This cookbook is a guide to creating CEP 11.1 HTML/JavaScript Extensions for Creative Cloud applications. CSXS is the old name before CS6, and CEP (Common Extensibility Platform) is new name from CS6. When we talk about CEP or CSXS, they refer to the same project.
CEP (formerly CSXS) Extensions extend the functionality of the Adobe point products in which they run. Extensions are loaded into applications through the PlugPlug Library architecture. Starting from CEP 4.0, HTML/CSS and JavaScript (ECMAScript 5) can be used to develop extensions.
These extension types are supported by CEP. You need to specify an extension's type in its manifest.xml.
- Panel
- The Panel type behaves like any other application panel. It can be docked, participates in workspaces, has fly-out menus, and is re-opened at start-up if open at shutdown.
- ModalDialog
- A Modal Dialog type opens a new extension window and forces the user to interact only with the extenison window before returning control to the host application. User can interact with host application only after closing extension window.
- Modeless
- A Modeless Dialog type opens a new extension window but doesn't force the user to interact with the extension window.
- Custom (Since CEP 5.0)
- This type is for invisible extensions. An invisible extension remains hidden and never becomes visible during its whole life cycle. Read "Invisible HTML Extensions" for more details.
These applications support CEP HTML extensions. (For older applications, see this chart.)
| Application | Host ID | CC 2019 Version | CC 2020 Version | FY 2020 | FY2021 |
|---|---|---|---|---|---|
| Photoshop | PHSP/PHXS | 20 (CEP 9) | 21 (CEP 9) | 22.0 (CEP 10) | TBD |
| InDesign | IDSN | 14 (CEP 9) | 15 (CEP 9) | 16.0 (CEP 10 ) | 16.3(CEP 11) - pre-release |
| InCopy | AICY | 14 (CEP 9) | 15 (CEP 9) | 16.0 (CEP 10 ) | 16.3(CEP 11) - pre-release |
| Illustrator | ILST | 23 (CEP 9) | 24 (CEP 9) | 25.0 (CEP 10) | 25.3(CEP 11) - pre-release |
| Premiere Pro | PPRO | 13 (CEP 9) | 14 (CEP 9) | 14.4 (CEP 10) | 15.4(CEP 11) - upcoming pre-release |
| Prelude | PRLD | 8 (CEP 9) | 9 (CEP 9) | 10.0(CEP 10) | 10.1(CEP 11) - upcoming pre-release |
| After Effects | AEFT | 16 (CEP 9) | 17 (CEP 9) | 17.1.4 (CEP 10) | 18.4(CEP 11) - upcoming pre-release |
| Animate | FLPR | 19 (CEP 9) | 20 (CEP 9) | 21.0 (CEP 10) | TBD |
| Audition | AUDT | 12 (CEP 9) | 13 (CEP 9) | 13.0.10 (CEP 10) | 14.4(CEP 11) - upcoming pre-release |
| Dreamweaver | DRWV | 19 (CEP 9) | 20 (CEP 9) | 21.0 (CEP 10) | TBD |
| Bridge | KBRG | 9 (CEP 9) | 10 (CEP 9) | 11.0 (CEP 10) | TBD |
| Rush | RUSH | 1 (CEP 9) | 1.2.1 (CEP 9) | 1.5.29 (CEP 10) | 2.1(CEP 11) - upcoming pre-release |
CEP HTML engine is based on Chromium Embedded Framework version 3 (CEF3). You can find more information about CEF here. Here are the versions used in CEP:
| Component | CEP 9.0 | CEP 10.0 | CEP 11.1 |
|---|---|---|---|
| CEF 3 | CEF 3 release branch 3163 | CEF 3 release branch 3729 | CEF 3 release branch - (88) |
| Chromium | 61.0.3163.91 | 74.0.3729.157 | 88 |
| Node.js | Node.js 8.6.0 | Node.js 12.3.1 | Node.js 15.9.0 |
| CEF/Node integration | Node-Webkit 0.25 | Node-Webkit 0.38 | Node-Webkit 0.50.1 |
| v8 | 6.3.292.49 | 7.4.288 | 8.7 |
CEP supports two kinds of cookies:
- Session Cookies - Temporary in-memory cookie which will expire and disappear when user closes extension
- Persistent Cookies - No expiry date or validity interval, stored in user's file system
Persistent Cookies location:
- Windows: `C:\Users\<USERNAME>\AppData\Local\Temp\cep_cache\`
- macOS: `/Users/<USERNAME>/Library/Caches/CSXS/cep_cache/`
Each persistent cookie is a file. File name is HostID_HostVersion_ExtensionName, such as PHXS_15.0.0_com.adobe.extension1.
-
Update the Node Modules: CEP 11 works with NodeJS 15.9.0 and Node-WebKit 0.50.1 versions. If extensions were build on older version of NodeJS there can be cases where these extensions may not load in CEP 11 due to incompatible node modules. In such cases, the node modules need to be updated using NodeJS version 15.9.0 or higher.
-
Content-Security-Policy Changes: With CEP 11.1 that integrates CEF #88, content-security-policy checks have been enforced strictly. Extensions that embed third-party sites may not be able to do so anymore if the site that is being embedded has a content-security-policy directive that doesn't allow sites to frame them.
-
Cookies: With #CEF v88 version, we have seen that when no SameSite attribute is set in a cookie, it defaults to SameSite=Lax. This will cause the cookie to be blocked in cross-site contexts. If your application needs cookies to be used in a cross-site context, we suggest you set SameSite=None.If component is not sending
SameSiteattribute in the cookie it would default toSameSite=Laxfor preventing to set the cookie in cross-site context. -
CORS checks: With CEP 11, with security updates, we are seeing that a lot more requests are being blocked by CORS checks. The extension may encounter this if it is trying to access an endpoint that does not have the Access-Control-Allow-Origin header. Please ensure that the endpoints that are being accessed have Access-Control-Allow-Origin headers set appropriately.
Also Refer Known issues in CEP 11
There is no migration required for the extensions working in CEP 9. Extensions running on CEP 9 should run on CEP 10 out of the box.
However few point to be noted with CEP 10
- Issues seen on NWJS version 0.38 will also be applicable in CEP 10
For example: There are errors noticed when node module "winston" is used with any other node module which uses duplex. Workaround: In case you are using the node module "winston" please add this below script to the extension.
var {Duplex} = require('stream')
var {Writable} = require('stream')
{
// Allow the keys array to be GC'ed.
const keys = Object.keys(Writable.prototype);
for (var v = 0; v < keys.length; v++)
{
const method = keys[v];
if (!Duplex.prototype[method])
Duplex.prototype[method] = Writable.prototype[method];
}
}
- Update your node modules
CEP 10 works with NodeJS 12.3.1 and Node-WebKit 0.38 versions. If extensions were build on older version of NodeJS there can be cases where these extensions may not load in CEP 10 due to incompatible node modules. In such cases, the node modules need to be updated using NodeJS version 12.3.1 or higher.
3.Issues in CSS properties due to deprecation in Chromium will be applicable in CEP 10
In such cases, extension developers will need to update the CSS for the extension, in order for the extension to render correctly in CEP 10. For eg: https://bugs.chromium.org/p/chromium/issues/detail?id=927066
- Also refer details for the disable nodejs option here.
There is no migration required for the extensions working in CEP 8. Extensions running on CEP 8 should run on CEP 9 out of the box.
Note:
<Size> element in manifest file is mandatory as per the documentation but it used to work by taking max or min size provided in the manifest till CEP 8, this bug has been fixed in CEP 9. Please ensure to define <Size> element in your manifest.
Below listed changes are applicable when --enable-nodejs is passed as CEFCommandline parameter.
CEP 8 and later versions introduces additional symbol: cep_node in the global context when nodejs is enabled. cep_node will hold following node specific symbols:
cep_node members Buffer, global, process and require
nodejs symbols will not be available in iframe's global context, only way you can access nodejs APIs within iframe's context is through cep_node. When --mixed-context is enabled, global node symbols and cep_node is available within iframe's global context as browser and node gets executed at the same context. It is also important to note that, if you have existing code to check or use nodejs symbols in iframe's global context, it used to work till last release and in the current release, it breaks in separate context mode.
Newly integrated nodejs adds "module" and "exports" symbols to the global context. Many libraries such as JQuery relies on its booting based on the availability of "module" symbol in the global scope. For ex, JQuery has following code while booting:
JQuery startup code
if ( typeof module === "object" && typeof module.exports === "object" ) {
// set jQuery in `module`
} else {
// set jQuery in `window`
}When this code is executed in CEP's browser with nodejs enabled, it will make JQuery to load in module context instead of Browser context. This would cause issues to extension's startup. Please use below code to handle such scenarios:
global symbols handling
<!-- Insert this line above script imports -->
<script>if (typeof module === 'object') {window.module = module; module = undefined;}</script>
<script>if (typeof exports === 'object') {window.exports = exports; exports = undefined;}</script>
<!-- extension's imports -->
<script src="scripts/jquery.min.js"></script>
<script src="scripts/csinterface.js"></script>
<!-- Insert this line after script imports -->
<script>if (window.module) module = window.module;</script>
<script>if (window.exports) exports = window.exports;</script>If you are not using "module" and "exports" in the extension, you could skip last 2 lines of above code. This logic is similar to how users handled nodejs's require while importing. If you are handling require already, you should continue to handle same way.
Nodejs 7.7.4 or higher versions requires path to be included as absolute path. So, if you have nodejs require for the js file,replace the path to absolute instead of relative. For eg. replace:
require("./js/lib/jquery.js");with
Win require absolute path for nodejs
var loc = window.location.pathname;
var dir = decodeURI(loc.substring(1, loc.lastIndexOf('/')));
require(dir + "/js/lib/jquery.js");On macOS (there is a change in index passed on to substring)
macOS require absolute path for nodejs
var loc = window.location.pathname;
var dir = decodeURI(loc.substring(0, loc.lastIndexOf('/')));
require(dir + "/js/lib/jquery.js");
(or)
require(__dirname + "/js/lib/jquery.js");CEP will have multiple contexts defined at the start. Following are the default contexts:
-
Browser Context
This is the default JS context associated with CEP's browser engine. Apart from Browser objects, CEP's file system APIs are accessed in this context.
-
Native Context
This is extended context through "cep" object to provide file system access via CSInterface.js.
-
Node Context (When nodejs is enabled through --enable-nodejs command line switch)
1. Separate Context 2. Mixed Context -
Host app context
This is the context which would be accessed through evalscript API.
Separate Context Mode
This context will be created with following command line switch: --enable-nodejs (and --mixed-context is not passed). When CEP is starting with this context, it would have following symbols:
-
cep_node: This is treated as global symbol available across frames and helps to access node symbols. Any modification to this symbol will be reflected at all the frame context. This symbols is introduced from CEP 8.
-
node globals: https://nodejs.org/api/globals.html could be accessed only in main frame (which is current document). If there is a new frame created in the current document, these symbols will not be available in frame's context. Please note that, this behavior is different CEP 7 for frames. In CEP 7, node globals are available for access in all frames.
