All times in this document are recorded in UTC+2 (CEST).
Incident summary
An issue was identified where fill-in-the-gap questions did not correctly compare answers containing characters with accents when the Case sensitive option was disabled.
When Case sensitive was disabled, answers that differed only in letter casing should have been considered equal. However, for certain characters with accents, the comparison did not work correctly. For example, the correct answer "Álvaro" was not considered equal to the submitted answer "álvaro", even though they differ only in letter casing.
As a result, participants may have received an incorrect score and mark.
A hotfix for this issue was deployed on July 6, 2026, at 13:12 to prevent the issue from affecting newly calculated results.
An impact analysis identified 706 impacted results across 211 assignments. The affected schools have been informed.
Lead-up
On July 6, 2026, at 10:31, a user contacted Support after noticing that a fill-in-the-gap question incorrectly marked an answer as incorrect when the correct answer started with a capital accented letter, even though the same letter was submitted in lowercase and the Case sensitive option was disabled.
The reported behaviour was promptly investigated by the technical team, who reproduced the issue using the example provided by the user. During the investigation, it was confirmed that answers containing characters with accents were not always recognised as equal when they differed only in letter casing.
Further investigation showed that this behaviour was caused by the implementation used for case-insensitive comparison, which did not correctly handle characters with accents. Once the cause had been identified, a hotfix was developed, reviewed, and deployed to production.
Fault
Fill-in-the-gap questions support both case-sensitive and case-insensitive answer comparison through the Case sensitive option. When Case sensitive is disabled, answers that differ only in letter casing should be marked as equivalent.
The implementation responsible for this comparison did not correctly normalise characters with accents before comparing them. As a result, answers that should have been considered equivalent, such as "Álvaro" and "álvaro", were incorrectly marked as different.
As a result, participants could receive fewer points than intended for affected fill-in-the-gap questions.
Impact
The impact of this incident could have affected student scores and marks.
An impact analysis identified 706 impacted results across 211 assignments.
All affected schools have been informed.
Detection
The issue was reported by a user via a support ticket. The technical team investigated the reported behaviour, reproduced the issue, and identified the underlying cause.
Response
The technical team investigated the issue immediately after it was reported and identified the underlying cause. A hotfix was developed, reviewed, and deployed on July 6, 2026, at 13:12 to ensure the correct comparison behaviour for characters with accents.
Following the deployment, the technical team performed a complete impact analysis to determine which assignments and participant results had been affected. Once the analysis had been completed, the affected schools were informed.
Recovery
A hotfix was deployed on July 6, 2026, at 13:12 to ensure that answers containing characters with accents are correctly compared when the Case sensitive option is disabled.
Following the deployment, newly calculated results are no longer affected by this issue.
Timeline
6th of July, 2026
- 10:31 - The Support team received a ticket reporting that a fill-in-the-gap question incorrectly marked an answer containing accented characters as incorrect.
- 10:42 - The technical team started investigating the reported behaviour.
- 11:28 - The technical team identified the underlying cause.
- 11:46 - The technical team started developing a hotfix.
- 13:12 - The hotfix was deployed to the production environment.
- 13:24 - The technical team started the impact analysis.
8th of July, 2026
- 08:52 - The impact analysis was completed.
- 11:44 - The affected schools were informed.
Reflection
This issue was caused by the implementation used for case-insensitive comparison in fill-in-the-gap questions. While the comparison worked correctly for standard Latin characters, it did not correctly handle characters with accents when the Case sensitive option was disabled.
The issue was limited to a specific combination of question settings and answer content, making it difficult to detect during normal usage. It was ultimately identified because a user noticed an unexpected scoring result and reported it to Support.
Once the issue had been reported, the technical team reproduced the behaviour, identified the root cause, and deployed a hotfix within a few hours. The impact analysis was then completed to determine which historical results had been affected, after which all affected schools were informed.
This incident highlights the importance of including characters with accents in automated testing for answer comparison. Additional regression tests will be added to ensure that future changes correctly handle case-insensitive comparison for all supported characters.
Comments
0 comments
Article is closed for comments.