TypeError due to assuming all <script> elements have a src attribute #117

steveme · 2015-02-25T10:58:47Z

Hi,

Am seeing a lot of :

Uncaught TypeError: Cannot read property 'replace' of null webpack-dev-server.js:1

Which is from here:
https://github.com/webpack/webpack-dev-server/blob/master/client/index.js#L5

As i've got script elements in my markup that don't have a src attribute.

We're using fluxible, so our HTML has inline state in a <script/> tag like: https://github.com/yahoo/flux-examples/blob/master/fluxible-router/components/Html.jsx

Thanks.

Steve

The text was updated successfully, but these errors were encountered:

jhnns · 2015-02-25T16:05:22Z

Yep, that's an error... that's probably not a good fallback to get the base path. @sokra when is __resourceQuery not set?

sokra · 2015-02-25T17:15:23Z

It only checks the current script tag, which should be the script tag for the webpack-dev-server client. __resourcePath is used when putting the webpack-dev-server/client?... in the entry. I don't think this is wrong. I think the error is somewhere else:

<script> tag to the WDS client is included with async, defer or inlined
webpack-dev-server/client is added to the entry point but no host argument is included.

voronianski · 2015-03-23T13:31:50Z

@sokra @steveme @jhnns I have very similar issue:

webpack-dev-server --hot --inline --progress --colors

70% 1/1 build modules
TypeError: Object /Users/me/github/app-webpack/node_modules/webpack-dev-server/client/index.js?http://localhost:8080,webpack/hot/dev-server,webpack/hot/dev-server has no method 'replace'
at Tapable.<anonymous> (/Users/me/github/app-webpack/node_modules/webpack/lib/NormalModuleFactory.js:42:26)

webpack config:

'use strict';

var webpack = require('webpack');

module.exports = {
    entry: [
        'webpack/hot/dev-server',
        './src/app.js'
    ],
    output: {
        path: __dirname + '/public/build/',
        filename: 'bundle.js',
        publicPath: '/build/'
    },
    resolve: {
        extensions: ['', '.js', '.jsx']
    },
    module: {
        loaders: [
            {
                test: /\.(js|jsx)$/,
                exclude: /node_modules/,
                loaders: ['babel-loader?experimental']
            }
        ]
    },
    plugins: [
        new webpack.HotModuleReplacementPlugin(),
        new webpack.NoErrorsPlugin()
    ]
};

base.html is served via express server where {{mainJS}} is passed to locals as 'http://localhost:8080/build/bundle.js' (also want to figure out how better serve this in production):

<!doctype html>
<html>
<head>
    <meta charset="utf-8">
    <title>App Webpack</title>
</head>
<body>
    <div id="app"></div>
    <script src="{{mainJS}}"></script>
</body>
</html>

Zache · 2016-01-19T10:06:42Z

This happens to me because Google Tag Manager inserts a script element at the bottom of the page and Webpack tries to replace the last script element, which no longer is the bundle...

iest · 2016-02-02T10:02:23Z

Get this exact error too, what @Zache said. Fixed by removing GTM in dev...

MattKunze · 2016-02-10T03:23:55Z

I'm getting this when I have the redux-devtools Chrome extension active - the extension injects a script into the document which doesn't have a src attribute

micha149 · 2016-02-10T10:31:54Z

I can confirm this problem with the redux devtools chrome extension

tiye · 2016-02-15T08:22:39Z

Because of defer in our case. I'd like to drop defer.

3LOK · 2016-02-20T23:34:20Z

We get it because we lazy load our scripts...

hijonathan · 2016-04-01T05:45:20Z

👀 me too

ptz0n · 2016-04-09T08:20:01Z

Same here. Seems to be related to the defer attribute of the script element.

Update: #447 solves the issue. 👍

brotzky · 2016-05-10T00:22:09Z

@Zache the GTM issue was never resolved, correct? I still get Uncaught TypeError: Cannot read property 'replace' of null with GTM in my index.html. This occurs on both 1.14.0 and 1.14.1 for me.

Zache · 2016-05-11T06:51:49Z

@brotzky I'm not sure, it wasn't GTM itself but rather the tags and our workaround was to remove them from our development env.

brotzky · 2016-05-11T15:57:46Z

@Zache, I think we will take the same approach. Thanks!

ptz0n · 2016-05-15T17:37:27Z

Sorry, but I don't get what's hindering us from merging either #447 or #394.

Please merge or clarify 😸

jhnns · 2016-05-15T19:26:33Z

I've just merged #447 – but now I think, this PR is only removing the error instead of being an actual solution to the problem.

When the url can't be retrieved from the script tag, it will default to / which might be correct in some cases, in some others, however, it will not be correct.

Wouldn't it be better to use document.currentScript? Unfortunately, there are no tests so I don't know if that change would break anything :(

achimnol · 2016-07-15T01:33:44Z

I have faced this issue with deferred ZenDesk script injection. Looking forward to find a good solution!

brotzky · 2016-07-18T15:46:34Z

@achimnol
@Zache's solution is the simplest. We have a dev and prod index.html that has different 3rd party scripts in them (GTM, LiveChat, Polyfills/Shims, etc). Use the dev index.html in development and then point webpack at the production index.html when running a build. Proven to be the simplest solution for now.

