a REPL for Nim

cicak dev log

Read, eval, print, loop, and keep the state.

Rendered with nimib

A REPL for Nim, where each input is compiled to native code and the session keeps its state. These posts follow its development, newest first. The code is on GitHub.

Date	Post	Summary
1 Nov 2026	Cicak, the chirping gecko	Naming the REPL cicak, after the chirping house gecko. Where it lives, what it is called, the stories around it, and why it fits a REPL.
31 Oct 2026	How could a Nim REPL work?	Research: prior art for Nim REPLs, the cells-as-modules design, a library and service architecture, and a reactive notebook.
31 Oct 2026	A working PoC with Hot Code Reloading and Incremental Compilation	First proof of concept: new cell modules are loaded into a running process with hot code reloading, earlier globals keep their values, and incremental compilation makes adding a cell cheap.

## The dev-log index: lists the rendered posts in docs/blog, newest first.
## Render the posts first; nothing here needs editing when a post is added.
import nimib
import std/[os, algorithm, re, strutils, times]
import theme

nbInit
nb.title = projectName & " dev log"
nb.useTheme(
  titleHtml = projectName & " <em>dev log</em>",
  eyebrow = "a REPL for Nim",
  epigraph = "Read, eval, print, loop, and keep the state.",
  meta = "Rendered with nimib",
  summary = "Notes from building a REPL for Nim.",
)

type Post = object
  file, title, summary: string
  date: DateTime

proc unescape(s: string): string =
  s.multiReplace(("&quot;", "\""), ("&#x27;", "'"), ("&lt;", "<"), ("&gt;", ">"), ("&amp;", "&"))

proc escape(s: string): string =
  s.multiReplace(("&", "&amp;"), ("<", "&lt;"), (">", "&gt;"), ("\"", "&quot;"))

proc readPost(path: string): Post =
  ## Date and link from the dated file name, title from `<title>`, summary from
  ## the description meta tag the theme writes.
  let html = readFile(path)
  result.file = path.extractFilename
  result.date = parse(result.file[0 ..< 10], "yyyy-MM-dd")
  var m: array[1, string]
  if html.find(re"<title>([^<]*)</title>", m) >= 0: result.title = m[0].strip.unescape
  if html.find(re"""<meta name="description" content="([^"]*)">""", m) >= 0:
    result.summary = m[0].unescape

var posts: seq[Post]
for path in walkFiles("blog/*.html"):  # nimib runs from docs/ (see nimib.toml)
  if path.extractFilename.match(re"\d{4}-\d{2}-\d{2}-.*\.html"):
    posts.add readPost(path)
posts.sort(proc (a, b: Post): int = cmp(b.file, a.file))

nbText: """
A REPL for Nim, where each input is compiled to native code and the session keeps its
state. These posts follow its development, newest first. The code is on
[GitHub](""" & repoUrl & """).
"""

var items = "<div class=\"posts-wrap\"><table class=\"posts\">\n" &
  "<thead><tr><th>Date</th><th>Post</th><th>Summary</th></tr></thead>\n<tbody>\n"
for p in posts:
  items.add "<tr><td class=\"date\"><time datetime=\"" & p.date.format("yyyy-MM-dd") & "\">" &
    p.date.format("d MMM yyyy") & "</time></td>\n<td class=\"post\"><a href=\"blog/" & p.file & "\">" &
    escape(p.title) & "</a></td>\n<td class=\"summary\">" & escape(p.summary) & "</td></tr>\n"
items.add "</tbody></table></div>"
nbRawHtml: items

nbSave

Made with nimib. Blocks marked "nim · ran" were executed when this page was built. Source: cicak.