Difference between revisions of "Basic Functions of a Wiki"
(→getText) |
(→useCurl) |
||
| Line 114: | Line 114: | ||
} | } | ||
</code> | </code> | ||
| + | |||
| + | The $url here is the selenium server url. curl_exc is calling that url. Next to check is what selenium server does with getText as the verb in the $url. | ||
=== assertElementContainsText === | === assertElementContainsText === | ||
Revision as of 15:07, 19 March 2007
each test independent of the others? can we test preview or save if edit isn't working?
- edit
- logged in & not logged in, click the "edit" link at the top & "type" the "action=edit" url & click section edit & type section edit url
- verify the textarea exists
- verify the textarea contents are correct
- verify that typing works
- logged in & not logged in, click the "edit" link at the top & "type" the "action=edit" url & click section edit & type section edit url
- save
- in edit mode, click the "Save page" button
- verify that the page saves
- we're now NOT in edit mode
- the contents of the page are what we expect
- preview
- in edit mode, click the "Show preview" button
- the page shows both the contents of the page in html AND the textarea we're editing
Contents
Next Actions
- need to carefully think through edit tests to make sure they cannot pass when they shouldn't - all 8 tests now pass!
- CFLN issue very interesting - I'm writing this
$sectionOne="==first section==" .chr(13).chr(10). "test text in page $testPage $timestamp";and then getting this from the test runner:
Failed asserting that <string:==first section== test text in page AboutUsTest.com 1173799512> contains "==first section== test text in page AboutUsTest.com 1173799512".
Done
- created files basic_wiki_functions_acceptance_test.sh & BasicWikiFunctionsTest.php
- testEditLinkWholePageIP() & testEditLinkWholePageLogin() both call editLinkWholePage() which does the actual work - the only difference is the login one logs in
- testEditLinkSectionIP() & testEditLinkSectionLogin() both call editLinkSection() - same as above
- basic structure of the file is complete now with 8 tests in two categories, logged in & IP; calling 4 functions in two categories, whole page or section; calling 2 functions in two categories, edit link or type edit URL; call DoTest(), so it's something like this
- doTest
- editLink
- editLinkWholePage
- IP
- logged in
- editLinkSection
- IP
- logged in
- editLinkWholePage
- editURL
- editURLWholePage
- IP
- logged in
- editURLSection
- IP
- logged in
- editURLWholePage
- editLink
- doTest
Notes
- carriage return line feed issue
- linux uses LF only, mac uses CR only, win uses CRLF [1]
- What do I do about this?
- What happens if we use CRLF ( chr(13).chr(10) )?
- whole page text, not just one section
- look for more duplication in code
getText
public function getText($locator)
{
return $this->selenium->getText($locator);
}
public function getText($locator)
{
return $this->getString("getText", array($locator));
}
private function getString($verb, $args = array())
{
try {
$result = $this->doCommand($verb, $args);
} catch (Testing_Selenium_Exception $e) {
return $e;
}
return substr($result, 3);
}
private function doCommand($verb, $args = array())
{
$url = sprintf('http://%s:%s/selenium-server/driver/?cmd=%s', $this->host, $this->port, urlencode($verb));
for ($i = 0; $i sessionId)) {
$url .= sprintf('&%s=%s', 'sessionId', $this->sessionId);
}
if ($this->driver == 'curl') {
$response = $this->useCurl($verb, $args, $url);
} else {
$response = $this->useNative($verb, $args, $url);
}
if (!preg_match('/^OK/', $response)) {
throw new Testing_Selenium_Exception('The Response of the Selenium RC is invalid: ' . $response);
}
return $response;
}
useCurl
private function useCurl($verb, $args, $url)
{
if (!function_exists('curl_init')) {
throw new Testing_Selenium_Exception('cannot use curl exntensions. chosse or "native"');
}
if (!$ch = curl_init($url)) {
throw new Testing_Selenium_Exception('Unable to setup curl');
}
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
$result = curl_exec($ch);
if (($errno = curl_errno($ch)) != 0) {
echo("curl_errno: " . $errno);
throw new Testing_Selenium_Exception('Curl returned non-null errno ' . $errno . ':' . curl_error($ch));
}
curl_close($ch);
return $result;
}
The $url here is the selenium server url. curl_exc is calling that url. Next to check is what selenium server does with getText as the verb in the $url.
assertElementContainsText
public function assertElementContainsText($locator, $text)
{
$this->assertContains($text, $this->getText($locator));
}
/r/n, /r, & /n
Very strange.
- Mediawiki doesn't care if I put /r or /n or /r/n into the wiki text, which is good, of course.
- If I put in /r/n or /r or /n and then change /r/n to /n and then /r to " " and /n to " " in my expected text, my tests pass as long as I don't have /r/r
- if I do have /r/r then that gets replaced in my expected test with two spaces, while in the getText output, there's only one space - I really need to look at docommand() to see why/how these end of line characters are being replaced/stripped
