In other news, the conversion to Racket is almost finished. I’m doing a test run on the actual data, and there’s plenty of bugs to fix/docs to write/&c., but I think I’m on the verge of throwing out the ruby!
Hah, thought the test run would be fast… b/c I thought ruby processing in 100s was slow. It’s been an hour now, there are 11k files to process… I’ve been slowly improving the generated code, trying to make it more efficient, while I wait. (Basically it’s Slack archives -> reasonable archive format -> pollen code -> HTML, and the slow bit is the last step.)
There are a couple of Pollen optimizations that I wrote in https://github.com/mbutterick/pollen-users/issues/43. Not sure if it will help.
Thanks, I’ll take a look. Using beeswax and generating more things ahead of time has been handy (e.g., generating the nav, since i know it up front, rather than each page re-computing it)
update: beeswax and some performance hacks seem to be making a very tangible difference. I’m about as far in 5 minutes as it took ~1h to get last time…
Is anyone going to be bothered if I completely skip the mailing list channel? I don’t see a good reason to archive mailing list announcements
I see no value too
ah, yes, that also looks… not worth spending time on. (I’m not removing anything from the archive data, for the record, just trimming down the list of output channels)
Have you been downloading recent data and merging it with the main archive, by the way? Cause we ran out of the free trial for a while now.
oh yes, samth has been creating archives. we haven’t set a regular cadence yet, but we have everything up to a few days ago IIRC. The current site (broken as it is by the length of some pages) is technically up to date
nice! and thank you for doing all of this :slightly_smiling_face:
The racket build will be easier to work with from day 1 because i broke up each channel’s messages by day :slightly_smiling_face: