Details
Task
Open
Minor
Description
If one includes characters from the Unicode Supplemental Multilingual Plane (code points U+10000 upwards) in a story file, if one then asks for an HTML report from the test run the characters will not be HTML-escaped correctly.
For example, given a story file with the following scenario:
------------
Scenario: Some scenario
Given some situation
When I do something
Then the result is 𐐆
------------
(The "dagger"-type character is actually code point U+10406 - see http://en.wikibooks.org/wiki/Unicode/Character_reference/10000-10FFF)
The resulting HTML report will have the "dagger" character escaped as �� - which represent surrogate-pair code points (used in UTF-16 only) and so is rendered as gibberish in HTML. The escape should be 𐐆
NOTE: This is NOT a bug in JBehave per se - the bug is in the StringEscapeUtils class of commons-lang. A related bug has already been raised (and fixed) in commons-lang: https://issues.apache.org/jira/browse/LANG-617. Although the commons-lang bug report relates to XML escaping rather than HTML escaping, it seems likely that the fix will cover both. Unfortunately, the fix is in commons-lang 3.0...
Activity
We use commons-lang to escape both html and xml. So, we'd have to wait for the fix in 3.0 to be released, unless you feel like digging into the common-lang trunk and cherry-pick the patch that fixed your unicode problem. We'd be happy to apply it as stop-gap until commons-lang is released.
Mauro Talevi
added a comment - We use commons-lang to escape both html and xml. So, we'd have to wait for the fix in 3.0 to be released, unless you feel like digging into the common-lang trunk and cherry-pick the patch that fixed your unicode problem. We'd be happy to apply it as stop-gap until commons-lang is released. The bug doesn't actually impede the functionality of any tests - it's purely cosmetic on the test results. Therefore, to me the importance of the bug doesn't outweigh the risk of introducing more bugs by coming up with a stopgap fix (which someone then has to remember to remove once commons-lang 3.x is introduced as a dependency in JBehave). But I did think that there was value in getting the issue logged so that anyone else encountering it had a reference point.
So I'm happy to wait for commons-lang 3.x and for the bug I've raised to be archived off in whatever way you think is best.
Alistair Dutton
added a comment - The bug doesn't actually impede the functionality of any tests - it's purely cosmetic on the test results. Therefore, to me the importance of the bug doesn't outweigh the risk of introducing more bugs by coming up with a stopgap fix (which someone then has to remember to remove once commons-lang 3.x is introduced as a dependency in JBehave). But I did think that there was value in getting the issue logged so that anyone else encountering it had a reference point.
So I'm happy to wait for commons-lang 3.x and for the bug I've raised to be archived off in whatever way you think is best. Changed to task as reminder to upgrade to commons-lang 3.0 when released.
Mauro Talevi
added a comment - Changed to task as reminder to upgrade to commons-lang 3.0 when released. Mauro Talevi
made changes -
| Field | Original Value | New Value |
|---|---|---|
| Issue Type | Bug [ 1 ] | Task [ 3 ] |
Mauro Talevi
made changes -
| Summary | Characters from the Unicode Supplemental Multilingual Plane included in story definitions get rendered incorrectly in HTML | Upgrade to commons-lang 3.0 to correctly escape characters from the Unicode Supplemental Multilingual Plane |
| Fix Version/s | 3.2 [ 16757 ] | |
| Component/s | Core [ 11086 ] |
Mauro Talevi
made changes -
| Fix Version/s | 3.x [ 16979 ] | |
| Fix Version/s | 3.2 [ 16757 ] |
Noting that Lang3 is released - waiting on it to sync to the central maven repo etc etc.
Henri Yandell
added a comment - Noting that Lang3 is released - waiting on it to sync to the central maven repo etc etc. Mauro Talevi
made changes -
| Assignee | Mauro Talevi [ maurotalevi ] | |
| Fix Version/s | 3.5 [ 17393 ] | |
| Fix Version/s | 3.x [ 16979 ] |
Unfortunately, a commons-lang 3.x is not backward compatible with 2.x. We need to provide a plugglable and backward compatible way to escape chars using a given version of commons-lang.
Mauro Talevi
added a comment - Unfortunately, a commons-lang 3.x is not backward compatible with 2.x. We need to provide a plugglable and backward compatible way to escape chars using a given version of commons-lang. Mauro Talevi
made changes -
| Fix Version/s | 3.x [ 16979 ] | |
| Fix Version/s | 3.5 [ 17393 ] |
Ah, web interface to JIRA is rendering the character escapes as HTML rather than escaping them. Here is what I meant to say in the penultimate paragraph:
The resulting HTML report will have the "dagger" character escaped as �� - which represent surrogate-pair code points (used in UTF-16 only) and so is rendered as gibberish in HTML. The escape should be 𐐆