ten or fifteen tabs open, finding anything becomes a guessing game.</p><p>Well, Chrome might finally have an answer.</p><p>It’s called <strong>Vertical Tabs</strong>.</p><p>One feature that give you a brand-new finding experience</p><h3>Vertical Tabs: A Feature We’ve Been Waiting For</h3><p>For anyone who regularly has dozens of tabs open, vertical tabs sound like a small change that makes a surprisingly big difference.</p><p>Microsoft Edge users have had this feature for a while, and Arc made vertical tab management a central part of its browser experience. But Chrome, despite being one of the most widely used browsers in the world, has taken its time getting here.</p><p>Now, the feature is finally making its way into Chrome.</p><p>So, what’s actually so good about it?</p><h3>1. More room for the things you’re actually working on</h3><p>Moving tabs from the top of the browser to the left side frees up the horizontal tab strip and gives you a cleaner browsing experience.</p><p>On a typical 16:9 laptop display, every bit of vertical screen space matters. If you’re reading long documentation, reviewing code, or comparing multiple pages, having less browser chrome at the top can make things feel noticeably less cramped.</p><h3>2. Tab titles are finally readable</h3><p>This is probably my favorite part.</p><p>With traditional horizontal tabs, Chrome quickly runs out of room. Eventually, all you’re left with is a row of tiny icons.</p><p>Is that GitHub? An API reference? A random article you opened three days ago?</p><p>Good luck figuring that out.</p><p>Vertical tabs give page titles much more room, making it easier to recognize what you’re looking for without clicking through half your browser.</p><h3>3. Managing dozens of tabs becomes much easier</h3><p>A vertical sidebar also makes it easier to organize your browsing sessions.</p><p>You can group related tabs together, collapse groups when you don’t need them, and keep different tasks separate.</p><p>You can also see which tabs are playing audio, which is surprisingly useful when some forgotten webpage suddenly starts making noise.</p><p>Your browser starts feeling less like a junk drawer and more like an organized workspace.</p><h3>How to Enable Vertical Tabs in Chrome</h3><p>The good news is that switching is pretty straightforward.</p><p><strong>Step 1.</strong> Right-click the Chrome tab bar.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*f5t0bcoyfb805PHQ.png" /></figure><p><strong>Step 2.</strong> Look for the option labeled <strong>“Show Tabs Vertical”</strong> and click it.</p><p>That’s it.</p><p>Your familiar row of horizontal tabs will move to the left side of the browser, where they’ll be displayed in a vertical list.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*U5JmA8DMRx1IJUdf.png" /></figure><p>Suddenly, finding a tab is no longer a mini-game.</p><p>Closing unwanted tabs is easier, too. You can actually see what you’re about to close instead of relying on a tiny favicon and a bit of blind faith.</p><p>It’s a small quality-of-life improvement, but once you get used to it, going back feels surprisingly uncomfortable.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/353/0*7wcG9BVc6JXoNn7h.png" /></figure><h3>What If You Prefer the Traditional Layout?</h3><p>Not everyone needs vertical tabs all the time.</p><p>If you only have a handful of tabs open, the traditional horizontal layout might still feel more natural.</p><p>Fortunately, Chrome doesn’t force you to stick with one layout.</p><p>You can switch back to horizontal tabs whenever you want.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/788/0*aqzAllSPAoCtdKvL.png" /></figure><p>So you can use vertical tabs when you’re juggling dozens of pages and switch back when things are quiet.</p><p>No commitment required.</p><h3>A Small Change That Makes a Big Difference</h3><p>Browser tabs might seem like a pretty minor part of the user experience, but think about how often we interact with them every single day.</p><p>even if I use most of Ai tools, the browser still is one of my most import tools</p><p>We open them, close them, switch between them, search for them, and somehow keep opening more even when we know we shouldn’t.</p><p>For developers, researchers, and anyone who practically lives inside a browser, tab management isn’t just a cosmetic detail. It’s part of the daily workflow.</p><p>Chrome’s vertical tabs won’t magically fix your 50-tab problem</p><p>Let me be honest: you’ll probably just find a way to open 80 tabs instead.</p><p>But at least you’ll be able to find them.</p><p>And that, in itself, feels like progress.</p><p>Next time you find yourself lost in that overcrowded tabs at the top of your browser, try this method</p><p>Your mouse finger deserves a break.</p><p>last but last least, thank you for reading.</p><p>if you like it,you can give a clap or follow that will encourage me to have more confidence create more better content for you</p><p>see you next time</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f7e0bafe9442" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Rust-Powered IntelliJ IDEA Officially Dropped]]></title>
