#memes #memestr

#RUNSTR #Strength

Got a little further on it, using a declarative map may help simplify the handling of the tags, since i can just drill down the list based on any arbitrary number of namespaces. For now though, I'm just storing the tags comma separated and using Internal Field Seperators when reading it back. I should be able to bring back in the tag requirements as well via increment counters if the key equals the required namespace(s). View quoted note →

Some more progress on #archivestr 's rebuild. IFS reads can use delimiters to split the line into multiple variables :drgn_blep: I do need to add some logic to rewrite them if the namespace does not exist, so that there's consistency, but this stops me having to 'find' the delimiter on my own. #grownostr #programming #artstr

I think my current workaround to the 't' tag problem is to use 'tag' instead for the unmolested tags, but continue to match the content body hashtags with matching 't' values, which can be ignored by tag parsing tools. View quoted note →

Starting on the rebuild of #archivestr , got a new outline on tag processing commented into the file. Ripped everything out from beyond line 60 in the old version in the update branch. Think It'll work. Main problem will still be handling special characters in the key pairs unless there's another way i can feed them into nak.

Coming back to running #Ollama on the new GPU again, I had already forgotten how fast 8b models can run on the latest AMD cards for text inference & generation.

#RUNSTR #Walking

Had to reinstall and redo permissions for #runstr , but i got it able pull step counts out of Gadgetbridge.