search

dyweb / weekly

DevComm-Shanghai Weekly 上海地区高校技术社团联合周报（欢迎投稿）
https://dyweb-weekly.netlify.com/
66 stars 5 forks source link

Weekly-103 #88

Closed gaocegege-bot closed 5 years ago

gaocegege-bot commented 6 years ago

联合周报第 103 期开始投稿 :tada:

at15 commented 6 years ago

yet another npm storyhttps://github.com/dominictarr/event-stream/issues/116 from @htfy96 twitter and I am just a carrier for @arrowrowe (didn't use the link format to avoid gh show issue reference on the other end)

gaocegege commented 6 years ago

gorilla/sessions provides cookie and filesystem sessions and infrastructure for custom session backends.

https://github.com/gorilla/sessions

at15 commented 6 years ago

@gaocegege 这个真心不怎么样 ... gorilla 下的除了 websocket (我没写过) 其他我觉得都很一般 ... 只是因为出来早, 性能和设计都很差 ....

at15 commented 6 years ago

https://aws.amazon.com/blogs/opensource/firecracker-open-source-secure-fast-microvm-serverless/ https://github.com/firecracker-microvm/firecracker (written in rust) also has adapter for containerd https://github.com/firecracker-microvm/firecracker-containerd

seems just a daemon to wrap KVM w/ REST API?

gaocegege commented 5 years ago

Mundane is a Rust cryptography library backed by BoringSSL that is difficult to misuse, ergonomic, and performant (in that order).

https://github.com/google/mundane

@codeworm96 即将要贡献代码的项目

gaocegege commented 5 years ago

https://github.com/firecracker-microvm/firecracker Secure and fast microVMs for serverless computing.

如何评价aws开源的firecracker，它对接下来的虚拟机、容器发展方向带来什么变化吗？

codeworm96 commented 5 years ago

https://github.com/facebookexperimental/MIRAI

Mirai is an abstract interpreter for the Rust compiler's mid-level intermediate representation (MIR). It is intended to become a widely used static analysis tool for Rust. The initial focus will be on taint analyis.

Cool.

at15 commented 5 years ago

https://github.com/google/wire Compile-time Dependency Injection for Go used by https://github.com/google/go-cloud saw it in meetup https://github.com/enocom/presentations/tree/master/wire

xplorld commented 5 years ago

https://github.com/dotnet/coreclr/blob/b3475eb87a5f28fb88cf23ebe30f174eff75bb46/Documentation/design-docs/GuardedDevirtualization.md?utm_source=wechat_session&utm_medium=social&utm_oi=27912371699712

瞻仰大佬是怎么做编译器优化的

htfy96 commented 5 years ago

@xplorld 这个东西对应的是gcc的-fdevirtualize-speculatively。这方面如果不算speculative的话，clang前两年有个slides讲得很好（是个bachelor大佬做的），spec似乎我记得只有gcc做了？

at15 commented 5 years ago

https://github.com/objectbox/objectbox-go ObjectBox is an easy to use, object-oriented lightweight database and a full alternative to SQLite.

htfy96 commented 5 years ago

Eat My Data: How Everybody Gets File IO Wrong (2008)

发现写的raft老是丢数据……读了这个才发现以前对“落盘”的理解都不对（

https://www.slideshare.net/nan1nan1/eat-my-data

xplorld commented 5 years ago

https://crawshaw.io/blog/utf7

imap 协议默认用的是一种 modified utf7

at15 commented 5 years ago

@htfy96 洛神在写什么的 raft? lab 还是实际的系统?

htfy96 commented 5 years ago

@at15 缩水版lab（

at15 commented 5 years ago

https://increment.com/documentation/ stripe 的人讲怎么写文档 (公司内部 & 开源) 有一系列文章

其中一篇 notes-on-the-synthesis-of-labyrinths 里面提的很好的一点是不应该只讲最好的解决方法，而是应该给出完整的 context。实际的开发流程往往是一个图的 dfs (打开 google 然后一路点进去，遇到错路再退回来) 而很多文档和文章只是给出一个线性的流程，这样大部分读者很难对里面描述的东西感到信服，也不能把里面的经验用到自己的开发当中。(不过也因项目而易，有的项目光 get start 就能让人看很久了 ...)

