Skip to content

Commit

Permalink
Switch from html2text to soundasleep/html2text
Browse files Browse the repository at this point in the history
  • Loading branch information
@larssandergreen
larssandergreen committed Oct 1, 2023
1 parent a8d60e4 commit 5bfe3f8
Show file tree
Hide file tree
Showing 8 changed files with 78 additions and 101 deletions.
3 changes: 1 addition & 2 deletions CRM/Utils/String.php
View file
Original file line number Diff line number Diff line change
Expand Up @@ -445,8 +445,7 @@ public static function strtoboolstr($str) {
*/
public static function htmlToText($html) {
$token_html = preg_replace('!\{([a-z_.]+)\}!i', 'token:{$1}', $html);
$converter = new \Html2Text\Html2Text($token_html, ['do_links' => 'table', 'width' => 75]);
$token_text = $converter->getText();
$token_text = \Soundasleep\Html2Text::convert($token_html, ['ignore_errors' => TRUE]);
$text = preg_replace('!token\:\{([a-z_.]+)\}!i', '{$1}', $token_text);
return $text;
}
Expand Down
5 changes: 1 addition & 4 deletions composer.json
View file
Original file line number Diff line number Diff line change
Expand Up @@ -102,7 +102,7 @@
"symfony/polyfill-php80": "^1.0",
"symfony/polyfill-php81": "^1.0",
"symfony/polyfill-php82": "^1.0",
"html2text/html2text": "^4.3.1",
"soundasleep/html2text": "^2.1",
"psr/container": "~1.0 || ~2.0",
"ext-fileinfo": "*"
},
Expand Down Expand Up @@ -275,9 +275,6 @@
"Update gitignore to ensure that sites that manage via git don't miss out on the important db.json file": "https://patch-diff.githubusercontent.com/raw/adrienrn/php-mimetyper/pull/15.patch",
"Apply patch to fix php8.2 deprecation notice on dynamic property $filename": "https://patch-diff.githubusercontent.com/raw/adrienrn/php-mimetyper/pull/17.patch"
},
"html2text/html2text": {
"Fix deprecation warning in php8.1 on html_entity_decode": "https://raw.githubusercontent.com/civicrm/civicrm-core/e758d20e9f613ca6c4cf652c23d2cd7e5d3af3ce/tools/scripts/composer/html2text_html2_text_php81_deprecation.patch"
},
"pear/db": {
"Apply patch to ensure that MySQLI reporting remains the same in php8.1": "https://patch-diff.githubusercontent.com/raw/pear/DB/pull/13.patch",
"Apply patch to fix deprecations in php8.2": "https://patch-diff.githubusercontent.com/raw/pear/DB/pull/14.patch",
Expand Down
98 changes: 56 additions & 42 deletions composer.lock
View file

