At the moment Scintilla V5 is still regarded as unstable, IIUC it will be V5.1 thats stable "soon".

This issue is to discuss how we should approach the Scintilla V5 changes, the API changes, the Lexilla/Scintilla split, how to make the process scalable, massive single commits are simply not viable.

Then there needs to be a discussion about importing Scintilla/Lexilla it will be done in Geany after the changes to separate Scintilla and Lexilla.

To be blunt, the mechanics of getting the parts Geany uses from the Scintilla repos is the least of a Scintilla upgrade, be it git subprojects or a bash/Python script doesn't matter, its a minor part of the work needed for the upgrade, even now Scintilla and Lexilla are separate. So the simplest for everyone to understand is the best.

The main part of a Scintilla upgrade is the manual work for each lexer:

1. checking if any of the syntactic elements from the lexers have changed, 2. changing the mapping of those to Geany entities in `highlightingmappings.h` 3. adding them to `filetypes.xxx`, 4. checking for any lexer properties that need to be added or removed.

These are manual actions at the moment and make up most of the work creating and reviewing a Scintilla upgrade and are why the upgrades take so long.

IIUC there are some additional scripts or similar that come with Scintilla/Lexilla 5 that may make at least part of that work automatic, identifying changes at least. That needs to be explored to allow supporting the current number of languages to remain viable.

ATM its just too much work and nobody has the time.

Also potentially making Lexilla changes separate from Scintilla would help, IIUC Neil will be issuing them as separate releases in the future, and even better if each filetype could be done by itself. Then whoever is interested in a language can submit a smallish PR that would be easy to review and commit.

Huge commit bombs are simply not viable into the future, thoughts anyone?