            <link>https://medium.com/@2084267015/rust-powered-intellij-idea-officially-dropped-b8f5c6fcc5e1?source=rss-844c6602f269------2</link>
            <guid isPermaLink="false">https://medium.com/p/b8f5c6fcc5e1</guid>
            <category><![CDATA[rust]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[intellij-idea]]></category>
            <category><![CDATA[productivity]]></category>
            <dc:creator><![CDATA[golang学习记]]></dc:creator>
            <pubDate>Tue, 03 Nov 2026 12:16:11 GMT</pubDate>
            <atom:updated>2026-11-03T12:16:11.965Z</atom:updated>
            <content:encoded><![CDATA[<h3>Weighing a Programmer’s “Age”: When You Open a Project, Does Your Fan Sound Like a Helicopter Taking Off?</h3><p>As someone who has spent years in the Java ecosystem, my relationship with IntelliJ IDEA is incredibly complicated. It truly gets me — its keymaps, its global search, its tool windows, are all ingrained in my muscle memory. But it’s also <em>heavy</em>. Every time I click that colorful little icon, I have to go brew a cup of coffee first. When the screen shows “Indexing…” and the progress bar crawls along, my 32GB of RAM seems to let out a desperate wail.</p><p>I often fantasize: What if IDEA could be as fast as lightning, like Zed or Alacritty?</p><p>Well, now this fantasy has been monetized (well, open-sourced) by a project called <strong>OpenIDEA</strong>.</p><h3>🚀 “The IDE You Know and Love, Rewritten in Rust”</h3><p>OpenIDEA has already racked up over 400 stars on GitHub just one week after open-sourcing.</p><p>Its ambition is incredibly bold: <strong>to port IntelliJ IDEA’s way of working (keymaps, tool windows, global search) exactly as it is, into a native application written in Rust and built on top of Tauri.</strong></p><p>In short, it wants to give you a replacement that “looks like IDEA, but runs like C.” You might think it’s just a VS Code clone with a new skin? No, it’s actually rewriting the wheel from scratch with Rust.</p><h3>📊 The Official “Brutal” Benchmark</h3><p>To prove it’s not just bluffing, the developer released a set of absolutely devastating benchmark comparisons (tested on a Python project called freqtrade containing 770 files).</p><p>Look at these brutal numbers (note the comparison target is IDEA 2026.2.3):</p><ul><li><strong>Opening the project window</strong>: OpenIDEA takes <strong>0.9 seconds</strong>, IDEA takes 3.1 seconds.</li><li><strong>Fully Ready (CPU stable)</strong>: OpenIDEA takes <strong>1.1 seconds</strong>, IDEA takes 13.2 seconds.</li><li><strong>Memory Usage (USS)</strong>: OpenIDEA takes a mere <strong>169 MB</strong>, while IDEA spikes to <strong>2011 MB</strong>.</li><li><strong>Quitting the app</strong>: OpenIDEA quits instantly (<strong>0.3s</strong>), IDEA takes 5.9 seconds.</li></ul><p>Looking at this data, I can only say: <strong>This isn’t optimization; this is dragging IDEA across the coals.</strong> What does 169MB of memory even mean? You probably use more than 169MB just opening a single webpage right now. Yet it carries a complete IDE with an editor, debugger, Git, terminal, and even a database client.</p><p>Out of curiosity, I downloaded OpenIDEA and used it for an afternoon, and the results were genuinely impressive.</p><p>For example, my most frequently used “Double Shift”:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*34-gPQkJmNCm9G-B.png" /></figure><p>Autocomplete from the LSP, which honestly feels even better than IDEA:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*CgnlFWbz8wafUUmW.png" /></figure><p>Finding where a symbol is used, almost exactly like IDEA:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*zZyqSlsNsfr4_5IG.png" /></figure><p>Running tests, with the familiar green button:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/785/0*onmlfvPIFgU7_J5p.png" /></figure><p>Breakpoint debugging, the experience is exactly the same as IDEA:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*TrTOTGMn6sA8_xHV.png" /></figure><p>It also has a terminal and an HTTP client. It’s worth noting that the HTTP client is usually a feature reserved for the paid version of IDEA, but OpenIDEA includes it — truly a kind gesture.</p><h3>🛠️ How Did They Do It?</h3><p>Looking at its codebase, you’ll find it’s a modern “Frankenstein” of tech:</p><ul><li><strong>Rust (33.9%)</strong>: Handles core logic and performance-sensitive parts.</li><li><strong>TypeScript (62%)</strong>: Handles UI and frontend interaction.</li><li><strong>Tauri Framework</strong>: Uses the system’s native WebView instead of the bloated Electron (yes, I’m looking at you, VS Code).</li></ul><p>It even supports over 60 languages, has built-in terminal themes (like Solarized and One series), and includes oxfmt and oxlint—the ridiculously fast next-gen frontend toolchains. The funniest part? If you look at the contributor list, nestled among the human developers is an AI named @claude.