Following representation provides visual context:
Separate Context
<html>
<head>
<!-- cep_node and node globals are available in this context -->
</head>
<body>
<!-- cep_node and node globals are available in this context -->
<iframe name="iframe_1">
<!-- only cep_node is available and node globals are not available within this iframe -->
</iframe>
<iframe name="iframe_2">
<!-- only cep_node is available and node globals are not available within this iframe -->
<!-- In the following line cep_node is modified and this change will be available for any frame's which is loading later than iframe_2 -->
cep_node.process.myVar = "Hello";
</iframe>
<iframe name="iframe_3">
<!-- cep_node is available with change: cep_node.process.myVar(iframe's loading order is important as change should be accessed after the modifications) and node globals are not available -->
</iframe>
</body>
</html>Mixed Context Mode
This context will be created with following command line switch: --enable-nodejs and --mixed-context. When CEP is starting with this context, it would have following symbols:
-
cep_node: This is treated as local symbol available within each frames and helps to access node symbols. Any modification to this symbol will not be reflected to other frames. This symbols is introduced from CEP 8.
-
node globals: https://nodejs.org/api/globals.html could be accessed in all the frames which is unique to each frame.
Following representation provides visual context
Mixed Context
<html>
<head>
<!-- cep_node and node globals are available in this context -->
</head>
<body>
<!-- cep_node and node globals are available in this context -->
<iframe name="iframe_1">
<!-- only cep_node is available and node globals are available within this iframe -->
</iframe>
<iframe name="iframe_2">
<!-- only cep_node is available and node globals are available within this iframe -->
<!-- In the following line cep_node is modified and this change will not be available for any other frames -->
cep_node.process.myVar = "Hello";
</iframe>
<iframe name="iframe_3">
<!-- cep_node is available which is unique to this context and it's built from start up time. So, it will not have the changes from previous iframe. node globals are available -->
</iframe>
</body>
</html>CEP HTML Extensions can be developed on both Windows and macOS platforms. The development machine needs to have the following applications in order to successfully develop CSXS extensions:
- Adobe Creative Cloud applications supporting CEP HTML extensions.
- HTML Extension Builder (Nice to have, but not mandatory).
- Adobe ExtendScript Tool Kit (This is available as an optional install).
- Chrome browser for debug workflow.
HTML Extension Builder is a tool set built on top of Eclipse and can be used for developing and debugging HTML extensions. Please download the Extension Builder 3 from here.
- Before you sign the extensions, you need to get or create the certificate file. Configurator and Adobe Exchange Packer can create certificates. Developers can get all information here after logging in.
- Three tools can be used to sign a HTML extension.
- Extension Builder 3
- CC Extensions Signing Toolkit (also on above labs web site)
- Example of using CC Extension signing toolkit: ccextensionswin64.exe -sign "d:\Adobe Layer Namer\Adobe Layer Namer"(input extension path) d:\AdobeLayerNamer.zxp (output zxp path) d:\sign.p12 (certificate path) 1 (certificate password) Adobe Exchange Packer (please sign in so that you can see it.)
- Adobe Exchange Packer (please sign in so that you can see it.)
If you are in the midst of development and are not using HTML Extension Builder for debug workflows and want to bypass the need to sign your extensions, you can bypass the check for extension signatures by editing the CSXS preference properties file, located at:
- Win: regedit > HKEY_CURRENT_USER/Software/Adobe/CSXS.11, then add a new entry PlayerDebugMode of type "string" with the value of "1".
- macOS: In the terminal, type:
defaults write com.adobe.CSXS.11 PlayerDebugMode 1(The plist is also located at /Users/<username>/Library/Preferences/com.adobe.CSXS.11.plist)
These entries will enable debug extensions to be displayed in the host applications. Please note that, CSXS.11 is given with the assumption that, you are developing the extension for CEP 11. If you are developing extension for previous version of CEP, replace 11 with corresponding version number.
Staring with macOS 10.9, Apple introduced a caching mechanism for plist files. Your modifications to plist files does not take effect until the cache gets updated (on a periodic basis, you cannot know exactly when the update will happen). To make sure your modifications take effect, there are two methods.
- Kill cfprefsd process. It will restart automatically. Then the update takes effect.
- Restart your macOS, or log out the current user and re-log in.
- More Information
CEP supports remote debugging for HTML extensions from 5.0.
- Create a “.debug” file to the extension root directory such as
Test_Extension\.debug. The .debug file contains remote debug ports. Developers must create this file and use valid debug ports because both remote debugging and dev tools are based on it. - ".debug" file name is special for both Windows and macOS platforms, it has to be created via command line.
- On Windows, use "copy con .debug" and "Ctrl+Z" to create an empty file.
- On macOS, use "touch .debug" to create an empty file.
- The value of Port should be between 1024 and 65535 (not include 65535), otherwise remote debugging and dev tools will not work.
- One extension bundle may have multiple extensions. The .debug file can specify debug ports for each extension. Here is an example file:
<?xml version="1.0" encoding="UTF-8"?>
<ExtensionList>
<Extension Id="com.adobe.CEPHTMLTEST.Panel1">
<HostList>
<Host Name="PHXS" Port="8000"/>
<Host Name="IDSN" Port="8001"/>
<Host Name="AICY" Port="8002"/>
<Host Name="ILST" Port="8003"/>
<Host Name="PPRO" Port="8004"/>
<Host Name="PRLD" Port="8005"/>
<Host Name="FLPR" Port="8006"/>
<Host Name="AUDT" Port="8007"/>
</HostList>
</Extension>
<Extension Id="com.adobe.CEPHTMLTEST.Panel2">
<HostList>
<Host Name="PHXS" Port="8100"/>
<Host Name="IDSN" Port="8101"/>
<Host Name="AICY" Port="8102"/>
<Host Name="ILST" Port="8103"/>
<Host Name="PPRO" Port="8104"/>
<Host Name="PRLD" Port="8105"/>
<Host Name="FLPR" Port="8106"/>
<Host Name="AUDT" Port="8107"/>
</HostList>
</Extension>
<Extension Id="com.adobe.CEPHTMLTEST.ModalDialog">
<HostList>
<Host Name="PHXS" Port="8200"/>
<Host Name="IDSN" Port="8201"/>
<Host Name="AICY" Port="8202"/>
<Host Name="ILST" Port="8203"/>
<Host Name="PPRO" Port="8204"/>
<Host Name="PRLD" Port="8205"/>
<Host Name="FLPR" Port="8206"/>
<Host Name="AUDT" Port="8207"/>
</HostList>
</Extension>
<Extension Id="com.adobe.CEPHTMLTEST.Modeless">
<HostList>
<Host Name="PHXS" Port="8300"/>
<Host Name="IDSN" Port="8301"/>
<Host Name="AICY" Port="8302"/>
<Host Name="ILST" Port="8303"/>
<Host Name="PPRO" Port="8304"/>
<Host Name="PRLD" Port="8305"/>
<Host Name="FLPR" Port="8306"/>
<Host Name="AUDT" Port="8307"/>
</HostList>
</Extension>
</ExtensionList>If you load an extension whose debug port is 8088, you can load the debugger through http://localhost:8088/ on Chrome.
Log files with useful debugging information are created for each of the applications supporting CEP extensions. The platform-specific locations for the log files are as follows:
- Win:
C:\Users\<USERNAME>\AppData\Local\Temp - macOS:
/Users/<USERNAME>/Library/Logs/CSXS
These files are generated with the following naming conventions:
- CEP 4.0 - 6.0 releases: csxs
<versionNumber>-<HostID>.log. For example, PlugPlug in Illustrator generates log file csxs8-ILST.log. - CEP 6.1 and later releases: CEP
<versionNumber>-<HostID>.log. For example, PlugPlug in Illustrator generates log file CEP8-ILST.log.
Logging levels can be modified as per the following levels:
- 0 - Off (No logs are generated)
- 1 - Error (the default logging value)
- 2 - Warn
- 3 - Info
- 4 - Debug
- 5 - Trace
- 6 - All
The LogLevel key can be updated at the following location (The application should be restarted for the log level changes to take effect):
- Win:
regedit > HKEY_CURRENT_USER/Software/Adobe/CSXS.11 - macOS:
/Users/<USERNAME>/Library/Preferences/com.adobe.CSXS.11.plist
For example of macOS, in the terminal do:
defaults write com.adobe.CSXS.11 LogLevel 6
Please note that, higher the log level will make the extension loading time slower.
In CEP 6.1 and later releases, CEPHtmlEngine generates logs. Each CEPHtmlEngine instance usually generate two log files, one for browser process, the other for renderer process. These files are generated with the following naming conventions:
- Browser process: CEPHtmlEngine
<versionNumber>-<HostID>-<HostVersion>-<ExtensionID>.log - Renderer process: CEPHtmlEngine
<versionNumber>-<HostID>-<HostVersion>-<ExtensionID>-renderer.log
For example:
- CEPHtmlEngine8-PHXS-18.1.1-com.adobe.DesignLibraries.angular.log
- CEPHtmlEngine8-PHXS-18.1.1-com.adobe.DesignLibraries.angular-renderer.log
They are also controlled by the PlugPlug log level.
In CEP 4.0 - 6.0, the Chromium Embedded Framework (CEF) in CEPHtmlEngine also generates a log:
- Win:
C:\Users\<USERNAME>\AppData\Local\Temp\cef_debug.log - macOS:
/Users/<USERNAME>/Library/Logs/CSXS/cef_debug.log
In CEP 6.1 and later releases, this log is merged into CEPHtmlEngine log.
CEP supports 3 types of extension folders.
-
Product extension folder. Here is a suggestion, but each point product can decide where this folder should be. Third party extension can't be installed at this location.
- ${PP}/CEP/extensions (PPs may use different folder.)
-
System extension folder
- Win(x64):
C:\Program Files (x86)\Common Files\Adobe\CEP\extensions, andC:\Program Files\Common Files\Adobe\CEP\extensions(since CEP 6.1) - macOS:
/Library/Application Support/Adobe/CEP/extensions
- Win(x64):
-
Per-user extension folder
- Win:
C:\Users\<USERNAME>\AppData\Roaming\Adobe\CEP/extensions - macOS:
~/Library/Application Support/Adobe/CEP/extensions
- Win:
How does CEP decide which extension to load?
- CEP first searches the product extension folder, then the system extension folder, and finally per-user extension folder.
- Extensions without an appropriate host application ID and version are filtered out.
- If two extensions have same extension bundle ID, the one with higher version is loaded.
- If two extensions have same extension bundle ID and same version, the one whose manifest file has latest modification date is loaded.
- If two extensions have same extension bundle ID, same version and same manifest modification date, CEP loads the first one that is found.
Extension Installation:
- Host application installers should install extensions to product extension folder.
- Extensions installed through https://creative.adobe.com/addons should get install at system extension folder or per-user extension folder.
Note:
- Character '#' is not allowed in extension folder path on both Windows and Mac OSX, since CEF treats '#' as a delimiter.
CEP provides some sample extensions for extension developers. Get its source code from: https://github.com/Adobe-CEP/CEP-Resources/tree/master/CEP_10.x/Samples
The manifest.xml file is required for every extension and provides the necessary information to configure a CEP extension. ExtensionManifest is the root element of the manifest.xml file. Extensions, ExtensionList, and DispatchList are the three child elements of the ExtensionManifest root element.
