RLTEST

This package implements a large subset of the tcltest::test command interface for use within AOLserver / NaviServer - the results are returned as a string rather than being printed to stdout.

It also provides some test helpers: stubbing of commands and methods, assertions about HTML and JSON documents.

License

This package is licensed under the same terms as the Tcl core.

Assert checks

The html_assert proc supports a very flexible set of tests with the -asserts option, and a simplified but more limited set with -checks

checks format

The checks option is a list of xpath expectation pairs. For each pair, the xpath is run against the document and the text value of every result must much match the expectation. Every xpath must also match some element to succeed, unless the expectation is a non-match.

The text value of an attribute is that attribute's value; while the text value of a element is the concatenation of all text nodes underneath that node.

The expectation is normally an ordinary string glob (that is, matched with string match). For particular use cases, several other syntaxes are supported:

• * - existence
  The matched path must exist. It can be blank.

• ! - nonexistence
  The matched path must not exist. A blank element still exists, and so will fail this check.

• ( words ) - word check
  The words string is broken at whitespace, and each resulting word is checked to exist in the matched value as an indivudual word. This is useful for checking for one or more html class values that may be in arbitrary order. All words must exist.

• !( words ) - word negative check.
  As word check, but all of the words must not exist. Any one of the words existing will cause the check to fail. An element or attribute that does not exist at all will implicitly match a negative check.

• + - nonblank
  The matched value must be nonblank.

• !+ - blank
  The matched value must be blank if it exists. To check that something exists and is blank, use multiple checks.

• &.... - subst value
  The expectation value is expanded with subst, with the result checked for a match as an ordinary glob. This is useful when you need to pull in a variable to the check without being forced to list-ify the entire check list.

Examples

For the example html result

<div id="d1">
    <img alt="title" src="/foo.gif">
    <img alt="title" title="title" src="/foo.gif">
	<span class="xs-3 sm-6">Caption</span>
	<span id="s2"></span>
</div>

• //div * - matches (a div exists)
• //div/@id * - matches (a div with an id attribute exists)
• //div/@class * - fails (no div with a class attribute exists)
• //img/@alt title - matches (every img tag has "title" as its alt attribute)
• //img/@title title - matches (every img tag with a title attribute has "title" as that attribute). Note that this does not verify that every img tag has a title attribute.
• //span + - fails (a blank span is present)
• //span/@class {(sm-6 xs-3)} - matches (all of the words are in the given attribute, although they are not in the given order)
• //span/@class "*s*" - matches, since the class attribute contains an "s"
• //span/@class {(s)}" - doesn't match, since the "s" in the attribute is not a word by itself.

There is no support for matching strings that would otherwise match the magic syntax. If that is needed, you will need to use the more general -asserts mechanism.