#dev 2022-06-18

2022-06-18 UTC
#
@OwenYoungZh ↩️ 怎么感觉前几天给博客加的Webmention派上用场了? https://indieweb.org/Webmention , 理论上Webmention不是评论，而是外链。。 https://www.owenyoung.com/en/blog/indieweb/ (twitter.com/_/status/1537973323929227266)
cybi, gerben, gRegor, geoffo_, [KevinMarks] and mro joined the channel
#
@realWeZZard Then people go to use WebMention? (twitter.com/_/status/1538069192301113344)
gRegor, mro, tetov-irc, cybi, mro_, gxt, geoffo_ and jacky joined the channel
#
jacky at the point where caching is important :)
#
jacky with my site at least
#
jacky the whole "use SQL to find properties over every entry" got a bit unwieldy
#
jacky at about ~2k entries with _at least_ 6 properties, right now, it'd bulk load all of those files + parse as JSON (filtering out the unparsable) and then do some checks based on what the query wants
#
jacky trying to see where I can put caching layers to speed things up (obvs filesystem)
jacky joined the channel
#
[KevinMarks] That seems small enough that mysql would cache it
geoffo_, mro, angelo, Seirdy, jacky and tbbrown joined the channel
#
jacky using SQLite
#
jacky and with this implementation I have, it's up to me to do caching, it doesn't assume a lot
#
jacky namely https://sqlite.org/vtab.html#the_xbestindex_method and https://sqlite.org/vtab.html#the_xfilter_method that does the heavy lifting
geoffo_ joined the channel
#
[KevinMarks] I've found with SQL that you can add indexes, but sometimes you need to tell the query which one to use
#
Zegnat I have done a couple of times where I read data into in-memory SQLite, never even putting it in the filesystem, which was usually the fastest way to do on-demand lookups
gRegor, geoffo_ and Seirdy joined the channel
#
@kevinmarks ↩️ You're missing the point. By owning the domain http://brid.gy could redeploy elsewhere. Like I did with http://kevinmarks.com when Heroku got hacked and I switched it to Netlify. Cloud services are fungible and substitutable. (twitter.com/_/status/1538287361456472064)
tetov-irc and sp1ff joined the channel