All HTML extensions must use 5.0 or above version. Get XSD from the GitHub repo: https://github.com/Adobe-CEP/CEP-Resources/tree/master/CEP_7.x (No elements change from CEP 7)
To check if the extension's manifest is in sync with the latest schema, perform the following steps:
- Download the latest schema (ExtensionManifest_
<version>.xsd). Please note that, schema is not changed in CEP 8, so refer to https://github.com/Adobe-CEP/CEP-Resources/blob/master/CEP_7.x/ExtensionManifest_v_7_0.xsd for CEP 8 and 9 as well. - Navigate here
- Upload the schema and your latest ExtensionManifest (from a real build to check the validity of the versions).
- Hit validate
Make sure correct point product versions are used. Here is an example.
<HostList>
<Host Name="PHXS" Version="[15.0,15.9]"/>
<Host Name="PHSP" Version="[15.0,15.9]"/>
</HostList>
This will support Photoshop version 15.0 up to, and including, 15.9. If you use the following syntax then you are supporting releases up to 15.9 but not including 15.9
<HostList>
<Host Name="PHXS" Version="[15.0,15.9)"/>
<Host Name="PHSP" Version="[15.0,15.9)"/>
</HostList>
Make sure correct CEP version is used.
<RequiredRuntimeList>
<RequiredRuntime Name="CSXS" Version="5.0"/>
</RequiredRuntimeList>
CEP 7.0 manifest files now support the use of a HostList specific to an extension in the bundle. For example, the following qualifies as the "default" HostList:
...
<ExecutionEnvironment>
<HostList>
<Host Name="DRWV" Version="15.0" />
<Host Name="FLPR" Version="15.0" />
<Host Name="IDSN" Version="11.0 "/>
<Host Name="AICY" Version="11.0" />
<Host Name="ILST" Version="19.0" />
<Host Name="PHSP" Version="16.0" />
<Host Name="PHXS" Version="16.0" />
<Host Name="PPRO" Version="9.0" />
<Host Name="PRLD" Version="4.0" />
<Host Name="AEFT" Version="13.0" />
<Host Name="DEMO" Version="1.0" />
<Host Name="AUDT" Version="8.0" />
<Host Name="LTRM" Version="7.0" />
<Host Name="MUSE" Version="2015" />
</HostList>
...
</Execution Environment>
...
However, you can also specify a custom HostList on a per extension basis as in the following example:
...
<DispatchInfoList>
...
<Extension Id="com.adobe.CEPHTMLTEST.Panel1">
<HostList>
<Host Name="PHXS" />
</HostList>
...
</Extension>
<DispatchInfo>
...
</DispatchInfo>
...
</DispatchInfoList>
HostLists specified under extension tags will override the default HostList specified under the execution environment tag. It is important to note that adding HostList nodes with no child nodes is akin to turning that particular extension off for all host applications; this behavior is intentional. Please refrain from specifying both HostList tags and Host attributes in DispatchInfo tags for each extension; choose one or the other. Specifying both will more than likely result in unexpected behavior. The Host attribute is only maintained for backward compatibility. New extensions should use HostList tags, not the Host attribute.
You can specify extension size, max size and min size in extension manifest. Size is mandatory; max size and min size are optional.
A modal or modeless dialog is resizable if there are different max and min size, otherwise it is un-resizable. When you move mouse pointer over dialog border, CEP shows different cursor for resizable and un-resizable dialogs.
<Geometry>
<Size>
<Height>580</Height>
<Width>1000</Width>
</Size>
<MaxSize>
<Height>800</Height>
<Width>1200</Width>
</MaxSize>
<MinSize>
<Height>400</Height>
<Width>600</Width>
</MinSize>
</Geometry>
This is only supported in InDesign and InCopy.
You can customize the extension menu by editing <menu/> item in manifest. Here is an example. In this example, the Adobe Add-ons extension is displayed under Windows main menu, rather than extensions menu under Windows. You can customize the location of extension to somewhere else by changing the value of attribute Placement in <menu/> item.
<?xml version="1.0" encoding="UTF-8"?>
<ExtensionList>
<Extension Id="Adobe Add-ons" Version="1.0"/>
</ExtensionList>
<ExecutionEnvironment>
<HostList>
<Host Name="IDSN" Version="8.0"/>
</HostList>
...
</ExecutionEnvironment>
<DispatchInfoList>
<Extension Id="com.adobe.CEPHTMLTEST.Panel1">
<DispatchInfo>
...
<UI>
...
<Menu Placement="'Main:&Window',600.0,'KBSCE Window menu'">Adobe Add-ons</Menu>
...
In high DPI display mode, panel extensions may want to use high DPI icons. You set these icons in extension's manifest.
<Icons>
<Icon Type="Normal">./images/IconLight.png</Icon>
<Icon Type="RollOver">./images/IconLight.png</Icon>
<Icon Type="DarkNormal">./images/IconDark.png</Icon>
<Icon Type="DarkRollOver">./images/IconDark.png</Icon>
</Icons>
You pack both normal icon files (IconLight.png and IconDark.png) and high DPI icon files ([email protected] and [email protected]) in your extension.
Host applications will be able to find and use
- IconLight.png and IconDark.png for normal display
- [email protected] and [email protected] for 200% high DPI display
@2X.ext is the industry standard. Please see more details on https://developer.apple.com/library/ios/qa/qa1686/_index.html.
Note: Photoshop supports _x2.ext format.
You can specify CEP dialog size as a percentage of screen size. Here is an example.
<UI>
<Type>Modeless</Type>
...
<Geometry>
<ScreenPercentage>
<Height>50%</Height>
<Width>50%</Width>
</ScreenPercentage>
...
</Geometry>
</UI>
(Since 5.2)
CEP 5.2 supports shortcut keys for HTML extensions. When focus is on HTML extensions, these shortcut keys are handled by extension.
| Windows Keys | macOS Keys | Function |
|---|---|---|
| Ctrl + A | Command + A | Select All |
| Ctrl + C | Command + C | Copy |
| Ctrl + V | Command + V | Paste |
| Ctrl + X | Command + X | Cut |
Other shortcut keys are handled by point products, such as pressing Ctrl + N to create a new document.
CEP JavaScript Libraries are counterparts of the Flex CSXS Library. They provide JavaScript APIs to access application and CEP information.
- CSInterface.js
- Vulcan.js
Get them from: https://github.com/Adobe-CEP/CEP-Resources/tree/master/CEP_10.x
To use them, please include these JavaScript files in your HTML extension.
The CEP JavaScript APIs keep changing in each new CEP release. The changes are guaranteed to be backward compatible. For newly added APIs, a version tag like "Since x.x.x" is added to its API comments indicating since which CEP version the APIs is available.
You will need to check the version tag against the version of CEP integrated by the Adobe Product you are using to make sure the API you want to use is available. To do so, use CSInterface.getCurrentApiVersion() to retrieve the version of CEP integrated by the Adobe Product. Please note this API itself is available only since 4.2.0. If you get an error saying getCurrentApiVersion is undefined, then you are running in CEP 4.0 or 4.1. Otherwise, the value returned will tell you the version of CEP integrated by the Adobe product.
CEP supports sending and receiving events within an extension, among extensions in an application, and among extensions in different applications. Since all extensions are based on the common communication layer with the same event data format, they can communicate with each other through CEP events, and even they can communicate with native side as long as point products invoke PlugPlugAddEventListener/PlugPlugDispatchEvent accordingly.
There are dispatchEvent/addEventListener/removeEventListener APIs available in JavaScript to dispatch and listen for events. Let's go through CEP event data format/structure, APIs to dispatch and listen event, and sample code snippet accordingly in JavaScript.
In terms of CSEvent, it just means CEP Event here. The data structure of CSEvent (CEP Event) is as below.
/**
* Class CSEvent.
* You can use it to dispatch a standard CEP event.
*
* @param type Event type.
* @param scope The scope of event, can be "GLOBAL" or "APPLICATION".
* @param appId The unique identifier of the application that generated the event. Optional.
* @param extensionId The unique identifier of the extension that generated the event. Optional.
*
* @return CSEvent object
*/
function CSEvent(type, scope, appId, extensionId)
{
this.type = type;
this.scope = scope;
this.appId = appId;
this.extensionId = extensionId;
};
You could create a CSEvent object and dispatch it by using CSInterface.dispatchEvent. Also you could access its property in your callback of CSInterface.addEventListener. Refer to the section addEventListener/dispatchEvent below for more details.
dispatchEvent/addEventListener/removeEventListener APIs are available in JavaScript world to dispatch and listen for CSEvent.
Here is the definition for addEventListener. Refer to CSInterface.js for more information:
/**
* Registers an interest in a CEP event of a particular type, and
* assigns an event handler.
* The event infrastructure notifies your extension when events of this type occur,
* passing the event object to the registered handler function.
* * @param type The name of the event type of interest.
* @param listener The JavaScript handler function or method.
* @param obj Optional, the object containing the handler method, if any.
* Default is null.
*/
CSInterface.prototype.addEventListener = function(type, listener, obj)
One thing needs to be mentioned here is both named and anonymous callback functions are supported in CSInterface.addEventListener.
- An example of how to use named callback function in
CSInterface.addEventListener.
function callback(event)
{
console.log(“type=” + event.type + “, data=” + event.data);
}
var csInterface = new CSInterface();
csInterface.addEventListener(“com.adobe.cep.test”, callback); //invoke the function
- An example of how to use anonymous callback function in
CSInterface.addEventListener.
var csInterface = new CSInterface();
csInterface.addEventListener(“com.adobe.cep.test”, function (event)
{
console.log(“type=” + event.type + “, data=” + event.data);
}
); // Anonymous function is the second parameter
The event.data can be an object (i.e. you could use an object as event.data).
Before CEP 6.1, we regarded every attribute in event.data object as a regular string, but from CEP 6.1, we revised the behavior that keep the type of each attribute in event.data as it was. If the value is a valid JSON string, CEP will parse it natively and convert it to an object. Here is an example on how to use it.
var csInterface = new CSInterface();
csInterface.addEventListener(“com.adobe.cep.test”, function (event)
{
var obj = event.data;
console.log(“type=” + event.type + “, data.property1=” + obj.property1 + “, data.property2=” + obj.property2);
}
); // Anonymous function is the second parameter
Here is the definition for CSInterface.dispatchEvent. Refer to CSInterface.js for more details.
/**
* Triggers a CEP event programmatically. Yoy can use it to dispatch
* an event of a predefined type, or of a type you have defined.
*
* @param event A \c CSEvent object.
*/
CSInterface.prototype.dispatchEvent = function(event)Here are three samples to demonstrate how to dispatch an event in JavaScript.
