[IDEA] It's requested to keep 1 .tid file per tiddler, if field-names violate v5.1.x rules about tiddlywiki5

[pmario]

@Jermolene -- I had to change the OP text a little bit because I did confuse "convert to JSON" with "convert to .txt and txt.meta" rules

[Jermolene]

Hi @pmario the discussion at #7799 is about the filesystemadaptor saving .tid files. It is not affected by the use of script tags with the new JSON store area.

[pmario]

I know. But the main reason for changing .tid files to .JSON files was the problem of the "save and reload" round-trip using HTML attributes to store field-names. Now since we do not have the old-store mechanism anymore, there is no need to ever transit from .tid file to .json except if there are multi-line values in fields.

My request is to remove the automatic transition from .tid to .json, just because the filed-name contains invalid characters.

[Jermolene]

Hi @pmario again, this is nothing to do with round-tripping fields to HTML attributes. There is entirely separate code to check for tiddlers that cannot be saved in a .tid file:

TiddlyWiki5/core/modules/utils/filesystem.js

Lines 225 to 232 in 4ebaba8

	var hasUnsafeFields = false;
	$tw.utils.each(tiddler.getFieldStrings(),function(value,fieldName) {
	if(fieldName !== "text") {
	hasUnsafeFields = hasUnsafeFields || /[\x00-\x1F]/mg.test(value);
	hasUnsafeFields = hasUnsafeFields || ($tw.utils.trim(value) !== value);
	}
	hasUnsafeFields = hasUnsafeFields || /:|#/mg.test(fieldName);
	});

As you can see, the conditions that trigger not being able to use a .tid file are:

• Any field name includes a control character from 0x00 to 0x1f
• Any field name starts or ends with whitespace
• Any field name includes the characters : or #