Changelog

Find here all changes tracked while developing picossg.

(Maybe) Future Version

High Prio

• [ ] build the docs kit so when passing no data it guides one to create a complete docs site

Eventually

• [ ] only allow-list files to be copied, not just copy all
• [ ] ignore .DS_Store files (or turn it around and allow only certain files)
• [ ] add sizes to the _file and _output, maybe also the time it took to process
• [ ] docs: make the docs correct and as I like them to be
• [ ] docs: add to the docs how to use PicoSSG when running npm install in a project
• [ ] docs: document the second parameter to the preprocess and postprocess functions
• [ ] provide a playground to see play with the converter online
• [ ] plugins: a link checker would be nice
• [ ] plugins: render social-exported.json (e.g. tweets.json) to HTML, so it can be used in the site
• [ ] plugins: a {% toc %} for md pages would be cool to render the table of contents
• [ ] add the copied files to the stats, so they can be used in the templates, #niceToHave
• [ ] idea: make plugins including as easy as plugin.<plugin-name>.preprocess(files, config)
• [ ] create nicer API interface, maybe: preCreateProcessors(), preWalkFiles(), preProcess(), preOutput(), postOutput(), ...
• [ ] provide fullwidth option, see https://github.com/orgs/GitbookIO/discussions/450
• [ ] docskit: render multiple code blocks directly following one another as tabs (or introduce a {% tabs %} tag?)
• [ ] docskit: add "✨ AI Summary" tab at the top of each page, which shows the AI-generated summary of the page, we do AI-summary once, so people don't need to waste the same amount of CO2 again
• [ ] introducing "kits", ready-made page site kits, just throw md files in there and out comes e.g. a docs page
• [ ] test the error message for nunjucks parsing errors
• [ ] add a types.d.ts (maybe refactor to classes first, so extending is as easy as class MySSG extends PicoSSG)

v6.0.1 (in the works)

• [x] the package moved to a new name: @pic0/ssg (v6.0.0 is already available there)
• [ ] when I write a blog in picostitch without any front matter the errors are misleading

v6.0.0 (539 LOC)

• [x] MUCH better error messages when a variable was not found or some nunjucks error happens
  • [x] make extended/included files also report errors nicely
  • [x] make extended files across multiple extends also report errors nicely
  • [x] nice error message for a pure.njk file if there is an error in the extended file
• [x] leaving out the extension, like index.njk or index.md which would end up without extension after processing gets the default .html, like all other tools do, this does still allow for file names like style.css.njk to be processed
• [x] docs: link docs in the changelog to the docs site
• [x] the _file.numLinesFrontmatter tells how many lines long is the frontmatter block, including the --- lines

_config.js changes (and internals)

• [x] introduced _file.extensionsToProcess
• [x] refactor internals to only process the files data as given, use _file.extensionsToProcess and be more flexible this way
• [x] add optional config.templatePaths, to allow multiple template paths, needed for the docs-kit
• [x] introduce user function preCreateProcessors() that can be used to modify the config object
• [x] _file.hasFrontmatterBlock was removed, find out if a frontmatter exists by looking into _frontmatter if you need to know it
• [x] add the to-copy files to the files map, so the user can modify these
• [x] docs: describe the new flow in file-mapping

v5.0.0 (349 LOC)

• [x] BREAKING: rename configureNjk() to configure() and pass it all processors, so the user can configure as they like
• [x] BREAKING: rename "utils" to "plugins"
• [x] docs: multiple updates
• [x] docs: for throwOnUndefined default configuration for nunjucks and why
• [x] docs: for stats plugin addStatsProperty(files)
• [x] plugins: applying config.js files per directory, that applies to all files in that directory, and the child dirs this is implemented in src/plugins/configs.js and can be called like this in a preprocess() loadDataFromConfigs(files, config) I am NOT documenting this plugin yet, I feel like it is not going to last
• [x] plugins: provide {{ _stats.lastModified.* }} and {{ _stats.files }}

v4.0.0 (339 LOC)

• [x] add npm run build:site to build the picossg.dev website, might move somewhere else later
• [x] use _config.js instead of _nunjuck-custom/filters.js (a breaking change)
• [x] add config as second parameter to preprocess and postprocess functions, sometimes you need to know the config instead of assuming e.g. the output-path
• [x] improved the docs significantly
• [x] move picossg.dev site into the repo root in the dir picossg.dev so we can
  • [x] use/link changelog, package.json and other files directly from the source
• [x] rename dist to _output, just to be consistent in naming
• [x] ensure that the path output files are written to are child paths of the output dir, it might happen that one adds a _output object which accidentally writes to places it should not (actually just happened to me), this is basically a safeguard I feel worth having

v3.0.2 (332 LOC)

• [x] add some real docs
• [x] 🐛 fixed a bug, npm run build did not work, I had the wrong path to build-cli.js in there

v3.0.1 (332 LOC)

• [x] update the README to be useful and correct
• [x] provide example(s)

v3.0.0 (332 LOC)

• [x] rebuild tests to be at least somehow grouped/sorted
• [x] make the components directory be _includes, just like in nunjucks no need to invent names actually this is good as it is, the components have files that will be built and served, so its mixed purpose
• [x] rewrite the processing to
  • [x] provide the metadata conveniently
    • [x] merge the data below on the root level deterministically, so the data are convenient to use in the template but also available in their original form (in the _* keys)
    • [x] provide _file with all info about the original file, like filename, size, content, etc.
    • [x] provide _frontmatter with all frontmatter data
    • [x] provide _output with the processed data, like url, content, etc.
    • [x] provide _site with all data from the site, like url, title, description, author, etc.
• [x] allow preprocess() and postprocess() functions that receive ALL files incl. data and content like so: preprocess(files: Map<filename, {content: string, data: object}>) this allows processing any file before and after the main processing, it provides access to all files e.g. to find related ones and do searches
• [x] no need for picossg.*, remove it, can be done inside the preprocess() function
• [x] copy files first, and do NOT put them in the files map, so they are handled first, and do not land in the processing function (preprocess and postprocess)
• [x] allow post-processing, e.g. to minify pages
• [x] add absolute and relative filePath to _output, to provide where the file will be stored to
• [x] remove -i command line option, just make the includes root be the content dir, simple

v2.0.0 (281 LOC)

• [x] make npx run @pic0/ssg -c content -o out work in other projects
• [x] allow loading njk filters
• [x] provide filter |md and |mdinline
• [x] make a frontmatter block work
  • [x] pass meta data to the template
  • [x] make "extends/layout" key in frontmatter block work
  • [x] allow "collecting" all metadata e.g. through picossg.pages("/blog") or so and iterate over them to generate a blog list or alike
    • [x] and provide attributes for each page such as "url", ????
  • [x] pass the data into the layout template too, so e.g. a title of a page will be seen in the _base.njk template
• [x] add _config.js/ts to allow pre-processing

v1.0.0 (138 LOC)

• [x] serve and process njk files
• [x] serve and process md files
• [x] serve and process njk.md files
• [x] serve assets, favicon, img.svg
• [x] exclude files with _
• [x] make includes work
• [x] make asset loading from components work, actually just a convention, no code for it
• [x] generate code from e.g. JSON files can be done in _config.js
• [x] allow running JS server side code to generate content use _config.js