[MDL-46256] count_words returns too few words when dealing with html tags - Moodle Tracker

XML

Word

Printable

Affected Branches:
MOODLE_27_STABLE, MOODLE_310_STABLE, MOODLE_311_STABLE, MOODLE_39_STABLE, MOODLE_400_STABLE
Fixed Branches:
MOODLE_310_STABLE, MOODLE_39_STABLE
Pull from Repository:
https://github.com/timhunt/moodle.git
Pull 3.9 Branch:
~~MDL-46256~~_39
Pull 3.9 Diff URL:
https://github.com/timhunt/moodle/compare/MOODLE_39_STABLE...MDL-46256_39
Pull 3.11 Branch:
~~MDL-46256~~_311
Pull 3.11 Diff URL:
https://github.com/timhunt/moodle/compare/MOODLE_311_STABLE...MDL-46256_311
Pull Master Branch:
~~MDL-46256~~
Pull Master Diff URL:
https://github.com/timhunt/moodle/compare/master...MDL-46256
Testing Instructions:

Hide

The existing and unit tests are sufficient.

Show
The existing and unit tests are sufficient.

The problem is strip_tags will join words together:

Some text
 
Some more text

How many words ? Moodle thinks it's 4.

Also count_letters includes the letters in HTML entities - eg   is 6 letters.

has been marked as being related by

MDL-71925 Inaccurate word count in Brickfield accessibility toolkit

is duplicated by

MDL-59989 count_words can miss things it is meant to count

Testing discovered

MDL-70796 Moodle word count does not match other software

Participants:: CiBoT, Damyon Wiese, Eloy Lafuente (stronk7), Jason Platts, Peter Spicer, Tim Hunt

Component watchers:: David Woloszyn, Huong Nguyen, Jake Dallimore, Meirza, Michael Hawkins, Raquel Ortega, Safat Shahin, Stevani Andolo

Estimated:

Not Specified

Remaining:

Logged:

45m