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

13 changes: 4 additions & 9 deletions ext/flexmailer/tests/phpunit/Civi/FlexMailer/ClickTracker/TextClickTrackerTest.php
View file
Original file line number Diff line number Diff line change
Expand Up @@ -45,19 +45,14 @@ public function getHrefExamples() {
'<p><a href=\'tracking(https://sub.example.com/foo.php?whiz=%2Fbang%2F&pie[fruit]=apple)\' rel=\'nofollow\'>Foo</a></p>',
];
$exs[] = [
// Messy looking URL, designed to trip-up quote handling
// Messy looking URL, designed to trip-up quote handling, no tracking as no http
'<p><a href="javascript:alert(\'Cheese\')">Foo</a></p>',
'<p><a href="tracking(javascript:alert(\'Cheese\'))" rel=\'nofollow\'>Foo</a></p>',
'<p><a href="javascript:alert(\'Cheese\')" rel=\'nofollow\'>Foo</a></p>',
];
$exs[] = [
// Messy looking URL, designed to trip-up quote handling
// Messy looking URL, designed to trip-up quote handling, no tracking as no http
'<p><a href=\'javascript:alert("Cheese")\'>Foo</a></p>',
'<p><a href=\'tracking(javascript:alert("Cheese"))\' rel=\'nofollow\'>Foo</a></p>',
];
$exs[] = [
// Messy looking URL, funny whitespace
'<p><a href="http://example.com/' . "\n" . 'weird">Foo</a></p>',
'<p><a href="tracking(http://example.com/' . "\n" . 'weird)" rel=\'nofollow\'>Foo</a></p>',
'<p><a href=\'javascript:alert("Cheese")\' rel=\'nofollow\'>Foo</a></p>',
];
$exs[] = [
// Messy looking URL, funny whitespace
Expand Down
30 changes: 8 additions & 22 deletions tests/phpunit/CRM/Mailing/BaseMailingSystemTest.php
View file
Original file line number Diff line number Diff line change
Expand Up @@ -163,14 +163,7 @@ public function testHtmlWithOpenTracking(): void {
// Default header
"Sample Header for TEXT formatted content.\n" .
// body_html, filtered
"You can go to Google \\[1\\] or opt out \\[2\\]\\.\n" .
"\n" .
"\n" .
"Links:\n" .
"------\n" .
"\\[1\\] http://example.net/first\\?cs=[0-9a-f_]+\n" .
"\\[2\\] http.*civicrm/mailing/optout.*\n" .
"\n" .
"You can go to \\[Google\\]\\(http://example.net/first\?cs=[0-9a-f_]+\\) or \\[opt out\\]\\(http.*civicrm/mailing/optout.*\\)\\.\n" .
// Default footer
"to unsubscribe: http.*civicrm/mailing/optout" .
";",
Expand Down Expand Up @@ -217,14 +210,7 @@ public function testHtmlWithOpenAndUrlTracking(): void {
$this->assertMatchesRegularExpression(
";" .
// body_html, filtered
"You can go to Google \\[1\\] or opt out \\[2\\]\\.\n" .
"\n" .
"\n" .
"Links:\n" .
"------\n" .
"\\[1\\] http.*(extern/url.php|civicrm/mailing/url)(\?|&)u=\d+&qid=\d+\n" .
"\\[2\\] http.*civicrm/mailing/optout.*\n" .
"\n" .
"You can go to \\[Google\\]\\(http.*(extern/url.php|civicrm/mailing/url)(\?|&)u=\d+&qid=\d+\\) or \\[opt out\\]\\(http.*civicrm/mailing/optout.*\\)\\.\n" .
// Default footer
"to unsubscribe: http.*civicrm/mailing/optout" .
";",
Expand All @@ -249,20 +235,20 @@ public function urlTrackingExamples() {
$cases[0] = [
'<p><a href="http://example.net/">Foo</a></p>',
';<p><a href="http://example\.net/">Foo</a></p>;',
';\\[1\\] http://example\.net/;',
';\\(http://example\.net/\\);',
['url_tracking' => 0],
];
$cases[1] = [
'<p><a href="http://example.net/?id={contact.contact_id}">Foo</a></p>',
// FIXME: Legacy tracker adds extra quote after URL
';<p><a href="http://example\.net/\?id=\d+""?>Foo</a></p>;',
';\\[1\\] http://example\.net/\?id=\d+;',
';\\(http://example\.net/\?id=\d+\\);',
['url_tracking' => 0],
];
$cases[2] = [
'<p><a href="{action.optOutUrl}">Foo</a></p>',
';<p><a href="http.*civicrm/mailing/optout.*">Foo</a></p>;',
';\\[1\\] http.*civicrm/mailing/optout.*;',
';\\(http.*civicrm/mailing/optout.*\\);',
['url_tracking' => 0],
];
$cases[3] = [
Expand All @@ -284,21 +270,21 @@ public function urlTrackingExamples() {
$cases[5] = [
'<p><a href="http://example.net/">Foo</a></p>',
';<p><a href=[\'"].*(extern/url.php|civicrm/mailing/url)(\?|&amp\\;)u=\d+.*[\'"]>Foo</a></p>;',
';\\[1\\] .*(extern/url.php|civicrm/mailing/url)[\?&]u=\d+.*;',
';\\(.*(extern/url.php|civicrm/mailing/url)[\?&]u=\d+.*\\);',
['url_tracking' => 1],
];
$cases['url_trackin_enabled'] = [
'<p><a href="http://example.net/?id={contact.contact_id}">Foo</a></p>',
';<p><a href=[\'"].*(extern/url.php|civicrm/mailing/url)(\?|&amp\\;)u=\d+.*&amp\\;id=\d+.*[\'"]>Foo</a></p>;',
';\\[1\\] .*(extern/url.php|civicrm/mailing/url)[\?&]u=\d+.*&id=\d+.*;',
';\\(.*(extern/url.php|civicrm/mailing/url)[\?&]u=\d+.*&id=\d+.*\\);',
['url_tracking' => 1],
];