- An example of how to dispatch event in JavaScript.
var csInterface = new CSInterface();
var event = new CSEvent("com.adobe.cep.test", "APPLICATION");
event.data = "This is a test!";
cSInterface.dispatchEvent(event);
- Another example of creating event object and setting property, then dispatch it.
var csInterface = new CSInterface();
var event = new CSEvent();
event.type = "com.adobe.cep.test";
event.scope = "APPLICATION";
event.data = "This is a test!";
cSInterface.dispatchEvent(event);
- An example of dispatching an event whose data is an object.
var event = new CSEvent("com.adobe.cep.test", "APPLICATION");
var obj = new Object();
obj.a = "a";
obj.b = "b";
event.data = obj;
cSInterface.dispatchEvent(event);
CEP event based communication between native point product and CEP extensions is simple. You can use PlugPlugAddEventListener/PlugPlugDispatchEvent to communicate with CEP extensions.
Please refer the following steps:
First, define a callback function in CEP extension:
function callback(event)
{
// do something here
}
Then call CSInterface.addEventListener with event type you want to listen and the callback function:
var csInterface = new CSInterface();
csInterface.addEventListener(“com.adobe.cep.test”, callback);
Following table lists the standard events supported by point products.
| Event Type | Event Scope | Description | Event Parameter | PS | ID | AI | AN | PR | PL | AU |
|---|---|---|---|---|---|---|---|---|---|---|
| documentAfterActivate | APPLICATION | Event fired when a document has been activated(after new/open document; after document has retrieved focus). | URL to the active document. If the doc was not save, the NAME will be set instead of the URL. |  |
 |
|
|
|
|
|
| documentAfterDeactivate | APPLICATION | Event fired when the active document has been de-activated.(after document loses focus) | URL to the active document. If the doc was not save, the name will be set instead of the URL. | |
|
|
|
|
|
|
| applicationBeforeQuit | APPLICATION | Event fired when the application got the signal to start to terminate. | none | |
 |
|
|
|
|
|
| applicationActivate | APPLICATION | Event fired when the Application got an "activation" event from the OS. | none | |
|
|
|
on macOS; on Windows |
on macOS; on Windows |
|
| documentAfterSave | APPLICATION | Event fired after the document has been saved | URL to the saved document. | |
|
|
|
|
|
|
( = supported, = not supported)
Note: CEP is unloaded from the point products ID and AI right after the event applicationBeforeQuit is emitted from the point products, therefore CEP may have no chance to get this event handled in HTML extensions.
In Photoshop, the following specific events are defined:
- com.adobe.PhotoshopPersistent
- com.adobe.PhotoshopUnPersistent
- com.adobe.PhotoshopWorkspaceSet
- com.adobe.PhotoshopWorkspaceGet
- com.adobe.PhotoshopWorkspaceAware
- com.adobe.PhotoshopWorkspaceData
- com.adobe.PhotoshopWorkspaceRequest
- com.adobe.PhotoshopRegisterEvent
- com.adobe.PhotoshopUnRegisterEvent
- com.adobe.PhotoshopLoseFocus
- com.adobe.PhotoshopQueryDockingState
For example, a CEP extension yields the mouse focus back to Photoshop by sending the com.adobe.PhotoshopLoseFocus event:
var csInterface = new CSInterface();
var event = new CSEvent("com.adobe.PhotoshopLoseFocus", "APPLICATION");
event.extensionId = csInterface.getExtensionID();
csInterface.dispatchEvent(event);
com.adobe.PhotoshopCallback will be removed in Photoshop 17.0 as adding a listener results in all CS Extensions receiving the event. As of Photoshop CC 2015 June release, developers can now use this alternative, which fixes the broadcast issue:
csInterface.addEventListener("com.adobe.PhotoshopJSONCallback" + gExtensionID, PhotoshopCallbackUnique);
First, define a callback function in CEP extension:
function evalScriptCallback(result)
{
// process the result string here.
}
Then call CSInterface.evalScript with the script you want to call and the callback function:
var script = "app.documents.add"; //Demo script
CSInterface.evalScript(script, evalScriptCallback);
Please be aware that the script in evalScript and the jsx file which is configured in <ScriptPath> in the extension's manifest are executed in host application's ExtendScript engine, which runs in host application's main thread. On the other hand, CEP event is also dispatched from host application's main thread. If the interaction between the script and CEP event is needed, please split the script into small parts and call them separately so that CEP event has a chance to be scheduled.
Starting with CEP 5.0, global CEP Events whose scope attribute is set to "GLOBAL" are no longer supported. Please use the APIs in Vulcan.js instead.
The data structure of Vulcan message in JavaScript is as below.
/**
* @class VulcanMessage
* Message type for sending messages between host applications.
* A message of this type can be sent to the designated destination
* when appId and appVersion are provided and valid. Otherwise,
* the message is broadcast to all running Vulcan-enabled applications.
*
* To send a message between extensions running within one
* application, use the <code>CSEvent</code> type in CSInterface.js.
*
* @param type The message type.
* @param appId The peer appId.
* @param appVersion The peer appVersion.
*
*/
function VulcanMessage(type, appId, appVersion)
{
this.type = type;
this.scope = VulcanMessage.SCOPE_SUITE;
this.appId = requiredParamsValid(appId) ? appId : VulcanMessage.DEFAULT_APP_ID;
this.appVersion = requiredParamsValid(appVersion) ? appVersion : VulcanMessage.DEFAULT_APP_VERSION;
this.data = VulcanMessage.DEFAULT_DATA;
}
VulcanMessage.TYPE_PREFIX = "vulcan.SuiteMessage.";
VulcanMessage.SCOPE_SUITE = "GLOBAL";
VulcanMessage.DEFAULT_APP_ID = "UNKNOWN";
VulcanMessage.DEFAULT_APP_VERSION = "UNKNOWN";
VulcanMessage.DEFAULT_DATA = "<data><payload></payload></data>";
VulcanMessage.dataTemplate = "<data>{0}</data>";
VulcanMessage.payloadTemplate = "<payload>{0}</payload>";
addMessageListener, removeMessageListener, dispatchMessage and getPayload APIs are available to dispatch and listen for Vulcan messages. The API definitions are as below. Refer to Vulcan.js for more information.
/**
* Registers a message listener callback function for a Vulcan message.
*
* @param type The message type.
* @param callback The callback function that handles the message.
* Takes one argument, the message object.
* @param obj Optional, the object containing the callback method, if any.
* Default is null.
*/
Vulcan.prototype.addMessageListener = function(type, callback, obj)
/**
* Removes a registered message listener callback function for a Vulcan message.
*
* @param type The message type.
* @param callback The callback function that was registered.
* Takes one argument, the message object.
* @param obj Optional, the object containing the callback method, if any.
* Default is null.
*/
Vulcan.prototype.removeMessageListener = function(type, callback, obj)
/**
* Dispatches a Vulcan message.
*
* @param vulcanMessage The message object.
*/
Vulcan.prototype.dispatchMessage = function(vulcanMessage)
/**
* Retrieves the message payload of a Vulcan message for the registered message listener callback function.
*
* @param vulcanMessage The message object.
* @return A string containing the message payload.
*/
Vulcan.prototype.getPayload = function(vulcanMessage)
Here is the example to demonstrate how to use the APIs in JavaScript.
var testVulcanMessage = new VulcanMessage(VulcanMessage.TYPE_PREFIX + "test");
testVulcanMessage.setPayload("To be or not to be that is a question!");
var callback = function (message) {
alert(VulcanInterface.getPayload(message));
};
VulcanInterface.addMessageListener(testVulcanMessage.type, callback);
VulcanInterface.dispatchMessage(testVulcanMessage);
...
VulcanInterface.removeMessageListener(testVulcanMessage.type, callback);
getEndPoints and getSelfEndPoint APIs are available to support point-to-point Vulcan message. The API definitions are as below. Refer to Vulcan.js for more information.
/**
* Gets all available endpoints of the running Vulcan-enabled applications.
*
* Since 7.0.0
*
* @return The array of all available endpoints.
* An example endpoint string:
* <endPoint>
* <appId>PHXS</appId>
* <appVersion>16.1.0</appVersion>
* </endPoint>
*/
Vulcan.prototype.getEndPoints = function()
/**
* Gets the endpoint for itself.
*
* Since 7.0.0
*
* @return The endpoint string for itself.
*/
Vulcan.prototype.getSelfEndPoint = function()
The steps to send point-to-point Vulcan message are as follows:
- Get all available endpoints.
- Select the destination endpoint and get appId and appVersion from it.
- Create Vulcan message with the destination
appIdandappVersion. - Dispatch Vulcan message.
var endPointList = VulcanInterface.getEndPoints();
var destIndex = 0;
var appId = GetValueByKey(endPointList[destIndex], "appId");
var appVersion = GetValueByKey(endPointList[destIndex], "appVersion");
var message = new VulcanMessage(VulcanMessage.TYPE_PREFIX + "test", appId, appVersion);
message.setPayload("blablabla...");
VulcanInterface.dispatchMessage(message);
There are two separate JavaScript engines here.
- JavaScript engine of host application - Application DOM/Extend script DOM
- JavaScript engine of CEP HTML runtime - HTML DOM
Application DOM is not available in CEP extension's engine and CEP DOM is not available in host application's engine.
To access Application DOM from CEP extensions, CEP JavaScript library provides an API, CSInterface.evalScript, to execute extend script to access the host application's DOM. Here is a brief diagram to indicate how to access Application DOM through this API.
Here is the sample JavaScript code snippet in HTML extension.
var csInterface = new CSInterface();
csInterface.evalScript('app.documents.add();', function(result){
alert(result);
});
There is no way to access HTML extension's JavaScript DOM directly from Application's ExtendScript. If you need to access it, CEP event based communication can be used as a substitution.
CEP creates a library which uses External Object mechanism of ExtendScript to send CSXS events. The external object provides an ExtendScript class CSXSEvent for creating and dispatching application-level CSXS events. On HTML extension side, event listeners can be registered via the addEventListener API in CSInterface.js to listen to the events.
Some CC applications (Photoshop, Illustrator, Premiere Pro) integrated PlugPlugExternalObject library and started to support this functionality in CC 2014 release. Audition supports this functionality since CC 2015.1 release.
ExtendScript developers need to create external object instance first.
var externalObjectName = "PlugPlugExternalObject";
var mylib = new ExternalObject( "lib:" + externalObjectName );
And then create the CSXSEvent instance.
var eventObj = new CSXSEvent();
eventObj.type="documentCreated";
eventObj.data="blahblah";
At last use this instance to dispatch event:
eventObj.dispatch();
Below is the sample code of ExtendScript.
...
var cs = new CSInterface();
cs.addEventListener("documentCreated", function(event){
alert('Cool!' + event.data);
});
var extendScript = 'var externalObjectName = "PlugPlugExternalObject"; var mylib = new ExternalObject( "lib:" + externalObjectName ); app.document.add(); var eventObj = new CSXSEvent(); eventObj.type="documentCreated"; eventObj.data="blahblah"; eventObj.dispatch();'
cs.evalScript(extendScript);
For fly-out menu on the native panel of HTML extension, it has been supported.
Two new interfaces are added to CSInterface.