</p><h3>💡 Should You Ditch Your IDEA?</h3><p>Don’t rush. OpenIDEA is currently only at v0.1.22. Despite its 437 stars, it still has a long way to go before becoming a mature productivity tool. Its biggest significance right now is sending a wake-up call to JetBrains: <strong>In today&#39;s world of being squeezed by AI and lightweight editors from both sides, you can&#39;t just rest on your laurels by being &quot;heavy&quot; and &quot;memory-hungry.&quot;</strong></p><p>If you love IDEA’s shortcuts but are fed up with its startup speed, go clone l0ng-ai/openidea from GitHub and give it a try. After all, watching an open-source project use Rust to completely flip the script on a commercial giant—that&#39;s a &quot;spectator&quot; thrill unique to programmers that&#39;s hard to resist.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b8f5c6fcc5e1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[10x Lighter Than VS Code: JetBrains Open-Sources a Brand New AI IDE]]></title>
            <link>https://medium.com/@2084267015/10x-lighter-than-vs-code-jetbrains-open-sources-a-brand-new-ai-ide-81ab80a95731?source=rss-844c6602f269------2</link>
            <guid isPermaLink="false">https://medium.com/p/81ab80a95731</guid>
            <category><![CDATA[coding]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[web-development]]></category>
            <dc:creator><![CDATA[golang学习记]]></dc:creator>
            <pubDate>Sat, 31 Oct 2026 03:37:07 GMT</pubDate>
            <atom:updated>2026-10-31T03:37:07.268Z</atom:updated>
            <content:encoded><![CDATA[<p>If you’ve used Cursor or the VS Code Agent, you’ve probably suffered through this: the AI goes on a wild coding spree, your main repository gets completely mangled, and when things inevitably break, you’re scrambling to revert everything. It feels like hiring a temp worker to renovate your house — they finish the job, leave a massive mess behind, and you’re the one who has to clean it up.</p><p>A few days ago, JetBrains launched a brand new AI IDE called <strong>ThinkRail</strong>, attempting a completely different approach to solve the pain points of Agent-driven development.</p><p>Unlike JetBrains’ previous AI tool “Air,” which was built on the heavy IntelliJ platform, ThinkRail is actually lightweight. The editor core is built directly on <strong>Monaco</strong> (the engine behind VS Code), and the underlying architecture is powered by a coding Agent engine called pi. The UI is just a thin shell. It doesn&#39;t secretly stitch together Prompts; model management, session compression, and tool orchestration are all handled by pi.</p><p>It’s also not like IDEA, where AI is treated as just another plugin. In ThinkRail, <strong>AI is the true protagonist.</strong></p><p>The most brilliant trick up its sleeve is <strong>Git worktree isolation</strong>. Traditional AI IDEs read and write code directly in your current working directory. If the Agent screws up, it directly corrupts your source code. ThinkRail creates a new, independent Git worktree for every Agent task, complete with its own dedicated branch and runtime directory.</p><p>Let’s take a look at this brand new IDE.</p><p>Even though it’s just been released, ThinkRail already supports installation on Windows, macOS, and Linux. You can download and install it from GitHub or their official site at thinkrail.ai.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/934/0*jVMfd2KOTdCzNByo.png" /></figure><p>After installing it, I realized how truly lightweight ThinkRail is — only about 40MB. Compared to VS Code’s 400MB+, it’s 10 times lighter.</p><p>Next, we arrive at the ThinkRail homepage.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ecWpyZrA2-4QaIn0.png" /></figure><p>It’s the first time I’ve seen an AI IDE this clean — just a single “Open Project” button right in the middle.</p><p>After opening a project, you’ll see this interface:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Tr-xlbtRCmQcIv2i.png" /></figure><p>Next, you can click on “Default.” This represents the default workspace, which embodies the main feature mentioned earlier: <strong>one workspace = one Git worktree.</strong></p><p>Once you click it, you’ll see the following interface:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*xF5aYYb8ExBW3XN_.png" /></figure><p>The core area is the chat, which is where you talk to the AI. Before you start chatting, you need to configure a provider.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*QbhBH-bFKBmMPDcE.png" /></figure><p>It also supports OpenRouter, allowing you to use free API keys. If you have a DeepSeek API key, you can just copy and paste it in to use.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/965/0*pdFAusr6pEWiE9yV.png" /></figure><p>Here is how I configured it with DeepSeek:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/996/0*_NLXJOBD7RI1Pk0X.png" /></figure><p>At the very bottom is the familiar terminal.