Posts

Recently I tried to optimize the startup of a Spring Boot 4 web application with JVM AOT Cache. The application itself is a normal Spring Boot web application. It uses PostgreSQL, JPA, Firebase Auth, and is built into a container image through bootBuildImage in GitHub Actions. I previously considered more aggressive startup optimizations, but adopting GraalVM Native Image requires a lot changes for my project and the benifit doesn’t worth it (for now). ...

Recently I reviewed the intern() method of Java String, which is a small method but sometimes makes string comparison questions quite confusing. In Java, string literals are special. For example: String a = "hello"; String b = "hello"; System.out.println(a == b); // true The result is true because both a and b refer to the same string literal in the string pool. However, if we create a string with new, the situation is different: String a = "hello"; String b = new String("hello"); System.out.println(a == b); // false System.out.println(a.equals(b)); // true The content of a and b is the same, so equals() returns true. But they are not the same object, so == returns false. This is where intern() can be used. ...

Recently I took a Next.js project from “source code in a repo” to a fully automated delivery path on AWS: CodeCommit → CodePipeline/CodeBuild → ECS (behind an existing ALB), provisioned with Terraform. I made some decisions regarding the architecture: remote state in S3, a centralized Terraform repository, dev/prod separation, IAM aligned with repository layout, and a practical approach to environment variables across build-time and ECS runtime. I used S3 with versioning enabled and tight bucket policies to store Terraform state. This is boring infrastructure, and that’s the point: it prevents high-cost incidents later. ...

Recently I modernized a YOURLS instance running on an aging stack: YOURLS 1.7.x + MySQL 5.7 + CentOS due to several risks: Security exposure: from older OS / packages Operational friction: (harder patching, fewer compatible components) Direct cost: continuing with an old managed MySQL version meant paying extended support fees To address those issues, I migrated to a clean, current setup: Latest YOURLS MySQL 8.4 Latest Amazon Linux Added Cloudflare Turnstile to reduce bot traffic and abuse I intentionally avoided “in-place upgrades”. Instead, I treated it as a rebuild + data migration, which tends to be safer and easier to roll back. ...

I decided to migrate my old blog from Jekyll to Hugo framework. Also I will use Cloudflare Pages over Github Pages. No special reasons, just for fun. It’s quite easy to build with Hugo & PaperMod. Cloudflare Pages is also extremely easy to set up. The structure of Jekyll posts is pretty similar to Hugo posts so a simple script is enough to migrate all of my previous posts. Overall I’m satisfied with the new blog, maybe I will add commenting system in the future.

This command searches all .log files which contains the target message in the current directory. It prints the filename, the matching line and the 30 lines that follow the matching line. grep -H -A 30 "target error message" *.log This command locates every line that contains the target message (-i here is for ignoring case distinctions) and prints it, along with the 30 lines that immediately follow each matching line. | less provides paging functions. grep -i -A 30 "target error message" sample.log | less tail -f starts continuously outputting new lines as they are written to the specified file. grep then, in real-time, scans these incoming lines and print necessary logs. This process continues indefinitely until we manually stop the command. tail -f sample.log | grep -A 30 "target error message" This command counts the occurrences of the target error message in the sample.log. grep -c "target error message" sample.log This command prints 30 lines of context both before and after the target message in the sample.log. grep -C 30 "target error message" sample.log Nowadays we have more useful tools such as ELK (Elasticsearch, Logstash, Kibana) stack. The configuration is more complicated so I guess I will add those contents in the future :)

DP code sample. Just a quick refresh. class Program { static void Main() { int goal = 97; int[] coins = {10, 25, 5, 1}; int []table = new int[goal + 1]; table[0] = 0; for (int i = 1; i <= goal; i++) { table[i] = int.MaxValue; } for (int j = 1; j <= goal; j++) { for (int k = 0; k < coins.Length; k++) { int selectedCoin = coins[k]; if (selectedCoin <= j) { table[j] = int.Min(table[j], table[j - selectedCoin] + 1); } } } foreach (var elem in table) { Console.Write(elem); Console.Write(" "); } Console.WriteLine(); Console.WriteLine(table.Last()); } }

Steps may change in the future. sudo snap remove firefox Follow https://support.mozilla.org/en-US/kb/install-firefox-linux#w_install-firefox-deb-package-for-debian-based-distributions However, in step 5: onfigure APT to prioritize packages from the Mozilla repository, use the command below: echo ' Package: * Pin: origin packages.mozilla.org Pin-Priority: 1001 Package: firefox Pin: version 1:1snap* Pin-Priority: -1 ' | sudo tee /etc/apt/preferences.d/mozilla-firefox sudo apt install firefox To ensure that unattended upgrades do not reinstall the snap version of Firefox: echo 'Unattended-Upgrade::Allowed-Origins:: "namespaces/moz-fx-productdelivery-pr-38b5/repositories/mozilla:mozilla";' | sudo tee /etc/apt/apt.conf.d/51unattended-upgrades-firefox

Minimal API is a fairly new way to implement Web APIs, introduced in .NET 6. I initially thought it was only suitable for small projects, as its name suggests. Recently, I read some articles and watched videos about how to organize Minimal API projects, and I have changed my mind. The main reason I believed Minimal API was only for small projects is that most demos are quite small and put all APIs in Program.cs, which is obviously not scalable and somewhat messy. By using extension methods, it is quite convenient to group APIs into different files, making them much easier to manage. A third-party library Carter also seems useful in grouping APIs. ...

Use escape sequences for special signs such as \n. Or use a verbatim statement // Escape sequences example Console.WriteLine("I need to use escape sequences for special signs such as \\ here."); // Verbatim statement example Console.WriteLine( @" My name is ""SilMOON"" and this is a verbatim statement. I can use special signs such as \ here. " ); Noted that C#11 introduced raw string literals which is more convenient. I’m actually suprised that they didn’t add this feature earlier. :) ...