kk

I want to use body filters to find out hidden parameters for api request in html or javascript when debug crawler code. But https://github.com/mitmproxy/mitmweb/issues/10 has been archived, the feature not...

@guyskk I got example at https://docs.mitmproxy.org/stable/addons-examples/#io-read-saved-flows , here is my modified code to filter response body. ```python import sys from mitmproxy import http, io def check_flow(flow: http.HTTPFlow, target: str): text...

这个也许有用：https://ask.dcloud.net.cn/question/91096 ```js // vue.config.js里面加上这个试试 module.exports = { configureWebpack: { resolve: { symlinks: false } } } ```

@euri10 👌 here is the PR: https://github.com/encode/uvicorn/pull/1288

> Adding `propagate = False` is not a fix. It turs off all uvicorn's logs. @gretkierewicz That's not correct, `propagate = False` just stop propagate logs to ancestors, the log...

@euri10 I see, will check on it.

Hi @euri10, the failures is because pytest caplog not able to capture `propagate=False` logger, see also: https://github.com/pytest-dev/pytest/issues/3697 And the [caplog_for_logger](https://github.com/encode/uvicorn/blob/master/tests/middleware/test_logging.py#L13) helper also not work on test config cases, because when...

Yes, the caplog_for_logger works when test logs after create Config object, but it can not capture logs inside `Config.__init__`. eg: `test_should_warn_on_invalid_reload_configuration`, `test_reload_dir_is_set` and `test_non_existant_reload_dir_is_not_set`

`docker rmi --no-prune ` 用这个命令删除，既可以回收空间，又可以保留层给下次打包复用。 @ChanDaoH @vangie

从 funcraft 过来的，其实 funcraft 用的挺好的，支持多云平台肯定是好事。我从用户的角度，觉得这些比较重要： 1. 稳定可靠：Serverless 本身不算成熟，作为生产工具必须把稳定可靠放在第一位，要帮用户把坑先填好，有坑的地方有提醒和可能的解决方案。 2. 抹平平台差异：如果不能抹平平台差异，就等于用户要为每个平台写一套配置，要去踩各个平台的坑。而且很显然不能直接把阿里ROS规范搬过来，否则别的平台不跟你玩，强行适配也会非常痛苦。 3. 关注业务场景，给业务带来更多价值，避免追求高大上、新奇的技术。例如冷启动时长，打包体积优化，如何支持更大的代码包，自动化部署，密钥管理，这些都是很需要解决的。 4. 没必要搞 k8s 和 Terraform，用 serverless 就是为了免运维，为了低成本的用阿里云腾讯云服务。把阿里云，腾讯云，AWS 做好、做完善就够了。如果公司有 k8s 运维，大家肯定就直接用 kubectl 了，何必再包一层 serveless 呢。