Conversation
| final += `export const SECOND: string = '${second}';\n`; | ||
| final += `// THIRD: codepoint >= 65536\n`; | ||
| final += `export const THIRD: string = '${third}';\n`; | ||
| require('fs').writeFileSync(path, final); |
There was a problem hiding this comment.
Do we expect the data file to ever change? If not I'd prefer to do a one time conversion and not include conversion scripts/data files in the repo
There was a problem hiding this comment.
It changes at least with every major release and the consortium updates the spec in a yearly fashion. In 2 months we should see unicode v11.0.
The changes address mostly new chars and updates of drafted specs, it seems the older a char the less likely is an update.
I am not a fan of the code generation script thingy either. It may introduce several problems due to inexpected changes in the spec - like the region_indicator from 9.0 to 10.0 where they even gave up the previous-next rule, which will not work correctly until explicitly handled in some corresponding code. Maybe we should tag a supported unicode version and stick to it for some time?
There was a problem hiding this comment.
Edit: The data generator could be parked in some dev module to avoid rewriting it once we decide to upgrade to a newer unicode version in future.
There was a problem hiding this comment.
xtermjs/node-grapheme-cluster-break?
|
Closing for now, needs a proper way of dealing with different unicode versions that has to come. See also #1470. |
WIP - dont merge.