CSInterface.prototype.setPanelFlyoutMenu = function(menu){
window.__adobe_cep__.invokeSync("setPanelFlyoutMenu", menu);
};
CSInterface.prototype.updatePanelMenuItem = function(menuItemLabel, enabled, checked){
var ret = false;
if (this.getHostCapabilities().EXTENDED_PANEL_MENU){
var itemStatus = new MenuItemStatus(menuItemLabel, enabled, checked);
ret = window.__adobe_cep__.invokeSync("updatePanelMenuItem", JSON.stringify(itemStatus));
}
return ret;
};
The "menu" parameter for setPanelFlyoutMenu is a XML string. Below is an example:
<Menu>
<MenuItem Id="menuItemId1" Label="TestExample1" Enabled="true" Checked="false"/>
<MenuItem Label="TestExample2">
<MenuItem Label="TestExample2-1" >
<MenuItem Label="TestExample2-1-1" Enabled="false" Checked="true"/>
</MenuItem>
<MenuItem Label="TestExample2-2" Enabled="true" Checked="true"/>
</MenuItem>
<MenuItem Label="---" />
<MenuItem Label="TestExample3" Enabled="false" Checked="false"/>
</Menu>
If user wants to be notified when clicking a menu item, user needs to register com.adobe.csxs.events.flyoutMenuClicked event by calling AddEventListener. When a menu item is clicked, the event callback function will be called. The data attribute of event is an object which contains menuId and menuName attributes.
To get notified when fly-out menu is opened and closed, register event listener for below event types respectively:
"com.adobe.csxs.events.flyoutMenuOpened"
"com.adobe.csxs.events.flyoutMenuClosed"
There are three APIs in CSInterface for developers to set and update the customized context menu.
CSInterface.prototype.setContextMenu = function(menu, callback){
window.__adobe_cep__.invokeAsync("setContextMenu", menu, callback);
};
CSInterface.prototype.setContextMenuByJSON = function(menu, callback){
window.__adobe_cep__.invokeAsync("setContextMenuByJSON", menu, callback);
};
CSInterface.prototype.updateContextMenuItem = function(menuItemID, enabled, checked){
var itemStatus = new ContextMenuItemStatus(menuItemID, enabled, checked);
ret = window.__adobe_cep__.invokeSync("updateContextMenuItem", JSON.stringify(itemStatus));
};
The "menu" parameter for "setContextMenu" is a XML string.
- Id - Menu item ID. It should be plain text.
- Icon - Menu item icon path. It is a path relative to the extension root path. For optimal display results please supply a 16 x 16px PNG icon as larger dimensions will increase the size of the menu item.
- Label - Menu item label. It supports localized languages.
- Enabled - Whether the item is enabled or disabled. Default value is true.
- Checkable - Whether the item can be checked/unchecked. Default value is false.
- Checked - Whether the item is checked or unchecked. Default value is false.
- The items with icons and checkable items cannot coexist on the same menu level. The former take precedences over the latter.
Here is an example.
<Menu>
<MenuItem Id="menuItemId1" Label="TestExample1" Enabled="true" Checked="false" Icon="./img/small_16X16.png"/>
<MenuItem Id="menuItemId2" Label="TestExample2">
<MenuItem Id="menuItemId2-1" Label="TestExample2-1" >
<MenuItem Id="menuItemId2-1-1" Label="TestExample2-1-1" Enabled="false" Checkable="true" Checked="true"/>
</MenuItem>
<MenuItem Id="menuItemId2-2" Label="TestExample2-2" Enabled="true" Checkable="true" Checked="true"/>
</MenuItem>
<MenuItem Label="---" />
<MenuItem Id="menuItemId3" Label="TestExample3" Enabled="false" Checked="false"/>
</Menu>
The "callback" parameter is the callback function which is called when user clicks a menu item. The only parameter is the ID of clicked menu item.
If you prefer to using a JSON string to set context menu, you can achieve it by calling setContextMenuByJSON.
The "menu" parameter for setContextMenuByJSON is a JSON string.
- id - Menu item ID. It should be plain text.
- icon - Menu item icon path. It is a path relative to the extension root path. For optimal display results please supply a 16 x 16px PNG icon as larger dimensions will increase the size of the menu item.
- label - Menu item label. It supports localized languages.
- enabled - Whether the item is enabled or disabled. Default value is true.
- checkable - Whether the item can be checked/unchecked. Default value is false.
- checked - Whether the item is checked or unchecked. Default value is false.
- The items with icons and checkable items cannot coexist on the same menu level. The former take precedences over the latter.
Here is an JSON example
{
"menu": [
{
"id": "menuItemId1",
"label": "testExample1",
"enabled": true,
"checkable": true,
"checked": false,
"icon": "./img/small_16X16.png"
},
{
"id": "menuItemId2",
"label": "testExample2",
"menu": [
{
"id": "menuItemId2-1",
"label": "testExample2-1",
"menu": [
{
"id": "menuItemId2-1-1",
"label": "testExample2-1-1",
"enabled": false,
"checkable": true,
"checked": true
}
]
},
{
"id": "menuItemId2-2",
"label": "testExample2-2",
"enabled": true,
"checkable": true,
"checked": true
}
]
},
{
"label": "---"
},
{
"id": "menuItemId3",
"label": "testExample3",
"enabled": false,
"checkable": true,
"checked": false
}
]
}
If developers do not set context menu, CEP shows default items (Back, Forward, View Source, etc.). This is compatible with previous releases. If developers set context menu which has no any default id, CEP removes all default items and show customized items only. If one of the following default ids is set, the default menu item against that id will be shown.
| id |
|---|
| "print" |
| "back" |
| "view source" |
| "forward" |
Notes:
- They are case-insensitive.
- The quotation marks is not a part of an id.
- The default callback associated to each id will be used and cannot be customised.
To disable the context menu, you can call setContextMenu by null.
Another way is to add oncontextmenu="return false;" to the HTML tag. For example,
<body oncontextmenu="return false;">
- Register "com.adobe.csxs.events.panelWindowStatusChanged" CSXS event
- Call isWindowVisible JavaScript API.
- Observe
com.adobe.csxs.events.panelWindowStatusChangedCSXS event, this is for PANEL extensions only. If user hides the panel window by clicking "X" or collapsing window, this event is going to be sent to observer with the "true" or "false" string in data attribute, while the event is not going to be sent if the extension is closed. That is to say, currently, only panel extensions which are running on Ai and persistent can receive this event, when the extension is hiding, event with "false" data is sent while the extension is shown, event with "true" is sent.
- Call
isWindowVisibleJS interface. Both dialog and panel extension enable to access this API, but it always returns true for modal and modeless dialog extensions while it always returns false for invisible extensions.
Getting extension content size can be done using window.innerWidth and window.innerHeight. However, if you are accessing these properties from inside an IFrame, you are actually accessing the properties of the IFrame's window object, not the ones for the HTML document. To access the top-most one, you will need to do parent.window.innerWidth and parent.window.innerWidth.
Changing modal and modeless extension content size is supported in all Adobe applications that supports CEP. However, changing panel HTML extension size is not supported in Premiere Pro, Prelude, After Effects and Audition.
CSInterface.prototype.resizeContent = function(width, height)
The width and height parameters are expected to be unsigned integers. The function does nothing when parameters of other types are passed.
Please note that extension min/max size constraints as specified in the manifest file apply and take precedence. If the specified size is out of the min/max size range, the min or max bounds will be used. When a panel is docked with other panels, there are chances that it won't resize as expected even when the specified size satisfies the min and max constraints. The restriction is imposed by host applications, not by CEP.
(Since 6.1) Register the invalid certificate error callback for an extension. This callback will be triggered when the extension try to access the web site that contains the invalid certificate on main frame. But if the extension does not call this function and try to access the web site containing the invalid certificate, a default error page will be shown:
CSInterface.prototype.registerInvalidCertificateCallback = function(callback)
(Since 6.1) Register an interest in some key events to prevent them from being sent to the host application:
CSInterface.prototype.registerKeyEventsInterest = function(keyEventsInterest)
This function works with modeless extensions and panel extensions. Generally all the key events will be sent to the host application for these two extensions if the current focused element is not text input or dropdown.
If you want to intercept some key events and you want them to be handled in the extension, please call this function in advance to prevent them being sent to the host application.
- keyEventsInterest: A JSON string describing those key events you are interested in. A null object or an empty string will lead to removing the interest
This JSON string should be an array, each object has following keys:
- keyCode: [Required] represents an OS system dependent virtual key code identifying the unmodified value of the pressed key.
- ctrlKey: [optional] a Boolean that indicates if the control key was pressed (true) or not (false) when the event occurred.
- altKey: [optional] a Boolean that indicates if the alt key was pressed (true) or not (false) when the event occurred.
- shiftKey: [optional] a Boolean that indicates if the shift key was pressed (true) or not (false) when the event occurred.
- metaKey: [optional] (macOS Only) a Boolean that indicates if the Meta key was pressed (true) or not (false) when the event occurred. On Macintosh keyboards, this is the command key. To detect Windows key on Windows, please use keyCode instead.
To learn all key codes:
- Windows
- macOS
/System/Library/Frameworks/Carbon.framework/Versions/A/Frameworks/HIToolbox.framework/Versions/A/Headers/Events.h- Install Key Codes from the Mac App Store.
An example JSON string:
[{
"keyCode": 48
},
{
"keyCode": 123,
"ctrlKey": true
}]
(Since 6.1) CEP 6.1 introduces two APIs to set and get the title of extension windows. Those functions work with modal and modeless extensions in all Adobe products, and panel extensions in Photoshop, InDesign, InCopy, Illustrator, Animate (Flash Pro) and Audition:
CSInterface.prototype.setWindowTitle = function(title){
window.__adobe_cep__.invokeSync("setWindowTitle", title);
};
CSInterface.prototype.getWindowTitle = function(){
return window.__adobe_cep__.invokeSync("getWindowTitle", "");
};
CEP JavaScript library provides APIs for detecting the availability of HI-DPI display on the macOS platform.
CSInterface.getScaleFactor()Use this function to retrieve the scale factor of the display on which the calling extension window is located.var scaleFactor = CSLibrary.getScaleFactor();CSInterface.setScaleFactorChangedHandler()Use this function to add a event handler that will be called when calling extension window is moved between HI-DPI and non-HI-DPI displays.
window.scaleFactorHandler = function(){
var scaleFactor = CSLibrary.getScaleFactor();
if (scaleFactor === 2){
imgSrc = "../img/PS_AppIcon_r.png"
} else {
imgSrc = "../img/PS_AppIcon.png"
}
document.getElementById("image").src = imgSrc;
}
CSLibrary.setScaleFactorChangedHandler(window.scaleFactorHandler);
CEP 5.2 has already supported HiDPI on Windows.
The JavaScript engine in CEP HTML engine had been extended to provide some APIs, including:
- local file access
- native process
- others
These APIs are in JavaScript DOM and can be used as other built-in JavaScript APIs. You do NOT need to include any JavaScript files.
API reference is as below.
CEPEngine_extensions.js is actually a CEF extension that is built in CEPHtmlEngine to expand the DOM of CEPHtmlEngine, like create/delete folder, read/write file, create/quit process, and so on. You can invoke these built-in APIs directly in your HTML extension without any JavaScript file reference.