</p><p>On the far right, there are sections for “spec” and “AI changes.” The “spec” part highlights ThinkRail’s main push: <strong>spec-driven development.</strong></p><p>Click the file icon on the right, and the project files will appear.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/629/0*Am7pdL_DY-fcCrHC.png" /></figure><p>We can also select a chunk of code to bring up an inline chat, similar to the VS Code Agent.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/828/0*ATUym-5HP-RYM10N.png" /></figure><p>Clicking “Command Palette” brings up the familiar command palette from VS Code.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1017/0*g_0SHPOpx1k664wP.png" /></figure><p>If you want to develop a new feature, you can click “Default” in the top left corner to add a new workspace.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/740/0*oSOhUQ0Gsf97iE7s.png" /></figure><p>Just like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/890/0*1kNjbk8QJxAJmfJ5.png" /></figure><p>It implemented it very quickly.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*uscoXJ78pgDS2rqj.png" /></figure><p>Click “Review turn” to see the changes made this time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/953/0*qxtfSrg_Dd0iOTsj.png" /></figure><p>Of course, it’s currently an incubated preview project under the InnovationHub. It’s not meant to replace official products like IDEA or PyCharm. It focuses on a browser-based UI, but also offers desktop installers for Windows/macOS/Linux. It’s completely open-source under the Apache-2.0 license.</p><p>Compared to the sea of AI editors on the market: others focus on the editor and treat AI as an add-on plugin. ThinkRail flips the script: <strong>the Agent is the core, and the editor is merely a window for you to watch the Agent work.</strong></p><p>Whether it can outrun Cursor and Zed remains to be seen, but the design philosophy of “don’t let the Agent pollute the main codebase” has already struck a chord with the real pain points of many developers.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=81ab80a95731" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Just Now，VS Code’s Website Got a Silent Update]]></title>
            <link>https://medium.com/@2084267015/just-now-vs-codes-website-got-a-silent-update-d085d7f92be7?source=rss-844c6602f269------2</link>
            <guid isPermaLink="false">https://medium.com/p/d085d7f92be7</guid>
            <category><![CDATA[vscode]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[programming]]></category>
            <dc:creator><![CDATA[golang学习记]]></dc:creator>
            <pubDate>Sat, 31 Oct 2026 03:29:14 GMT</pubDate>
            <atom:updated>2026-10-31T03:29:14.338Z</atom:updated>
            <content:encoded><![CDATA[<p>Let me ask you a soul-searching question: How long has it been since you actually <em>looked</em> at the VS Code official website?</p><p>For veteran netizens like me, code.visualstudio.com is like the refrigerator at home—as long as it still keeps things cold (or in this case, still lets me download the installer), I couldn&#39;t care less if the exterior has turned yellow.</p><p>I open VS Code every day, but I rarely visit its official website. Usually, I only go there to check the docs or read the latest release notes.</p><p>A few days ago, I dropped by to look up a keyboard shortcut, and I froze on the spot: <strong>Is this still the dark, “straight-male-aesthetic” website I used to know?</strong></p><p>This update was completely silent. No official announcement, no option to revert to the old version — they just swapped it out on you. But after taking a closer look, I genuinely don’t want to go back.</p><p>Here is what changed.</p><h3>1. No More Getting Blinded: Seamless Light/Dark Mode</h3><p>The old website was a bit “aloof,” defaulting to a light background. Staring at it during the day felt like making your eyes work overtime. Sure, there was a moon icon to switch to dark mode, but it felt like an afterthought.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/448/0*dnkutb4AQc97isaq.png" /></figure><p>Now, the new version <strong>natively follows your system’s light/dark theme</strong>. The little sun/moon toggle in the top right is gone — it just adapts.</p><p>Personally, I prefer setting my system to dark mode.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*o_Gx8Vnay8Ypt9r9.png" /></figure><p>And here is what the new homepage looks like:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*2ZVCk9ZbdHfeZ5q7.png" /></figure><p>Not only that, but the entire UI has become more rounded, with more whitespace. The large banner screenshots have been replaced with more refined versions. It used to look like a rough manual; now, it has a hint of Apple-like lightweight elegance.