你看了什么帖子

@oooolongtea 没用 kindle 了

@chztv
@seanzxx 问题是不看虚拟内存，就看活动监视器显示的内存，也占了 4G 多，这也很大吧

@oooolongtea 用得感觉挺好的，没遇到品控问题，而且我是实用主义的，对品控也不在乎

我是 git log 显示中文乱码

@RiverMud 说是 2024 年 6 月 30 才停止服务

13 寸哪来的 m2max

我 y 般 b 用 dou 阴

@paynezhuang 已经尝试过但是没用

@imnpc opencore 的貌似是存储在 plist 文件里面的吧，我后来重新关闭 SIP 也是改 plist

@chouchen 有什么更好的替代呢

你眼睁睁看着蟑螂通过后盖的缝隙钻进去了？

@meishiwanwan520 你的意思是白斑作为性能故障在保内换机吗？我的意思是我已经过保了，但是出现了白斑。想换电池来换整机，不知道会不会因为有白斑而被拒绝

@yankebupt

你说的这个是你自己想的吧。我的意思是第一个筷子先阻塞式锁，第二个筷子非阻塞锁。你的说法是第一个筷子也非阻塞锁。

不过感觉先锁桌子，然后左右 tryLock 也是可以的。

@lessMonologue 笑死人，已 block😁

@Nerv 那你的意思是对凡是已经有定论的的东西都绝对不要提出个人想法是么

@Nerv 来 v 站发帖不就是为了看大家意见从而学习吗？光思考的话为什么要发帖子呢？我认为自己是对的不就行了

我总结一下死锁和活锁的对比，以下来自于《 C++ Concurrency in Action 》

1.死锁没法救活，除非编程时就避免。这一点活锁有优势，因为活锁就算你编程时不避免，由于 CPU 的调度存在一定的偏差，也有可能可以自行脱离活锁。（可以看引用的“not-so-serious cases”）
2.死锁发生时因为是等待，所以不会自旋，不占 CPU ，这方面是比活锁好的。


> Deadlock - As you saw in chapter 3, in the case of deadlock, one thread is waiting for another, which is in turn waiting for the first. If your threads deadlock, the tasks they're supposed to be doing won't get done. In the most visible cases, one of the threads involved is the thread responsible for the user interface, in which case the interface will cease to respond. In other cases, the interface will remain responsive, but some required tasks won't complete, such as a search not returning or a document not printing.

> Livelock - Livelock is similar to deadlock in that one thread is waiting for another, which is in turn waiting for the first. The key difference here is that the wait is not a blocking wait but an active checking loop, such as a spin lock. In serious cases, the symptoms are the same as deadlock (the app doesn't make any progress), except that the CPu usage is high because threads are still running but blocking each other. In not-so-serious cases, the livelock will eventually resolve because of the random scheduling, but there will be a long delay in the task that got livelocked, with a high CPU usage during that delay.