For example, you want to
- create a folder.
var path = "/tmp/test";
var result = window.cep.fs.makedir(path);
if (0 == result.err){
...// success
} else {
...// fail
}
- write a file.
var data = "This is a test.";
var path = "/tmp/test";
var result = window.cep.fs.writeFile(path, data);
if (0 == result.err){
...// success
} else {
...// fail
}
- Write file with base64 encoding mode. To use this mode, you need to convert the input string to a base64-encoded string before calling
writeFile(). The following is an example.
var data = "This is a test.";
var path = "/tmp/test";
data = cep.encoding.convertion.utf8_to_b64(data);
var result = window.cep.fs.writeFile(fileName, data, cep.encoding.Base64);
if (0 == result.err) {
...// success
} else {
...// fail
}
- read a file.
var path = "/tmp/test";
var result = window.cep.fs.readFile(path);
if(result.err === 0){
//success
alert(result.data); //result.data is file content
} else {
...// fail
}
- Read file with base64 encoding mode in which the read data after
readFilecalled is converted to a base-encoded string. You need to decode this string to any format you want. The following is an example
var path = "/tmp/test";
result = window.cep.fs.readFile(path, cep.encoding.Base64);
if(result.err === 0){
//success
var base64Data = result.data;
var data = cep.encoding.convertion.b64_to_utf8(base64Data);
} else {
...// fail
}
- Create a process and check if it's running.
var result = window.cep.process.createProcess("usr/X11/bin/xterm");
if(result.err === 0){
var pid = result.data;
result = window.cep.process.isRunning(pid);
if(result.data === true){
// running
}
}
You could use other APIs like delete folder, rename folder, set file permission, delete file, show file open dialog, quit process, etc.
We have the following samples that demonstrate use of some of these APIs:
- https://github.com/Adobe-CEP/Samples/tree/master/Flickr
- https://github.com/Adobe-CEP/Samples/tree/master/Collage
- https://github.com/Adobe-CEP/CEP-Resources/tree/master/CEP_10.x/Samples
In order to support localization, both the extension and the host application must provide locale information. There are two distinct types of locale information.
- The License Locale (returned as the applicationLocale by the AMT library)
- The Effective/Language/UI Locale (which is controlled by the user in the OS settings).
The extension must provide the list of supported locales for both the License Locale and the Language Locale via the HostEnvironment. This is particularly important in cases where the extension has features for a specific locale. PlugPlug library expects the host application to provide the locale information as part of the environmental data.
JavaScript API HostEnvironment has the appLocale property in place.
CEP checks the License Locale of host application against supported locales declared in extensions locale list to determine if the extension is loadable for the host application.
Each property file should be placed in its corresponding locale folder. For example, the en_US property file should be <YourExtension>/locale/en_US/messages.properties. Users can define a default property file (<YourExtension>/locale/messages.properties), which will be used when the corresponding locale file is not defined.
YourExtension/
|-csxs/
|-locale/ <-- Directory for localized resources
|-- messages.properties <-- The one to fallback to if no localized resources is provided for a locale
|-- en_US/
| |- messages.properties
|-- zh_CN/
|- messages.properties
A locale file contains multiple lines of <key>=<value>. There should be a new line below the last property key/value.
key1=value1
key2=value2
key3.value=value3
key4.innerHTML=value4
CEP provides a JS interface named initResourceBundle to initialize the locale resources. This should be called during the loading of the extension. CEP initializes the resource bundle for the extension with property values for the current application and UI locale. Then users can access the resource bundle (object) to get the localized strings.
var csInterface = new CSInterface();
csInterface.initResourceBundle();
CEP 6.0 provided a mechanism to allow multiple locales to share the same localization resources (messages.properties).
For example, you want es_MX to use the messages.properties for es_ES. To do so, supply a file named fallback.properties in the es_MX folder as illustrated below.
YourExtension/
|-csxs/
|-locale/
|-- messages.properties
|-- es_ES/
| |- messages.properties
|-- es_MX/
|- fallback.properties
In the fallback.properties file, you specify which locale's localized resources you want es_MX to use, in below format
fallback=es_ES
Side Notes:
- The
fallback.propertiesfile takes precedence when bothmessages.propertiesandfallback.propertiesexist at the same time. - If
fallback.propertiesis malformed, or it specifies a non-existent fallback locale, the messages.properties file in the same directory will be used.
In manifest, it supports to use locale string as menu. For example, in ShareOnBehance's manifest, it is using %UI_share_on_Behance. UI_share_on_Behance is defined as UI_share_on_Behance=xxx in messages.properties.
<UI>
<Type>ModalDialog</Type>
<Menu>%UI_share_on_Behance</Menu>
...
</UI>
var cs = new CSInterface();
// Get properties according to current locale of host application.
var resourceBundle = cs.initResourceBundle();
// Use the localized strings.
<script type="text/javascript">document.write(resourceBundle.key1);</script>
data-locale is the custom HTML element attribute and you can add to each HTML element that you want to localize.
In this example, there is key3.value=value3 in the property file. In the HTML file, the input widget has attribute data-locale with key3, then its value is set to value3.
In this example, there is key4.innerHTML=value4 in the property file. In the HTML file, the text area widget has attribute data-locale with key4, then its innerHTML is set to value4.
<script type="text/javascript">
var cs = new CSInterface();
// Get properties according to current locale of host application.
var resourceBundle = cs.initResourceBundle();
// Use the localized strings.
document.write(resourceBundle.key1);
document.write(resourceBundle.key2);
</script>
<input type="submit" value="" data-locale="key3"/>
<textarea rows="10" cols="80" data-locale="key4"></textarea>
Use parameters ($1, $2, ...) in localized strings.
var localize = function(key){
var cs = new CSInterface();
var resourceBundle = cs.initResourceBundle();
var localizedStr = resourceBundle[key];
if (localizedStr){
var index = 1;
while (localizedStr.indexOf("$" + index) !== -1){
localizedStr = localizedStr.replace("$" + index, arguments[index]);
index++;
}
return localizedStr;
} else {
return '';
}
};
MENA stands for "Middle East and North Africa". Support needs to be provided for Arabic, Hebrew, and NA French languages.
| Language | ISO Code |
|---|---|
| Arabic (Middle East Enabled English Arabic) | en_AE |
| Hebrew (Middle East Enabled English Hebrew) | en_IL |
| NA French | fr_MA |
If an extension needs to be loaded in host applications in MENA locales, MENA locales must be added to the supported locale list of the extension manifest file. For example:
<LocaleList>
...
<Locale Code="en_AE"/>
<Locale Code="en_IL"/>
<Locale Code="fr_MA"/>
...
</LocaleList>
Suppose your extension has this directory layout
Extension/
|-csxs/
|-locale/ <-- Directory for localized resources
|-- messages.properties <-- The one to fallback to if no localized resources is provided for a locale
|-- fr_FR/
| |- messages.properties
|-- en_GB/
|- messages.properties
When CSInterface.initResourceBundle() is called, CEP uses the app UI locale (not app locale) reported by PP to load localized resources. If there is no localized resources for an app UI locale, for example fr_MA, then CEP will fall back to use messages.properties located under the locale folder.
With MENA feature, point products map en_AE/en_IL to en_US and fr_MA to fr_FR for app UI locale. In this case, for en_AE and en_IL build of point product, en_US resources will be used if provided and for fr_MA build of point product, fr_FR resources will be used if provided. What extension team needs to do in this case is to provide en_US version of resources for en_AE/en_IL and fr_FR version of resources for fr_MA.
CEP 5.0 and higher versions supports playing video and audio encoded in below formats.
| Format | MIME-Type | Misc. |
|---|---|---|
| MP4 | video/mp4 | MPEG 4 files with H.264 video codec and AAC audio codec |
| Ogg | video/ogg | Ogg files with Theora video codec and Vorbis audio codec |
| mp3 | audio/mpeg |
Here is an example of playing video in your extension:
<video poster="http://www.html5rocks.com/en/tutorials/video/basics/star.png" controls>
<source src="http://www.html5rocks.com/en/tutorials/video/basics/Chrome_ImF.mp4" type='video/mp4; codecs="avc1.42E01E, mp4a.40.2"'/>
</video>
One thing to note is that because HTML extensions are hosted in integrating application's windows, video cannot be played in full-screen mode.
WebRTC is targeting to serve stream audio, video capture, like online video conference. WebRTC is not enabled by default. To enable it, the schema below need to be added in manifest file. For details, refer to Customize CEF command parameters.
<CEFCommandLine> <Parameter>--enable-media-stream</Parameter> </CEFCommandLine>
For WebRTC related development, CEP runtime just keeps the same experiences as the usage in Chrome. Please refer to samples for the implementation.
On macOS, scroll bars of panel are hidden by OS (since Lion by design). It can be always shown by settings as below.
-
Click the Apple menu at the top-left of the screen, then select System Preferences.
-
Next, select the General preferences pane; it’s the very first one, up at the top.
-
Under the “Show scroll bars” heading, you’ll find three options: “Automatically based on input device,” “When scrolling,” and “Always.” Chose "Always."
An HTML extension can be invisible during its whole life cycle. This means
- It always runs in the background
- It is never visible
To make an HTML extension invisible
- Set extension manifest version to "5.0" or higher.
- Specify its window type as 'Custom' in the manifest file.
- Set
<AutoVisible>to false in the manifest file. - If you do not want the extension to appear in the Window->Extensions menu, do not add the
<Menu>tag. - If you want the extension to start on specific types of events, specify those events using
<StartOn>tag.
Here is an example:
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<ExtensionManifest xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" ExtensionBundleId="IamInvisible" ExtensionBundleVersion="1.0" Version="5.0">
<ExtensionList>
<Extension Id="IamInvisible" Version="1.0"/>
</ExtensionList>
<ExecutionEnvironment>
<HostList>
<Host Name="PHXS" Version="13.0"/>
</HostList>
<LocaleList>
<Locale Code="All"/>
</LocaleList>
<RequiredRuntimeList>
<RequiredRuntime Name="CSXS" Version="5.0"/>
</RequiredRuntimeList>
</ExecutionEnvironment>
<DispatchInfoList>
<Extension Id="IamInvisible">
<DispatchInfo>
<Resources>
<MainPath>./html/index.html</MainPath>
</Resources>
<Lifecycle>
<AutoVisible>false</AutoVisible>
<StartOn>
<!-- Photoshop dispatches this event on startup -->
<Event>applicationActivate</Event>
<!-- Premiere Pro dispatches this event on startup -->
<Event>com.adobe.csxs.events.ApplicationActivate</Event>
<!-- You can add more events -->
<Event>another_event</Event>
</StartOn>
</Lifecycle>
<UI>
<Type>Custom</Type>
<Geometry>
<Size>
<Height>1</Height>
<Width>1</Width>
</Size>
</Geometry>
</UI>
</DispatchInfo>
</Extension>
</DispatchInfoList>
</ExtensionManifest>
One important thing to note is that not all host applications support Invisible HTML Extension. See table below for more information:
| Point Product | Supports Invisible Extension | Misc. |
|---|---|---|
| Photoshop | Yes | |
| Premiere Pro | Yes | |
| Prelude | Yes | |
| Animate (Flash Pro) | Yes | |
| Audition | Yes | |
| InDesign | Yes | However, as of 2019 you must include a width and height. |
| InCopy | Yes | However, as of 2019 you must include a width and height. |
| Illustrator | Yes |
Chromium/CEF command line parameters can be passed to CEPHtmlEngine, like --enable-media-stream. Available Chromium command line parameters.
