Set up the RegEx to be able to match characters between spoiler characters (for now, the spoiler characters are > and <, so any text between >these two symbols< will be matched by the RegEx. This is for https://linear.app/comm/issue/ENG-1981/set-up-regex-to-match-text-between-two-spoiler-characters, and subsequent parts of the task will be stacked on top of this diff.
Details
Details
Tested the RegEx on a random sentence with the two characters.
The result:
[NODEM] [ [NODEM] '>spoilers here<', [NODEM] 'spoilers here', [NODEM] index: 31, [NODEM] input: 'This is my testing phrase with >spoilers here<', [NODEM] groups: undefined [NODEM] ]
Also a link to an interactive playground with the regex demo: https://regex101.com/r/EC2mJt/1
Diff Detail
Diff Detail
- Repository
- rCOMM Comm
- Lint
Lint Not Applicable - Unit
Tests Not Applicable
Event Timeline
Comment Actions
Nice, looks good. I'm assuming you already spoke to atul/ashoat about using >SPOILER< to mark spoilers. I know Discord uses ||SPOILER|| to mark spoilers and some other places use >! SPOILER to mark spoilers
Comment Actions
Doesn't seem like spoiler is included in either the CommonMark or GitHub Flavored Markdown spec (which AFAIK are the two "big ones").
So I guess it's up to us to decide what we want to do... looks like reddit does >! SPOILER !< in addition to the ones @ginsu mentioned. Maybe we support all of them to reduce the learning curve for new users? Not sure
CC @ashoat to get his thoughts
Comment Actions
Tested the RegEx on a random sentence with the two characters.
Thoughts on adding some unit tests for this?
Comment Actions
Sure, happy to add unit tests, I can do that a bit after as a separate task once we decide on the format for spoilers - originally messaged on Comm, but this might be a better place so others can contribute on what they think.
Looks like reddit does >! SPOILER !< in addition to the ones @ginsu mentioned. Maybe we support all of them to reduce the learning curve for new users? Not sure
I think there are a few ways to do this, I think something that will be easy on mobile will make the most sense since a more complicated pattern could make it harder for mobile users. Both >! SPOILER !< and || SPOILER || seem fine to me based on Ginsu's suggestion - my opinion is the last one is a bit easier from a user perspective.
Comment Actions
Temporarily setting the spoiler RegEx to match text between || and ||, since using > causes problems with other Markdown. Waiting on final confirmation still on how we want to detect spoiler text.
Comment Actions
Sending back with a question. Agree with this decision on the format given overlap with quotation regex
Sure, happy to add unit tests, I can do that a bit after as a separate task once we decide on the format for spoilers
Can you create a Linear task for this and link it here?
| lib/shared/markdown.js | ||
|---|---|---|
| 253 ↗ | (On Diff #17462) | What's the question mark doing here? |
Comment Actions
Sure, here's the task: https://linear.app/comm/issue/ENG-2000/add-unit-tests-to-test-the-regex-for-spoiler-text
Will update it with some more details
| lib/shared/markdown.js | ||
|---|---|---|
| 253 ↗ | (On Diff #17462) | I wanted a lazy evaluation of the text in order to allow for multiple spoilers within a text. As opposed to a greedy evaluation where it will go to the last instance of || in the sentence. So from my understanding (feel free to correct me here): Lazy: This is some ||spoiler text|| and this is ||more text to hide||. Greedy: This is some ||spoiler text|| and this is ||more text to hide||. |
Comment Actions
Add a caret to the spoiler regex to prevent simple-markdown errors and for
consistency with other regex expressions in the file