[U] Update channel content
This commit is contained in:
github-actions
+10
-1
@@ -2,11 +2,20 @@
|
||||
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="zh-cn">
|
||||
<id>https://aza.moe/life</id>
|
||||
<title>小桂桂的回忆录 📒</title>
|
||||
<updated>2024-11-24T04:11:22.995704+00:00</updated>
|
||||
<updated>2024-11-24T06:11:59.779467+00:00</updated>
|
||||
<link href="https://aza.moe/life" rel="alternate"/>
|
||||
<generator uri="https://lkiesow.github.io/python-feedgen" version="0.9.0">python-feedgen</generator>
|
||||
<logo>https://aza.moe/meru_256px.png</logo>
|
||||
<subtitle>「我们所经历的每个平凡的日常，也许就是连续发生的奇迹」</subtitle>
|
||||
<entry>
|
||||
<id>4683</id>
|
||||
<title>小桂桂的回忆录 📒 #4683</title>
|
||||
<updated>2024-11-24T05:52:34+00:00</updated>
|
||||
<content><p>唉 这就是 AI hype 的 nonsense 吗</p>
|
||||
<p><a href="https://arxiv.org/abs/2308.09597">ChatHaruhi</a> 这篇没写完的论文，结果评估都没有，居然就发 preprint 了，居然已经有 15 citation 了... <i class="custom-emoji" emoji-src="emoji/6170424195260877091.webm">🤗</i></p>
|
||||
<p>非常可惜我也会给它贡献一个 citation</p></content>
|
||||
<link href="https://aza.moe/life?post=4683" rel="alternate"/>
|
||||
</entry>
|
||||
<entry>
|
||||
<id>4682</id>
|
||||
<title>小桂桂的回忆录 📒 #4682</title>
|
||||
|
||||
File diff suppressed because one or more lines are too long
Binary file not shown.
|
After Width: | Height: | Size: 28 KiB |
Binary file not shown.
|
After Width: | Height: | Size: 1.1 KiB |
@@ -64156,7 +64156,7 @@
|
||||
"id": 4670,
|
||||
"date": "2024-11-20T15:44:45",
|
||||
"text": "呜呜 当时写 AquaDX 的时候选择了「先用最方便快捷的方式把网页做出来，等有问题的时候再优化」，然后就随手写出了一个请求需要把数据库两个表完整拉下来的代码 <i class=\"custom-emoji\" emoji-src=\"emoji/6323267916523374069.webp\">🥺</i>\n\n然后现在真的有问题了，现在请求一次排行榜要拉 1452 个用户和 221448 条游玩记录... 优化之前要整整 30 秒，稍微加了缓存和 batching 优化了一下也依然要十秒 <i class=\"custom-emoji\" emoji-src=\"emoji/6323582802050680516.webp\">🛌</i>",
|
||||
"views": 75,
|
||||
"views": 77,
|
||||
"forwards": 0,
|
||||
"reply": {
|
||||
"id": 3430,
|
||||
@@ -64180,7 +64180,7 @@
|
||||
"id": 4671,
|
||||
"date": "2024-11-21T03:25:16",
|
||||
"text": "今天花了一整天调查这个数据库问题， Hibernate 全责 <i class=\"custom-emoji\" emoji-src=\"emoji/6323582802050680516.webp\">🛌</i> 以后的项目再也不用 Java EE 了\n\n我试了一下把这整个函数用 native mysql 写出来，发现这条大查询只用 1.19 秒，那为什么我的 kotlin 代码运行速度差这么大呢？\n\n开了 SQL debug 之后我发现即使只是 .findAll() 都要 N 条查询，按理来说 select * from table 只要一条查询就行啊，为什么呢？仔细看看这 N 条查询发现它把 OneToOne fk 都一个一个拉下来了。原本首先花了一条查询 SELECT * FROM parent，然后又花了 N 条查询拉 SELECT * FROM child WHERE id=i...\n\n查了一下这个问题，网上说开 Hibernate bytecode enhance 就好了，它会修改编译完的字节码让 fk 链接用到的时候再查询而不是预先把所有东西查询好。问题是它不太兼容 kotlin，绕过了三个不同的问题才成功编译，之后加上 lazy fetch 终于可以让它在用到 fk 的时候再拉数据了... \n\n但是即使这样依然要把需要拉所有 fk 数据的地方改成手动 mysql 查询，因为 orm/jpql 没办法让 findAll 把单独的查询优化成 join <i class=\"custom-emoji\" emoji-src=\"emoji/6323582802050680516.webp\">🛌</i>\n\n唉 感觉凡是透明一点的 ORM 比如 ktorm 或者 exposed 都不会有这样的问题，以后不会再用 Java EE 了",
