From 773621f61050b827d58103cca124b4592223a368 Mon Sep 17 00:00:00 2001 From: sorrel Date: Wed, 13 Dec 2023 22:14:10 -0500 Subject: [PATCH] update the readme so it's not lying --- README | 97 ++++++++++++++++++++++++++--------------------------- publish.rkt | 9 +---- 2 files changed, 48 insertions(+), 58 deletions(-) diff --git a/README b/README index 25812f1..7e566e2 100644 --- a/README +++ b/README @@ -1,6 +1,6 @@ -------------- - | λ.sorrel.dev |\ + | λ sorrel.dev |\ \--------------\ \ \______________\| @@ -21,14 +21,18 @@ so |----------------| ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯ it's a racket web server that uses xml library to generate some valid html. -there's a dirty little dsl in there so that i can type some plain texty stuff and -have my plain texty stuff get turned into sexprs and then into xexprs and then written -as new-fangled html. typical lispy stuff. -``` -(require web-server ; like so much of this module - xml) -``` -that's all compiled and hosted in my friend maren's bedroom. it's all very cute. +i write my little posts in xml-like s-expressions. (not for everyone! but i like the +expressivity!) someday, i might turn it into a proper dsl with support for htmx's +~funky~ syntax and a little more distance from it's host language, but not this day. + +i make generous use of `racket/web-server` with a dispatcher for `hx/` prefixed content +(snippets of html), and plain `/` (for rendering full pages.) handlers build the +content from the little sexpr .txt files (they're not valid racket on they're own, +they get read and evaluated as data.) it's all real small and everything, but i figured +why not save the generated html in a `static/` directory so it doesn't need to be built +on each request? so that's what happens. + +that's the server! it's compiled and hosted on bunk.computer. it's all very cute. _________________ |-----------------| @@ -46,15 +50,16 @@ right now, the plan's for ~SETTLED THOUGHTS~ it's like a portfolio. links to stuff. deployed stuff. pictures of stuff. a - home for everything that's done enough to need a home. + home for everything that's done enough to need a home. or projects with + "minimally kickable tires" - ~THOUGHTS ABOUT WHO EVEN AM I~ - yr typical about page, y'know with some biographical information, some - musings about life's meaning. some links to things i like or do as a - person with other people in the world. - - ~some kinda tag system for things - ~some kinda apology for even serving you some javascript + ~THOUGHTS ABOUT THOUGHTS?~ + there's some "root" resources. yr typical about page, y'know with some + biographical information, or whatever. "errata" like: + ~a page about the site itself. + ~maybe a link page with some links to things i like or do as a person with other people + in the world + ~a tag index and indices for tags the structure of things is such that: there's no templating per se. if the request has the path prefix "/hx/" you just get little snippets of html. the htmx is there to @@ -70,43 +75,35 @@ or footer or any of that fanciness. | HOW TO POST | |-------------| ¯¯¯¯¯¯¯¯¯¯¯¯¯ -remember up at the top of the README i said there was a DSL for this thing? i meant a -domain specific language. the domain is hella specific. like just my particular web -page specific. there's little grammars for each little part of it. i'll talk about -those in a moment. +let's talk about the posting. so you write up a little bit of s-expressified xml-like +plain text works of astounding eloquence; how do you get the server to generate some +stuff for you? i wrote a little utility for that. it's called "publish" and it uses a +#lang package a friend made for making cli tools called +[cli](https://docs.racket-lang.org/cli/index.html) -for right now i'm going to talk about the posting. so you write up a little bit of -plain text according to the grammar for the type of post yr going to do. how do you get -the server to generate some stuff for you? that's where the second part of the process -comes in. there's a little racket utility for that. if it were an ~unsettled thought~ -for example it might look something like: +the tool is called "publish" and it needs some flags to build a thing right. ``` -$ mkpost -s source/source-file.txt -t public_html/blog/target-file.html +$ publish -i in-progress/ex.txt \ # -i is for the input (i keep em in in-progress/") + -r unsettled \ # -r is for the resource "type" (see "WHAT'S ON THERE?") + -t ecology,art,computer \ # -t is for comma separated tags + -l "An Example Text" \ # -l is for head_l_ine (it means title) + -d "a short description-like a sentence long ideally" + # there's also an "-x" flag to do a "test run" before publishing. it'll generate + # stuff in publish-test/ including a "staging database" (it's csv files) ``` -there's a couple things this'll do. the most obvious is that it takes yr plain text -and it generates some html. you expected that. the other thing it does is it writes -the name of that file to a little lookup table with all it's tags. [i haven't -implemented this part yet] - -so that's simple enough. now for the - - __________________________ -|--------------------------| -| DOMAIN SPECIFIC LANGUAGE | -|--------------------------| - ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯ -the main pages are just handwritten xexprs. not worth spending time on something you'll -only write once. - -but for the posts: - -post = post type, tags - -post type = ( "blog" | "project" ) "\n" ; - -tags = { tag identifier, white space } "\n" ; -tag identifier = "#", { alphabetic character} +there's a couple things this'll do. +- first it'll take yr post and do some processing, add a footer with metadata and tags. +- it'll also update the "database" of csv files for the resource type and for all of + the associated tags. +- this generates new index files for the resource type, and the tags. +- if the post introduces a new tag, a new tag index gets made. +- the atom feed gets updated with the uri and the relevant info about the post. + (it's on the to-do list to embed the post in the feed, cause i like to read my rss + in cli, but i'm also a hypocrite that wants people to visit my page!) +- finally it moves the input file to an archive folder. + (in the future i'll make an update tool for posting updates to existing stuff!) +so that's simple enough. diff --git a/publish.rkt b/publish.rkt index 70cff52..eeee312 100644 --- a/publish.rkt +++ b/publish.rkt @@ -365,11 +365,4 @@ (displayln "publish was successful"))) -;;; (run publish) - -(run publish #("-i" "in-progress/beginning-latl.txt" - "-r" "unsettled" - "-t" "latl,conlang" - "-l" "Beginning LATL" - "-d" "beginning the process of thinking through an environment for conlanging and other language shenanigans" - "-x")) +(run publish)