CEP filters out some parameters due to various reasons:
| Parameters | What is it filtered out? |
|---|---|
| --remote-debugging-port | This could overwrite the one in .debug file. Filter out to avoid conflict. |
| --ignore-certificate-errors | This ignores SSL certificate errors. It is a security concern to ignore invalid server certificate, which allows extensions to load files from malicious sites. |
All other parameters are passed to underlying CEF. It is up to CEF to decide whether a parameter is supported and what is the behavior.
- Add
<CEFCommandLine><Parameter>--param1<Parameter/> ... </CEFCommandLine>in manifest. - For
key=valueparameter, add<CEFCommandLine><Parameter>--param1=value1<Parameter/> ... </CEFCommandLine>in manifest.
Here is an example:
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<ExtensionManifest xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" ExtensionBundleId="xx.yy.zz" ExtensionBundleVersion="1.0" Version="5.0">
<ExtensionList>
<Extension Id="xx.yy.zz" Version="1.0"/>
</ExtensionList>
<ExecutionEnvironment>
<HostList>
<Host Name="PHXS" Version="13.0"/>
<Host Name="PPRO" Version="6.0"/>
</HostList>
<LocaleList>
<Locale Code="All"/>
</LocaleList>
<RequiredRuntimeList>
<RequiredRuntime Name="CSXS" Version="5.0"/>
</RequiredRuntimeList>
</ExecutionEnvironment>
<DispatchInfoList>
<Extension Id="xx.yy.zz">
<DispatchInfo>
<Resources>
<MainPath>./html/index.html</MainPath>
<CEFCommandLine>
<Parameter>--enable-media-stream</Parameter>
</CEFCommandLine>
</Resources>
...
</DispatchInfo>
</Extension>
</DispatchInfoList>
</ExtensionManifest>
| Parameters | Notes |
|---|---|
| --enable-media-stream | Enable media (WebRTC audio/video) streaming. |
| --enable-speech-input | Enable speech input (x-webkit-speech). |
| --persist-session-cookies | Persist session cookies. |
| --disable-image-loading | Disable loading of images from the network. A cached image will still be rendered if requested. |
| --disable-javascript-open-windows | Disable opening of windows via JavaScript. |
| --disable-javascript-close-windows | Disable closing of windows via JavaScript. |
| --disable-javascript-access-clipboard | Disable clipboard access via JavaScript. |
| --enable-caret-browsing | Enable caret browsing. |
| --proxy-auto-detect | This tells Chrome to try and automatically detect your proxy configuration. See more info at http://www.chromium.org/developers/design-documents/network-settings. |
| --user-agent | A string used to override the default user agent with a custom one. |
| --disable-application-cache | Disable the ApplicationCache. |
| --enable-nodejs | Enable Node.js APIs in extensions. Supported since CEP 6.1. |
| --disable-pinch | Disable compositor-accelerated touch-screen pinch gestures. |
| --mixed-context | Enable the "mixed context" mode. Supported since CEP 7.0. |
The purpose of persistent is to force not reload HTML extension when it is closed or hidden. Photoshop has provided persistent since the version 14.2. Please note that, this feature is product specific and may not be available in few host applications.
InDesign and InCopy support perisitent feature from version 13.1. Following are the specific event names to be used: Indesign: com.adobe.InDesignPersistent com.adobe.InDesignUnPersistent Incopy: com.adobe.InCopyPersistent com.adobe.InCopyUnPersistent
To make an HTML extension persistent in Photoshop
- Upgrade Photoshop version to 14.2 or later
- Dispatch the event
com.adobe.PhotoshopPersistentfrom HTML extension to Photoshop to request persistent
Sample code:
var Persistent = function(inOn){
if(inOn){
var event = new CSEvent("com.adobe.PhotoshopPersistent", "APPLICATION"); //Photoshop
//var event = new CSEvent("com.adobe.InDesignPersistent", "APPLICATION"); //Indesign
//var event = new CSEvent("com.adobe.InCopyPersistent", "APPLICATION"); //Incopy
} else {
var event = new CSEvent("com.adobe.PhotoshopUnPersistent", "APPLICATION"); //Photoshop
//var event = new CSEvent("com.adobe.InDesignUnPersistent", "APPLICATION"); //Indesign
//var event = new CSEvent("com.adobe.InCopyUnPersistent", "APPLICATION"); //Incopy
}
event.extensionId = <extension id as per manifest definition>;
csInterface.dispatchEvent(event);
}
Persistent(true); //persistent to prevent extension from unloading
…
Persistent(false); //unpersistent
In HTML extension, URL link could be opened in the default browser by calling window.cep.util.openURLInDefaultBrowser(‘http://example.com'):
<li><button onclick="window.cep.util.openURLInDefaultBrowser('http://www.adobe.com')">Open browser</button></li>
When an extension is about to be unloaded, the communication channel between extension and point product has been torn down, so there's no easy way for point product to get the extension-related information such as the extension's status.
To fulfill this functionality, the cooperation between extension and point product is needed.
From extension's side, a JavaScript function registerExtensionUnloadCallback is provided to register its callback function. The definition can be found at https://github.com/Adobe-CEP/CEP-Resources/blob/CEP8/CEP_9.x/CEPEngine_extensions.js. There're some limitations as below:
- If registerExtensionUnloadCallback is called more than once, the last callback function that's successfully registered will be used.
- The callback function must be context-independent and it is executed in the mainframe's context.
Example
window.cep.util.registerExtensionUnloadCallback(function(){
window.cep.fs.writeFile("D:\\1.txt", "Hello");
});
From point product's side, csxs::event::EVENT_TYPE_CSXS_EXTENSION_UNLOADED ("com.adobe.csxs.events.ExtensionUnloaded") is dispatched. When point product receives the event, it can read the file written by extension to get the extension-related information.
Since the execution in the extension and the point product is in parallel, the point product needs to take care of the order.
The above method does not work due to the multi-process architecture in CEF 3.
One of the most prominent feature in CEP 5.0 and higher versions are allowing Node.js APIs to be used in HTML extensions. Most of the built-in APIs in Node.js version 0.8.22 are available to HTML extensions, with the below exceptions:
- Cluster APIs are not supported.
- Console APIs are not supported on Windows.
Other things to note:
- CEP injects following symbols into the root HTML DOM:
- global, GLOBAL, root - same with the window object
- Buffer - node's Buffer class
- process - node's process class
- require - the magic function that bring you node API
- module - in node the main script is a special module, this is for the compatibility
- Conflicts with Web-Based Require Function
- If your app uses libraries like
RequireJSthat inserts a require function into DOM, you should consider renaming CEP's require function to something else before migrating.
- If your app uses libraries like
<script type="text/javascript">windows.nodeRequire=window.require && window.require=undefined</script>
<script type="text/javascript" src="your/require/js/file.js"></script>
- Conflicts with Web-Based module Function
- If your app uses
JQuerythat is trying to register itself asnodejsplugin, then you will have to add the script below inside script tag to define window.module as undefined.
- If your app uses
<script type="text/javascript">window.module = undefined</script>
- Disable Node.js APIs in iframe
- Because of security consideration, CEP provides an option to disable Node.js APIs in iframe. To do so, add a
nodejs-disabled="true"attribute to iframe tag. For example:
- Because of security consideration, CEP provides an option to disable Node.js APIs in iframe. To do so, add a
<iframe id="xxx" class="xxxxx" nodejs-disabled="true">
- Forcing the environment implementation. If you are using
RequireJS, and the text plugin tries to detect what environment is available for loading text resources,Node,XMLHttpRequest (XHR)orRhino, but sometimes theNodeorRhinoenvironment may have loaded a library that introduces anXHRimplementation. You can force the environment implementation to use by passing an"env"module config to the plugin:
requirejs.config({
config: {
text: {
//Valid values are 'node', 'xhr', or 'rhino'
env: 'rhino'
}
}
});
All third-party node JavaScript modules are supported. The root search path of third-party modules is the directory which contains your HTML file. For example, when you do require in file:///your_extension/index.html, CEP will lookup modules under file:///your_extension/node_modules, this rule is exactly the same with upstream node.
Node.js native modules are not directly supported since CEP is using a different V8 version from the official node.
process.env.ENV_VARIABLE // ENV_VARIABLE is the name of the variable you want to access.
CEP 6.1 upgraded its HTML engine to CEF 2272 (based on Chromium 41.0.2272.104) with IO.js version 1.2.0 integrated.
Other things to note:
- Node.js APIs are disabled by default
- Due to security consideration, node.js APIs are disabled by default (prior to CEP 6.1, they were enabled by default) both on the extension level and iFrame level.
To enable Node.js APIs:
- Set
ExtensionManifestversion andRequiredRuntimeversion 5.0 or higher. - Specify
--enable-nodejsin extension manifest. See section Customize CEF Command Line Parameters for details. - To use
Node.jsAPIs in IFrames, add propertyenable-nodejsto it and to all its ancestor iFrames. If any of its ancestors don't have this property specified,Node.jsAPIs won't work
<iframe id="xxx" class="xxxxx" enable-nodejs>
- The old
nodejs-disabledCEF command line parameter and iFrame property are no longer supported and ignored by the new HTML engine. - Node context and Browser context
- The way
io.jswas integrated into CEF introduced two types of JavaScript contexts, one for browser, the other forio.js. Global objects created in HTML pages are in browser context, whilerequiredjs files run inio.jscontext. These two contexts don't have direct access to each other's data. To share data, pass reference to objects between the two contexts:- Accessing objects in
io.jscontext from browser context - For example, in browser context,var backbone = require('backbone');executes the backbone module's code and then pass the result object to browser context. - Accessing objects in browser context from
io.jscontext - Browser context'swindowglobal object is injected toio.jscontext, providing a way to access objects in browser context fromio.jscontext. For example, if you want to access a global object namedlocaleStringsdefined in browser context from yourio.jsmodule, usewindow.localStringsin yourio.jsmodule.
- Accessing objects in
- The way
In addition to the existing Node.js support in CEP 6.1, CEP 7.0 provided a new mixed context Node.js mode. Unlike the separate context mode in CEP 6.1 where a required node module is in a separate JavaScript context, a required node module and the JavaScript code that requires it are in the same context in the new mixed context mode, eliminating all the inconveniences in the old separate context mode.
This mode is disabled by default. To enable it, add command line parameter --mixed-context to your extension manifest.
<Parameter>--mixed-context</Parameter>
In addition to existing Node.js support in CEP 9, CEP 10 provide a new disable-nodejs attribute for iframe. When this attribute is used in iframe, Node.js will be disabled in separate context and mixed-context both.
<iframe name="iframe-1">
\\ Node will be accesible here
</iframe>
<iframe name="iframe-2" disable-nodejs>
\\ Node will not be accesible here
</iframe>
There is a known limitation of cep.process.stdout/stderr which is targeting to capture one time of stdout/stderr output.
For applications that has not integrated CEP 5, there are two workarounds suggested as the following.
