salve-dom is a library able to validate XML documents stored in DOM trees according to a Relax NG schema. It uses salve to perform the validation and has the same general limitations. This library was extracted from the wed XML editor. The algorithms used here have been in production for years.

Note that the validation facilities provided by this library are geared towards validating a document while it is being edited. The API provides functions that are meaningful in that kind of context. However, there's nothing that prevents using this library to perform a "batch" validation of the kind you'd do by running a validator at the command line (for instance).

Using

The library is written in TypeScript but can be used by JavaScript code. The wed editor mentioned above was still implemented in ES5 when it started using salve-dom, though it has since been converted to TypeScript.

However, if you are using JavaScript you still need to pay attention to the TypeScript annotations and pay special attention to:

1. What is public and what is not public in the API. Anything that the TypeScript type annotations do not expose to TypeScript code is private. Relying on private parts of the code is an API violation.

2. What is marked readonly. Writing a variable marked readonly is an API violation.

3. Objects marked readonly by their name. For instance, TypeScript has a ReadonlyArray generic which can be used to mark an array as being immutable. The thing is though that the underlying JavaScript array is still mutable. Mutating objects that are meant to be immutable is an API violation.

The nitty-gritty details are in the typedoc comments that addorn the source code. You should rely on these. Note that typedoc is still at an early stage of development compared to tools like JSDoc. Some of the documentation tags are not recognized by typedoc but are nonetheless used to record useful information.

The big picture:

• You create a Validator object that will validate a DOM tree according to your Relax NG schema. Let's say your new validator is validator.

• You install event listeners to listen for events on validator.events.

• You start it with validator.start().

• You can stop it before it is done with validator.stop().

• The current state of validation is available with validator.getWorkingState() and validator.errors.

• Editors that do validation during editing will want to use methods like possibleAt, possibleWhere, speculativelyValidate, etc.

• The Validator class calls a method named _runDocumentValidation when it is done with a document. The default implementation of _runDocumentValidation is a no-op. A class extending Validator could override this protected method to perform validation that cannot be modeled in Relax NG. The implementation may call the protected method _processError to add errors to the validation results.

Caveats

1. The schema must have been processed with salve in order to be used by the Validator class.

2. Errors report the DOM Node where the errors occur. This works fine if you are showing the actual document being validated to the user: you can just highlight the erroring node. (This is what wed does.)

   In a "batch" mode of operation, however, where we typically want to report errors with reference to line/column numbers, there is a problem because DOM nodes do not record line and column information. The issue here is largely a consequence of validating DOM trees. You could fairly easily compute a location from the erroring node and the serialization of the DOM tree (from outerHTML). However, the serialization of the DOM tree is not necessarily the same sequence of bytes as the string that was originally parsed into a tree. So this may not be particularly useful.

   Trying to match the erroring node to the source in a robust way (i.e. which would work with any XML thrown at it) would probably be as complicated as writing an XML parser. If you must report errors with line and column numbers that refer to the source as provided by the user, your best bet is to use salve in conjuction with a library like sax. Salve's own source code and test suite have examples of such usage.

Browser Requirements

You should check the requirements provided in salve's own README if you want details regarding what the JavaScript environment must support. Since version 4 salve-dom is compiled to generate es6 code, and does not include any polyfills for old browsers.

salve-dom passes its test suite on Chrome, Firefox, Edge, Opera and Safari form Sierra onwards. It has been tested on "relatively recent versions".

Performance Note on FF

The performance on FF 52 is bad. Most of the browsers mentioned above are in the same ballpark performance-wise. However, when the test suite is run on FF 52, it takes easily 3 to 4 times as long as on other browsers. There seem to be a noticable performance degradation bump at FF 49, getting progressively worse with newer versions.

We unfortunately do not have time to track down what could be causing this performance drop on FF.

Loading

Two general options:

1. Load it as a UMD module. You must configure your loader to load the salve-dom*.js bundles in the root of the installed salve-dom package. This probably your best bet if you want to use the library "as-is". The salve module is required by these bundles, so it must be available through whatever module loader you use.

2. Load it as a collection of modules. You'll find the modules in the lib subdirectory of the installed package. These modules are in the CommonJS format. This allows creating custom bundles (e.g. combine salve-dom and salve in a single bundle).

Building

You need at least Node 10 to build it.

Acknowledgments

salve-dom is tested using BrowserStack. BrowserStack provides this service for free under their program for supporting open-source software.

salve-dom is also tested using Sauce Labs. Sauce Labs provides this service for free under their Open Sauce program.

Credits

salve-dom is designed and developed by Louis-Dominique Dubeau, Director of Software Development for the Buddhist Translators Workbench project, Mangalam Research Center for Buddhist Languages.

This software has been made possible in part by a Level I Digital Humanities Start-up Grant and a Level II Digital Humanities Start-up Grant from the National Endowment for the Humanities (grant numbers HD-51383-11 and HD-51772-13). Any views, findings, conclusions, or recommendations expressed in this software do not necessarily represent those of the National Endowment for the Humanities.