SpaceK33z · 2016-08-21T13:49:31Z

@jhnns, document.currentScript looks exactly like the thing we want. It also supports scripts that run with async, so it will solve many other issues that were reported.

The only problem is that document.currentScript does not work for IE 6-11. There is a polyfill available, which supports IE6-10, but not IE11.

So what should we do here? Do we support IE?

reem · 2016-08-31T02:06:28Z

I and several other members of my team are hitting this issue as well and I believe it is just because of chrome extensions that we have. It would be really great if the fix from #447 was backported to v1 even if it's only a temporary solution. Right now we are just stuck refreshing the page until it works, which is a very frustrating side effect of an otherwise amazing tool!

SpaceK33z · 2016-09-01T15:25:32Z

This issue was fixed in PR #496. It now behaves like this:

Use document.currentScript if available. This is the most reliable way to get the current script, and works in most browsers.
Fallback to document.scripts.
If that fails, throw an exception.

I'm not sure about backporting this to the 1.x branch, because it can potentially break existing setups.

jhnns · 2016-09-01T15:27:53Z

Awesome! 👍 Great work!

SpaceK33z · 2016-09-06T08:54:09Z

Released in 2.1.0-beta.3.

Braincompiler · 2016-09-20T09:36:56Z

Is it also fixed in 1.16.x? We cannot use webpack2 (it tries to resolve our dynamic module loading approach using System.import and we cannot prevent webpack2 from doing this).

SpaceK33z · 2016-09-20T09:39:07Z

@Braincompiler, yes, this was backported in 1.15.2.

aryzing · 2018-09-25T17:28:48Z

Not sure if this issue is back, but I'm seeing the error

Warning: React DevTools encountered an error: TypeError: Cannot read property 'replace' of null

as reported by #381. This error is reported the very first time webpack HMRs. Subsequent HMRs do not cause this error to show up. Let me know if you need a repo reproducing the issue.

Zache referenced this issue Jan 20, 2016

Fix socket connection for non-inlined connections

3d3c000

jhnns mentioned this issue May 15, 2016

fix undefined src attribute in scriptHost #394

Closed

SpaceK33z added the type: bug label Aug 21, 2016

SpaceK33z mentioned this issue Aug 21, 2016

Getting errors on client build. #367

Closed

SpaceK33z added the priority: 2 (urgent) label Aug 21, 2016

narkq mentioned this issue Aug 30, 2016

Avoid accessing document if __resourceQuery is defined #577

Merged

SpaceK33z closed this as completed Sep 1, 2016

evenstensberg mentioned this issue Dec 16, 2016

flags issue webpack/webpack#3499

Closed

subodhpareek18 mentioned this issue Feb 4, 2017

JWT Authentication setup failing feathersjs-ecosystem/authentication#411

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

TypeError due to assuming all <script> elements have a src attribute #117

TypeError due to assuming all <script> elements have a src attribute #117

steveme commented Feb 25, 2015

jhnns commented Feb 25, 2015

sokra commented Feb 25, 2015

voronianski commented Mar 23, 2015

Zache commented Jan 19, 2016

iest commented Feb 2, 2016

MattKunze commented Feb 10, 2016

micha149 commented Feb 10, 2016

tiye commented Feb 15, 2016

3LOK commented Feb 20, 2016

hijonathan commented Apr 1, 2016

ptz0n commented Apr 9, 2016

brotzky commented May 10, 2016

Zache commented May 11, 2016

brotzky commented May 11, 2016

ptz0n commented May 15, 2016

jhnns commented May 15, 2016 •

edited

Loading

achimnol commented Jul 15, 2016

brotzky commented Jul 18, 2016

SpaceK33z commented Aug 21, 2016

reem commented Aug 31, 2016

SpaceK33z commented Sep 1, 2016 •

edited

Loading

jhnns commented Sep 1, 2016

SpaceK33z commented Sep 6, 2016

Braincompiler commented Sep 20, 2016

SpaceK33z commented Sep 20, 2016

aryzing commented Sep 25, 2018

TypeError due to assuming all <script> elements have a src attribute #117

TypeError due to assuming all <script> elements have a src attribute #117

Comments

steveme commented Feb 25, 2015

jhnns commented Feb 25, 2015

sokra commented Feb 25, 2015

voronianski commented Mar 23, 2015

Zache commented Jan 19, 2016

iest commented Feb 2, 2016

MattKunze commented Feb 10, 2016

micha149 commented Feb 10, 2016

tiye commented Feb 15, 2016

3LOK commented Feb 20, 2016

hijonathan commented Apr 1, 2016

ptz0n commented Apr 9, 2016

brotzky commented May 10, 2016

Zache commented May 11, 2016

brotzky commented May 11, 2016

ptz0n commented May 15, 2016

jhnns commented May 15, 2016 • edited Loading

achimnol commented Jul 15, 2016

brotzky commented Jul 18, 2016

SpaceK33z commented Aug 21, 2016

reem commented Aug 31, 2016

SpaceK33z commented Sep 1, 2016 • edited Loading

jhnns commented Sep 1, 2016

SpaceK33z commented Sep 6, 2016

Braincompiler commented Sep 20, 2016

SpaceK33z commented Sep 20, 2016

aryzing commented Sep 25, 2018

jhnns commented May 15, 2016 •

edited

Loading

SpaceK33z commented Sep 1, 2016 •

edited

Loading