我正在尝试解决内存问题,但我真的不知道从哪里开始。当我运行测试时,我遇到了JS堆问题,并且即使不运行,测试也会出错。我的直觉是,这是由我用于国际化的非常大的字符串文件引起的。这是一个继承的项目,因此我将概述当前发生的情况,也许有人可以帮助我找出一种使之变得更好的方法。
在主应用程序中
文件inject.en.js
是一个2.2MB的json
文件,其中包含应用程序的翻译字符串。
var INJECT_en = { ... };
要尽早使用它,它作为全局var
包含在应用程序index.html
<script src="js/languages/hybrid.inject.en.js" type="text/javascript"></script>
然后用于构建全局配置对象。
var windowJS = new WindowService().nativeWindow;
var HybridInject = windowJS.HybridInject_en; // custom app strings
var INJECT = windowJS.INJECT_en;
var startInjector = _.merge(_.clone(HybridInject), _.clone(INJECT));
export const CONFIG: ConfigType = {
...
strings: populateStrings(startInjector),
...
}
用于测试
出于多种原因,我将Jest用于测试此应用程序。我尝试了一些方法来设置Jest,以便在构建CONFIG
时以相同的方式使这些文件可用,并为我的jestGlobalMocks.ts
确定了以下内容。
我将文件复制到测试目录中,并将其转换为export
对象。
export const INJECT_en = { ... }
然后通过在窗口上将其提供给CONFIG
构建来使它们可用。
Object.defineProperty(window, 'HybridInject_en', {value: HybridInject_en});
Object.defineProperty(window, 'INJECT_en', {value: INJECT_en});
结果
运行测试时,我得到以下输出
[BABEL] Note: The code generator has deoptimised the styling of "jest/configs/inject.en.ts" as it exceeds the max of "500KB".
<--- Last few GCs --->
[208:0x37d62f0] 86678 ms: Scavenge 1384.7 (1422.5) -> 1383.8 (1423.0) MB, 11.5 / 0.1 ms (average mu = 0.143, current mu = 0.101) allocation failure
[208:0x37d62f0] 86694 ms: Scavenge 1384.7 (1423.0) -> 1383.9 (1423.5) MB, 13.5 / 0.1 ms (average mu = 0.143, current mu = 0.101) allocation failure
[208:0x37d62f0] 86710 ms: Scavenge 1384.9 (1423.5) -> 1384.2 (1424.0) MB, 13.5 / 0.1 ms (average mu = 0.143, current mu = 0.101) allocation failure
<--- JS stacktrace --->
==== JS stack trace =========================================
0: ExitFrame [pc: 0x255fdc5dbe1d]
Security context: 0x0e1a9b79e6e1 <JSObject>
1: SourceMapConsumer_allGeneratedPositionsFor [0x32dd014cb6a1] [/var/jenkins_home/workspace/base_jenkins-build/node_modules/source-map/lib/source-map-consumer.js:~178] [pc=0x255fdcb3491a](this=0x0319d2aa8031 <BasicSourceMapConsumer map = 0x68f75371d81>,aArgs=0x1954c0c022a9 <Object map = 0x68f75372671>)
2: /* anonymous */(aka /* anonymous */...
FATAL ERROR: Ineffective mark-compacts near heap limit Allocation failed - JavaScript heap out of memory
1: 0x8daaa0 node::Abort() [node]
2: 0x8daaec [node]
3: 0xad73ce v8::Utils::ReportOOMFailure(v8::internal::Isolate*, char const*, bool) [node]
4: 0xad7604 v8::internal::V8::FatalProcessOutOfMemory(v8::internal::Isolate*, char const*, bool) [node]
5: 0xec4c32 [node]
6: 0xec4d38 v8::internal::Heap::CheckIneffectiveMarkCompact(unsigned long, double) [node]
7: 0xed0e12 v8::internal::Heap::PerformGarbageCollection(v8::internal::GarbageCollector, v8::GCCallbackFlags) [node]
8: 0xed1744 v8::internal::Heap::CollectGarbage(v8::internal::AllocationSpace, v8::internal::GarbageCollectionReason, v8::GCCallbackFlags) [node]
9: 0xed43b1 v8::internal::Heap::AllocateRawWithRetryOrFail(int, v8::internal::AllocationSpace, v8::internal::AllocationAlignment) [node]
10: 0xe9d834 v8::internal::Factory::NewFillerObject(int, bool, v8::internal::AllocationSpace) [node]
11: 0x113cf9e v8::internal::Runtime_AllocateInNewSpace(int, v8::internal::Object**, v8::internal::Isolate*) [node]
12: 0x255fdc5dbe1d
Aborted
npm ERR! Test failed. See above for more details.
答案 0 :(得分:1)
在package.json
的脚本部分中,您应该增加堆(下面的示例,但是您需要适应您的用途,我从来没有用过玩笑)
"gulp": "node --max_old_space_size=2560 ./node_modules/gulp/bin/gulp",
"run-build": "npm run gulp",
"run-build-with-switches": "npm run gulp -- myspecialtarget"
第一行设置了一个脚本,该脚本在每次使用npm run gulp
引用时都会以增加的堆运行gulp。
第二行将运行不带参数的第一个脚本(对于gulp而言,这意味着默认目标)。
第三行将使用参数myspecialtarget
运行gulp。
答案 1 :(得分:0)
经过一段时间的努力,我设法解决了这个问题,方法是将package.json
文件中所有依赖项中的所有 ^ 替换为〜。>
替换
"dependencies": {
"@agm/core": "^1.0.0-beta.5",
"@angular/animations": "^7.2.0",
"@angular/common": "^7.2.0",
"@angular/compiler": "^7.2.0",
"@angular/core": "^7.2.0",
"@angular/forms": "^7.2.0",
...
},
收件人
"dependencies": {
"@agm/core": "~1.0.0-beta.5",
"@angular/animations": "~7.2.0",
"@angular/common": "~7.2.0",
"@angular/compiler": "~7.2.0",
"@angular/core": "~7.2.0",
"@angular/forms": "~7.2.0",
...
},
我认为这个问题是由于不同版本的库之间存在冲突引起的。