- Embed cep.process.stdout/stderr:
var getSTDOutput = function(){
console.log("getSTDOutput");
window.cep.process.stdout(pid, function(output){
console.log(output);
// your code is here
});
var result = window.cep.process.isRunning(pid);
if (result.data === true){
setTimeout(getSTDOutput, 1000);
}
}
- Join all stdout output as one, like below
var str1 = 'abcdef';
var str2 = '12345';
var str3 = 'gghhtt';
console.log(str1 + str2 + str3);
An example on how to get curl downloading progress through stderr:
<script>
var downloadPid = -1;
function getStdErrOutput()
{
window.cep.process.stderr(downloadPid, function(progress) {
var keys = progress.split(new RegExp('[# ]', 'g'));
for(i=0; i<keys.length; i++){
if (keys[i] != '') {
console.log(keys[i]);
}
}
});
var result = window.cep.process.isRunning(downloadPid);
if (result.data == true)
{
setTimeout(getStdErrOutput, 100);
}
}
var doDownload = function() {
var qURL = 'http://code.jquery.com/jquery-1.11.0.min.js';
var dest = '/tmp/test.js';
console.log("ext download (curl) " + qURL + " " + dest);
var result = window.cep.process.createProcess('/usr/bin/curl', qURL, '-#', '-o', dest);
downloadPid = result.data;
console.log("download pid is " + downloadPid);
getStdErrOutput();
};
doDownload();
</script>
Since CEP 5, Node.js is integrated into CEP runtime and users could invoke the standard APIs of Node.js in extension directly. For applications that has integrated CEP 5, refer to http://nodejs.org/api/process.html on how to use the global process object in Node.js.
Sample extension is at https://github.com/Adobe-CEP/Samples/tree/master/ExchangeExample.
JSX files normally define some functions and objects which are intended to be executed in host application's ExtendScript environment.
There are two approaches for HTML extension to trigger JSX files to be loaded in host application's ExtendScript.
- Define
<ScriptPath>node inmanifest.xml, and the value of the node is the relative path for the JSX file. For example:
Extension Id="com.adobe.CEPHTMLTEST.Panel1">
<DispatchInfo>
<Resources>
<MainPath>./html/index.html</MainPath>
<CEFCommandLine>
<Parameter>--enable-speech-input</Parameter>
<Parameter>--enable-media-stream</Parameter>
</CEFCommandLine>
<ScriptPath>./jsx/example.jsx</ScriptPath> <!-- ExtensionRootPath/jsx/example.jsx -->
</Resources>
...........................
</DispatchInfo>
</Extension>
- HTML extension uses CEP JavaScript API
evalScriptwith the script$.evalFile(jsxFile)to trigger JSX files to get loaded in host application's ExtendScript. Developers may need more script code to determine the path ofjsxFilebefore theevalScriptcall. For example, they may refer to$.fileNameto find out the path and the value of$.fileNameshould be the currently executedJSXfile path. Here is an example:
var extensionPath = $.fileName.split('/').slice(0, -1).join('/') + '/'; // The value of $.fileName should be ExtensionRootPath/jsx/example.jsx as mentioned above, while the value of extensionPath should be "ExtensionRootPath/jsx/"
$.evalFile(extensionPath + 'example1.jsx');
$.evalFile(extensionPath + 'example2.jsx');
$.evalFile(extensionPath + 'example3.jsx');
But if the $.fileName is referred in the FIRST LOADED JSX file, the value is not correct. That is to say, if the snippet above runs in example.jsx which is referred in the manifest.xml, the error will arise. So, PLEASE AVOID using $.fileName in the FIRST LOADED JSX file, maybe this is a limitation in ExtendScript. The workaround is to refer it in the second loaded and afterward JSX files. For example:
// After finishing loading the jsx file refered in the manifest.xml, please use evalScript of CSInterface to load other jsx files.
// "anotherJSXFile" is not the first loaded jsx file, so the value of "$.fileName" in it's stage is correct.
CSInterface.evalScript('$.evalFile(anotherJSXFile)', callback);
// Or in the first loaded jsx file, load another jsx file, and the value of "$.fileName" is correct in this file.
// Given the code is running this example.jsx which is referred in the manifest.xml.
// In the stage of "hardCodeJSXFile", the value of "$.fileName" is correct too.
$.evalFile(hardCodeJSXFile);
Note: Since the script in this section is intended to be executed in host application's ExtendScript, $ in this section is different from jquery global variable $.
Please use "namespace" if the developers want to define new variable/function/object in Global Space or $ object. If the same name defined in multiple JSX files, the definition in the last loaded JSX file will take effect, and the definition in the previous loaded JSX files will be overridden. For example, $.ext is defined in a.jsx, b.jsx and c.jsx, and in a.jsx $.ext is a function, in b.jsx $.ext is an object and in c.jsx $.ext is a string, and the load sequence is a.jsx->b.jsx->c.jsx, after loading, $.ext is a string, rather than an object or a function. And this behavior will be across multiple extension running in the same point product, for example, if a.jsx, b.jsx and c.jsx belong to extension a,b,c separately, and extension loading order is extension a-> extension b-> extension c, $.ext will be still a string, rather than a function or an object.
CEP 5.2 support HTML 5 Drag and Drop. There are four types.
-
Drag and drop inside HTML extension.
-
Drag and drop between two HTML extensions
-
Drag and drop between HTML extension and its host application.
-
Drag and drop between HTML extension and operating system (e.g. Desktop or Browser).
To learn about HTML 5 Drag and Drop and how to use it by JavaScript, please refer to http://www.w3.org/TR/html5/editing.html#dnd.
Here are some demos.
Extension developers can disable the default behavior of DnD by JavaScript.
Method 1
<body ondragover="return false" ondrop="return false">
Method 2 (using jQuery)
$(document.body).on('dragover drop', function(e) {
e.preventDefault();
});
Please read HTML 5 standard for more details. http://www.w3.org/TR/html5/editing.html#event-dragenter
CEP HTML Engine does not restrict using any extension JavaScript libraries. As long as a library can be used in CEF Client or Chrome browser, it should be usable in CEP HTML Engine.
Here are some JavaScript which had been used successfully
- JQuery - http://jquery.com/
- Please refer to Node.js section about resolving symbol conflicts.
- RequireJS - http://requirejs.org/
- Please refer to Node.js section about resolving symbol conflicts.
- spin.js - http://fgnass.github.com/spin.js/
- Modernizr - http://modernizr.com/
- Modernizr is a JavaScript library that detects HTML5 and CSS3 features in the user’s browser.
The support for NPAPI plug-ins had been dropped since CEP 4.2. Although it is still supported by CEP 4.0 and 4.1, but as all the products are moving to CEP 4.2, we suggest developers not use it.
Don't embed Java Applet in your extension as they all depend on NPAPI plug-in.
There are couples of JavaScript ways to increase or decrease font size in HTML panel either by plain JavaScript or JQuery. The following is the two pieces of sample snippet to achieve this. One is in plain JavaScript and the other uses JQuery library.
- Plain JavaScript - Use
document.body.style.fontSizeto change font size in page.
<script type="text/javascript" language="javascript">
window.onload = function() {
var fontchange = document.createElement("div");
var fontchangelink = function(fontsize, desc) {
var a = document.createElement('a');
a.href="#";
a.style.margin = "5px";
a.onclick = function() {
document.body.style.fontSize = fontsize + 'pt';
};
a.innerHTML = desc;
return a;
};
fontchange.appendChild(document.createTextNode("Change font size:"));
fontchange.appendChild(fontchangelink(9, "1"));
fontchange.appendChild(fontchangelink(11, "2"));
fontchange.appendChild(fontchangelink(13, "3"));
document.body.insertBefore(fontchange, document.body.childNodes[0]);
};
</script>
- JQuery - Use
$('html').css('font-size', size)to change font size in page.
<script type="text/javascript" language="javascript">
$(document).ready(function(){
// Reset Font Size
var originalFontSize = $('html').css('font-size');
$(".resetFont").click(function(){
$('html').css('font-size', originalFontSize);
});
// Increase Font Size
$(".increaseFont").click(function(){
var currentFontSize = $('html').css('font-size');
var currentFontSizeNum = parseFloat(currentFontSize, 10);
var newFontSize = currentFontSizeNum*1.2;
$('html').css('font-size', newFontSize);
return false;
});
// Decrease Font Size
$(".decreaseFont").click(function(){
var currentFontSize = $('html').css('font-size');
var currentFontSizeNum = parseFloat(currentFontSize, 10);
var newFontSize = currentFontSizeNum*0.8;
$('html').css('font-size', newFontSize);
return false;
});
});
</script>
The ways above are the common used solution for JavaScript developers to increase or decrease fonts.
if (navigator.onLine === true)
{
//system is online
}
else
{
//system is offline
}
Please refer to http://jsfiddle.net/BfLAh/1390/
$(document).mousemove(function(e){
$("#image").css({left:e.pageX, top:e.pageY});
});
Due to the X-Frame-Options header a number of HTTPS websites are unavailable to host in an iframe. An alternative to displaying HTTPS content is to use the window.location.href e.g.
iFrame alternative for HTTPS content
// html
<body onLoad="onLoaded()">
<!--iframe src="https://www.trello.com"></iframe--> <!-- this line can be deleted as HTTPS blocks content being displayed in an iframe -->
</body>
// javascript
function onLoaded() {
window.location.href = "https://www.trello.com";
}
CEP 5.2 supports HTML title attribute to show the tooltip on Windows. However, it's not supported on macOS due to off-screen rendering. The alternative is use JavaScript instead, please refer to http://www.a2zwebhelp.com/bootstrap-tooltips for good examples.
CEP and the underlying Chromium have multi-process architecture. Each CEP extension runs in a separate CEP HTML Engine, which might have three processes - one main/browser process, one renderer process, and one GPU process. On Max OSX, the main process is named CEPHtmlEngine, and the other two process are named CEPHtmlEngine Helper. On Windows, all the three processes are named CEPHtmlEngine.exe. As long as the extensions are well implemented, having multiple processes should not be a performance problem.
You can find what extension a CEP HTML Engine process belongs to by checking its command line parameters. On Mac OSX, please use ps -ef | grep CEPHtmlEngine in a terminal to check the command line. On Windows, please refer to http://superuser.com/questions/415360/how-do-i-find-out-command-line-arguments-of-a-running-program to check the command line. The browser process does not have --type in the command line, while the renderer process and the GPU process have --type=renderer and --type=gpu-process in the command line.
If you use TCPView to monitor CEPHtmlEngine process, you may see some ports are opened in localhost. Most of the ports are opened internally in Chromium code for websocket, which is initiated by HTML extension instead of CEPHtmlEngine itself. You can use RawCap to capture the data in *.pcap file and open it in Wireshark to examine the details.
Developers can use the following forums and github repositories to ask questions and create bugs/feature requests
- Getting Started Guides
- https://forums.adobe.com/community/creativesuites/extensionbuilder
- https://github.com/Adobe-CEP/CEP-Resources/issues
- https://github.com/Adobe-CEP/Samples/issues
- https://www.adobeprerelease.com
If you like to contribute for extension development, please raise your PR at https://github.com/Adobe-CEP