|
||||
"views": 77,
|
||||
"views": 82,
|
||||
"forwards": 0,
|
||||
"media_group_id": 13857276129125749,
|
||||
"images": [
|
||||
@@ -64373,7 +64373,7 @@
|
||||
"id": 4682,
|
||||
"date": "2024-11-23T01:14:05",
|
||||
"text": "时隔一年半的后续：今天 Zotero 告诉我说之前遇到的这篇造假的论文被 retract 了，可喜可贺\n\n看到 Retract notice 里面指出了一个 \"Paper Mill: A business that creates and sells research papers where the buyer(s) claim authorship; the business is generally unacknowledged as a contributor.\" ...居然这个世界上还有这种业务存在，好可怕。",
|
||||
"views": 63,
|
||||
"views": 66,
|
||||
"forwards": 1,
|
||||
"reply": {
|
||||
"id": 2202,
|
||||
@@ -64391,5 +64391,24 @@
|
||||
"thumb": "media/4682.jpg_thumb.jpg"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"id": 4683,
|
||||
"date": "2024-11-24T05:52:34",
|
||||
"text": "唉 这就是 AI hype 的 nonsense 吗\n\n<a href=\"https://arxiv.org/abs/2308.09597\">ChatHaruhi</a> 这篇没写完的论文，结果评估都没有，居然就发 preprint 了，居然已经有 15 citation 了... <i class=\"custom-emoji\" emoji-src=\"emoji/6170424195260877091.webm\">🤗</i>\n\n非常可惜我也会给它贡献一个 citation",
|
||||
"views": 22,
|
||||
"forwards": 0,
|
||||
"images": [
|
||||
{
|
||||
"width": 534,
|
||||
"height": 244,
|
||||
"date": "2024-11-24T05:52:34",
|
||||
"media_type": "photo",
|
||||
"original_name": "photo_2024-11-24_05-52-34.jpg",
|
||||
"url": "media/4683.jpg",
|
||||
"size": 28671,
|
||||
"thumb": "media/4683.jpg_thumb.jpg"
|
||||
}
|
||||
]
|
||||
}
|
||||
]
|
||||
@@ -12,7 +12,15 @@
|
||||
<link>https://aza.moe/life</link>
|
||||
</image>
|
||||
<language>zh-cn</language>
|
||||
<lastBuildDate>Sun, 24 Nov 2024 04:11:22 +0000</lastBuildDate>
|
||||
<lastBuildDate>Sun, 24 Nov 2024 06:11:59 +0000</lastBuildDate>
|
||||
<item>
|
||||
<title>小桂桂的回忆录 📒 #4683</title>
|
||||
<link>https://aza.moe/life?post=4683</link>
|
||||
<description><p>唉 这就是 AI hype 的 nonsense 吗</p>
|
||||
<p><a href="https://arxiv.org/abs/2308.09597">ChatHaruhi</a> 这篇没写完的论文，结果评估都没有，居然就发 preprint 了，居然已经有 15 citation 了... <i class="custom-emoji" emoji-src="emoji/6170424195260877091.webm">🤗</i></p>
|
||||
<p>非常可惜我也会给它贡献一个 citation</p></description>
|
||||
<guid isPermaLink="false">4683</guid>
|
||||
</item>
|
||||
<item>
|
||||
<title>小桂桂的回忆录 📒 #4682</title>
|
||||
<link>https://aza.moe/life?post=4682</link>
|
||||
|
||||
Reference in New Issue
Block a user