仪器结果自动解析
我正在使用 Jenkins 和 xcodebuild 命令行工具自动化我的 iOS 项目的构建过程。我还可以使用 UIAutomation 测试脚本（instruments 命令行实用程序）在模拟器上测试应用程序。但我想使用其他工具，如时间分析器、泄漏等，以及自动化功能来自动测试不规则行为。
虽然我可以制作自定义模板并自动运行它，但问题是我无法解析这些其他仪器的结果。例如，有没有办法知道我的代码中存在泄漏，并触发电子邮件（来自詹金斯）？
I am working on automating my iOS project's build process using Jenkins and xcodebuild command line tool. I am also able to test the application on simulator using UIAutomation testing scripts (instruments command line utility). But I would like to use other instruments like Time Profiler, Leaks, etc.. along with the automation to automatically test for irregular behavior.
Although I can make a custom template and run it with automation, but the problem is I can not parse the results of these other Instruments. Is there a way, for example, to know that there is a leak in my code, and trigger an email (from jenkins)?
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
这是可能的（我已经做到了），但是要做好却非常复杂。。您需要一些东西来将 Instruments 结果 plist 转换为 Jenkins 可以理解的内容。您可以使用 XSLT（我不推荐它），或者在 perl/python/等中编写脚本。
Jenkins XCode 插件是 Jenkins 自动化 Xcode 的优秀工具。
https://wiki.jenkins-ci.org/display/JENKINS/Xcode+Plugin除此之外
，围绕 UIAutomation 构建 CI 环境是可行的，但存在很多问题 - Jenkins 用户必须拥有正确的权限集才能使用开发工具，必须处于正确的位置组，并且必须登录到 UI 才能使模拟器从 Instruments 正确运行。确保您能够做到这一点，然后再投入太多时间来解析 Instruments 运行的输出。
It's possible (I've done it), but it is terribly complicated to do well. You need something to translate the Instruments results plist to something Jenkins can understand. You can use XSLT (I wouldn't recommend it), or write a script in perl/python/etc.
The Jenkins XCode plugin is an excellent tool for automating Xcode from Jenkins.
https://wiki.jenkins-ci.org/display/JENKINS/Xcode+Plugin
That aside, building a CI environment around UIAutomation is doable, but there are a lot of gotchas - the Jenkins user has to have the correct set of permissions to use the dev tools, has to be in the right group, and has to be logged into the UI for the simulator to run correctly from Instruments. Make sure you can get that far before you invest too much time into parsing the output of the Instruments run.