$cases[7] = [
// It would be redundant/slow to track the action URLs?
'<p><a href="{action.optOutUrl}">Foo</a></p>',
';<p><a href="http.*civicrm/mailing/optout.*">Foo</a></p>;',
';\\[1\\] http.*civicrm/mailing/optout.*;',
';\\(http.*civicrm/mailing/optout.*\\);',
['url_tracking' => 1],
];
$cases[8] = [
Expand Down
17 changes: 4 additions & 13 deletions tests/phpunit/CRM/Utils/HtmlToTextTest.php
View file
Original file line number Diff line number Diff line change
Expand Up @@ -25,27 +25,18 @@ public function htmlToTextExamples() {

$cases[] = [
"\n<p>\n" .
"This is a paragraph with <b>Bold</b> and <i>italics</i>\n" .
"This is a paragraph with <b>Bold</b> and <i>italics</i>.\n" .
"Also some <a href=\"http://www.example.com\">hrefs</a> and a\n" .
"few <mailto:\"[email protected]\">mailto</mailto> tags.\n" .
"This is also a really long long line\n" .
"\n",
"This is a paragraph with BOLD and _italics_ Also some hrefs [1] and a few\n" .
"mailto tags. This is also a really long long line\n" .
"\n" .
"Links:\n" .
"------\n" .
"[1] http://www.example.com\n" .
"",
"This is a paragraph with Bold and italics. Also some [hrefs](http://www.example.com)" .
" and a few mailto tags. This is also a really long long line",
];

$cases[] = [
"<p>\nA <a href=\"{action.do_something}\">token</a>\nis not treated as a relative URL",
"A token [1] is not treated as a relative URL\n" .
"\n" .
"Links:\n" .
"------\n" .
"[1] {action.do_something}\n",
"A [token]({action.do_something}) is not treated as a relative URL",
];

return $cases;
Expand Down
11 changes: 3 additions & 8 deletions tests/phpunit/CRM/Utils/TokenConsistencyTest.php
View file
Original file line number Diff line number Diff line change
Expand Up @@ -1122,7 +1122,7 @@ public function testEscaping(): void {
]);
$context['eventId'] = $this->eventCreateUnpaid([
'title' => 'The Webinar',
'description' => '<p>Some online webinar thingy.</p> <p>Attendees will need to install the <a href="http://telefoo.example.com">TeleFoo</a> app.</p>',
'description' => '<p>Some online webinar thingy.</p><p>Attendees will need to install the <a href="http://telefoo.example.com">TeleFoo</a> app.</p>',
])['id'];

$messages = $expected = [];
Expand All @@ -1138,15 +1138,10 @@ public function testEscaping(): void {
$messages['event_text'] = 'You signed up for this event: {event.title}: {event.description}';
$expected['event_text'] = 'You signed up for this event: The Webinar: Some online webinar thingy.
Attendees will need to install the TeleFoo [1] app.
Links:
------
[1] http://telefoo.example.com';
Attendees will need to install the [TeleFoo](http://telefoo.example.com) app.';

$messages['event_html'] = '<p>You signed up for this event:</p> <h3>{event.title}</h3> {event.description}';
$expected['event_html'] = '<p>You signed up for this event:</p> <h3>The Webinar</h3> <p>Some online webinar thingy.</p> <p>Attendees will need to install the <a href="http://telefoo.example.com">TeleFoo</a> app.</p>';
$expected['event_html'] = '<p>You signed up for this event:</p> <h3>The Webinar</h3> <p>Some online webinar thingy.</p><p>Attendees will need to install the <a href="http://telefoo.example.com">TeleFoo</a> app.</p>';

$rendered = CRM_Core_TokenSmarty::render($messages, $context);

Expand Down
2 changes: 1 addition & 1 deletion tests/phpunit/Civi/Token/TokenProcessorTest.php
View file
Original file line number Diff line number Diff line change
Expand Up @@ -524,7 +524,7 @@ public function getFilterExamples(): array {
$testCases['TextMessages with HtmlData'] = [
'text/plain',
[
'This is {my_rich_text.and_such}...' => 'This is TESTING & SUCH...',
'This is {my_rich_text.and_such}...' => 'This is testing & such...',
'This is {my_rich_text.and_such|lower}...' => 'This is testing & such...',
'This is {my_rich_text.and_such|upper}!' => 'This is TESTING & SUCH!',
],
Expand Down

0 comments on commit 5bfe3f8

Please sign in to comment.