create a skills matrix

| Topic | Skill | Level | Supporting Evidence |

Nim memory management is tied to the type you use

You either use:

• an object. Which is on the stack and is either trivial or uses destructors and is suitable for embedded due to deterministic memory management
• a pointer object. Which is a raw pointer like C *. Managed directly via raw malloc/free or Nim malloc/free. Suitable for embedded
• a reference type. Which is managed by one of Nim GC or is an error if you use bc:none

via on HN.

You can use fly proxy command to setup a SSH proxy to a running Fly.io instance.

In the screenshot below, I used the connection to download the tohray database (a sqlite file) to my local machine for development.

tohray now has RSS feed

I added the RSS feed endpoint to this blog as well as the software.

Also used the newly created Convert to WebP MacOS QuickAction to convert the above screenshot to .webp!

Fontmatter

fontmatter is a new project to collect nice fonts, with sample of how they look, with a link where they can be downloaded, with licensing information.

The grid layout on the page itself was created using Claude 3.5 (It will soon become implied that every coding project has Claude's "hands").

Check it out and suggest your favourite Open Source font (or commercial if it's reasonably obtainable).

tohray now has edit and delete functionality.

The edit post and delete post links show up for every entry when you are logged in.

You can see the commit here on github.

I also want to thank Nilesh for sending a PR to add docker compose

This website gets excellent pagespeed scores.

100 on everything except 85 on a11y, where I was dinged for terminal.css's low contrast navigation links, and some links like "github" being too small for touch screen to click on :)

Scinim's Ecosystem overview is a great collection of Nim related libraries for deep-learning, matmul, nlp etc.

Adding /export endpoint to tohray

I added the /export endpoint to tohray. This allows the user to backup the whole blog in either JSON on Markdown format.

A GET call to /export returns JSON (default format). And user can get all the posts in markdown format by attaching ?format=md to the above endpoint, ie., /export?format=md.

Added to urls.nim:

pattern("/export", exportAll, HttpGet, name="export"),

Added to views.nim:

proc exportAll*(ctx: Context) {.async.} =
  let
    format = ctx.getQueryParams("format", "json")
    db = open(consts.dbPath, "", "", "")
    rows = db.getAllRows(sql"SELECT slug, created, content FROM post ORDER BY created DESC")
  defer: db.close()
  if format == "json":
    ctx.response.addHeader("Content-Type", "application/json")
    var jsonArray = newJArray()
    for row in rows:
        var jsonObject = %*{
          "slug": row[0],
          "created": row[1],
          "content": row[2]
        }
        jsonArray.add(jsonObject)
    resp jsonResponse(%*{"posts": jsonArray})
  elif format == "md":
    ctx.response.addHeader("Content-Type", "text/plain")
    var res:string
    for row in rows:
      res = res & "\n" & row[0] & "\n" &  row[1] & "\n\n" & $rows[2] & "\n\n" & chr(28) # ascii file separator
    resp res

Compiling nim programs with C, CPP and ObjC backends

Nim supports multiple compiler backends like C, C++, ObjC and JavaScript.

I did not try compiling the program with JS because many OS facing APIs are not available to the JS backend.

• C* -- nim c -d:release tohray -- 2117992
• CPP -- nim cpp -d:release tohray -- 2443528
• ObjC -- nim objc -d:release tohray -- NA

NA: ObjC compile failed with:

/Users/pradeep/.choosenim/toolchains/nim-2.0.8/lib/pure/math.nim(281, 11) template/generic instantiation of `isNaN` from here
/Users/pradeep/.choosenim/toolchains/nim-2.0.8/lib/pure/math.nim(204, 20) Error: undeclared identifier: 'c_isnan'
candidates (edit distance, scope distance); see '--spellSuggest':
 (1, 3): 'isNaN'

The CPP generated executable is slightly larger. I do not know if it's faster for the same compilation settings.