</p><h3>2. Saving the “Archaeologists”: The Blog Finally Has an Archive</h3><p>In the past, if I wanted to find a deep-dive article on VS Code written six months ago, clicking into the blog would just throw me straight to the latest post. <strong>Where did the old articles go? The answer: lost at sea.</strong> No directory, no list — digging through history was more painful than writing Assembly.</p><p>The new version rolls out an <strong>independent blog homepage</strong>:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*XsCE8zw200bpoZ9X.png" /></figure><p>And it uses a card-based layout:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*yY09Q0ADHQddXZik.png" /></figure><p>However, VS Code opted not to use pagination to display the articles, which I’m personally not a huge fan of.</p><h3>3. Documentation Navigation: AI Finally Finds a Home</h3><p>This is the best part for those who write tutorials or frequently read the docs.</p><p>Previously, reading the documentation meant dealing with a sidebar where languages and plugin development categories were all piled up. It was densely packed, and finding something felt like playing a game of “spot the difference.” As for AI-related content? It was scattered all over the place, here, there, and everywhere.</p><p>The new version puts sub-modules like documentation into the “Resource” tab:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/795/0*B-fDEmefBAbhhZ6K.png" /></figure><p>The left-hand directory is much clearer now:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/367/0*TZiIl2HXh8ANN3Ms.png" /></figure><p>Finding things is much clearer now — follow the trail, and you won’t get lost.</p><h3>4. The Features Page</h3><p>The old website didn’t have a Features page. In the new version, it’s placed directly in the first tab on the left.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/468/0*sd-oicfR45AWzc9h.png" /></figure><p>You can see that VS Code is being very smart here. It divides its features into two blocks: one is AI development, and the other is the traditional editor. It even dropped the word “code” from the title.</p><p>Let’s start with the editor part, which we are all most familiar with.</p><p>Here is the homepage for it:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*QSBlGiuM4RDYY-DS.png" /></figure><p>It emphasizes the fundamental capabilities of using VS Code for editing: Edit, Build, Debug.</p><p>Then it specifically highlights the most commonly used features:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*rxAASEfFPAlcTdEq.png" /></figure><p>It also specifically mentions the terminal later on:</p><h3>5. The Soul-Searching Question: What is the Homepage Actually Selling?</h3><p>If you compare the copywriting of the new and old homepages, you’ll realize that VS Code’s “persona” has completely changed.</p><p>The <strong>old version</strong> used to say: “Look, I’m lightweight, I’m open-source, I’m an AI code editor.” The <strong>new version</strong> slaps its face right on the screen: “Look, I have Multi-agent, I have Copilot, I am the ultimate AI coding workbench!”</p><p>AI has gone from being a secondary selling point to taking center stage. The download button has also become incredibly prominent, fixed in the top right corner, as if they’re terrified you won’t find it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/778/0*YWuW2iIvj_ZYiX-H.png" /></figure><p>The other block is dedicated to agents, which is a key section for VS Code:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*JmFoXzWhZfy_9a6c.png" /></figure><p>The three main stars in here are the Agents window, the Terminal, and the Editor window, which takes third place:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/856/0*Q88LlaUJ_c6_nAUB.png" /></figure><p>Finally, it’s worth highlighting the newly added global search. Click the search button to search across all content on the VS Code website:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/508/0*16bD3WQ4d26C1h8q.png" /></figure><p>VS Code uses a tool called <strong>docfind</strong> for search. Even if you misspell a word, it can still locate the document. It’s incredibly useful.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/918/0*AN-AvXuYatpGthGe.png" /></figure><h3>Summary: From “Editor” to “AI Workbench”</h3><p>What hasn’t changed is the domain — it’s still code.visualstudio.com, and the download link and Marketplace are all still there.</p><p>But its soul has evolved. <strong>The old website was a pure “code editor website,” while the new website has completely transformed into an “AI intelligent coding workbench website.”</strong></p><p>Moreover, the overall structure is much clearer, making it easier for VS Code users to understand the product.</p><p>If you have some free time, I highly recommend taking a stroll through the new website. After all, even Microsoft itself is saying that if you can’t write a couple of prompts nowadays, you shouldn’t even call yourself a programmer.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d085d7f92be7" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>