现在越来越觉得维护一个好的文档是十分重要的, 能想清一个问题并且把各种方法的利弊讲出来其实比写出恰好能够工作的代码更重要 (即使是只有自己开发和自己使用的坑(伤感 ...ry))。当然上班的时候更多还是就写一些明明知道以后会坑别人的代码, 毕竟我已经被坑的麻木了 ..

tq5124 commented 5 years ago

提供一个思路：Design doc应该由大组编写一个template，我见过的常见section有：

  1. project staus (eg. draft, reviewed, implementing, experimenting, finalized).
  2. Contact info (eg. Eng, PM).
  3. TL;DR
  4. Background
  5. Objection (including goal, non-goal).
  6. Design Overview
  7. Design Details
  8. Other design options (and why we don't do it).
  9. Testing plan
  10. Rollout plan
  11. Q&A

感觉这样每个Eng都会cover这些design options，然后每一份design都会收录在组里文档目录里，方便后面的人review。

Add a link to keep the content in weekly: https://google.com

gaocegege commented 5 years ago

tq 学长的 section 基本涵盖了我读过的 design proposal 里的所有内容

gaocegege commented 5 years ago

https://github.com/wowlusitong/awesome-education

学生福利-收集支持教育优惠的软件与服务

at15 commented 5 years ago

https://live.bilibili.com/4630260 tidb hackthon 有人做了 timeseries with TiKV (一天速成 rust 和失败 ...

xplorld commented 5 years ago

http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2018/p1190r0.html

herb：6个operator太多了，我们搞一个 spaceship operator 抄一抄 java 本文作者：可是这样 operator!= 就会变慢，比如两个不等长的 vector 本可以直接返回 true 现在就得找一下是大于还是小于

然后提了100种听起来都怪怪的解决方案。为啥不直接允许用户重写 operator== 然后比 spaceship 优先级高呢？

xplorld commented 5 years ago

Eat My Data: How Everybody Gets File IO Wrong (2008)

发现写的raft老是丢数据……读了这个才发现以前对“落盘”的理解都不对（

https://www.slideshare.net/nan1nan1/eat-my-data

tl;dr Use fsync

at15 commented 5 years ago

from https://github.com/zenany/weekly/commit/ce8caf00a487942881333cdd5a56b80e3e53e9f2

https://github.com/elsewhencode/project-guidelines 主要是针对 js 项目的，没什么特别新的东西，但是能完全做到(还是得看队友 ....

at15 commented 5 years ago

from https://github.com/zenany/weekly/commit/ce8caf00a487942881333cdd5a56b80e3e53e9f2

https://thecreativeindependent.com/guides/how-to-balance-full-time-work-with-creative-projects/ 虽然不是讲软件的但是还是讲的很好的, 比如拒绝周末跟别人出去玩(这点可以丧起来 @xplorld ) done is better than perfect etc... 当然看文章总是容易，做到里面说的就还是需要很大的毅力了 ....

xplorld commented 5 years ago

虽然不是讲软件的但是还是讲的很好的, 比如拒绝周末跟别人出去玩(这点可以丧起来 @xplorld ) done is better than perfect etc... 当然看文章总是容易，做到里面说的就还是需要很大的毅力了 ....

@at15 并没有这种担忧，周末么有人和我玩的。

我主要是不知道做啥 project 好。。

at15 commented 5 years ago

from https://news.ycombinator.com/item?id=18508284 a mac ascii art software, no longer under develop though, it's now $9 https://monodraw.helftone.com/

htfy96 commented 5 years ago

@xplorld fsync的语义也是奥妙重重……在Mac上fsync也没用（

https://wiki.postgresql.org/wiki/Fsync_Errors

gaocegege commented 5 years ago

https://github.com/ithinco/i-am-chinese-the-dragonfly-must-go-on

我其实不是太懂，搜索引擎也不是只有百度啊

xplorld commented 5 years ago

@gaocegege 意思是百度太恶心，bing 太烂，duckduckgo 国内没听过