Collect file extension stats in gh-pages

.github/workflows/deploy.yml

@@ -1,10 +1,11 @@
-name: GitHub Pages
+name: Build Stats
 
 on:
   push:
     branches:
       - main
   pull_request:
+  workflow_dispatch:
   schedule:
     # run every Sunday at 22:12 UTC
     - cron: '12 22 * * 0'
@@ -13,12 +14,11 @@ jobs:
   deploy:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v2
+      - uses: actions/checkout@v3
 
-      - uses: actions/checkout@v2
-        with:
-          repository: mozilla/gecko-dev
-          path: gecko-dev
+      - name: Clone `gecko-dev`
+        run: |
+          git clone https://github.com/mozilla/gecko-dev.git
 
       - run: npm install -g mustache
 
@@ -29,6 +29,10 @@ jobs:
           cat build/data.json
           mustache build/data.json index.mustache > build/index.html
           cat build/index.html
+
+      - name: Gather commits by day and file extension statistics
+        run: |
+          ./01stats.sh gecko-dev build
 
       - name: Deploy
         uses: JamesIves/[email protected]

.gitignore

@@ -0,0 +1 @@
+gecko-dev/

01stats.sh

@@ -0,0 +1,21 @@
+#!/bin/bash
+
+REPODIR=${1:-.}   # default to . if first argument is not set
+SCRIPTDIR=$(dirname "$(readlink -f "${BASH_SOURCE[0]}")")
+BUILDDIR=$(readlink -f "${2:-$SCRIPTDIR}")
+
+[[ -d $BUILDDIR ]] || { echo "$BUILDDIR dir does not exist"; exit 1; }
+
+(cd "$REPODIR" || exit
+
+  if $(git rev-parse --is-shallow-repository); then
+    echo "WARNING: Repository checkout is shallow. Data is incomplete."
+  fi
+
+  echo "writing $BUILDDIR/commits-per-day.txt"
+  git log --date=short --format="%cd" --reverse --topo-order |
+    uniq -c > "$BUILDDIR"/commits-per-day.txt
+
+  echo "writing $BUILDDIR/extensions.txt"
+  "$SCRIPTDIR"/dev/git-file-stats > "$BUILDDIR/extensions.txt"
+)

README.md

@@ -1,13 +1,12 @@
 # firefox-lang-stats
 
-How much Rust in Firefox? You can see the chart [here][gh-pages].
+How much Rust in Firefox? See [the chart].
 
-Repository contains scripts collecting the language statistics for
-[mozilla/gecko-dev] repository. Gathered data then assembled into a pie chart
-`index.mustache`.
-
-Page is updated weekly with a Travis Cron job. See the `date` meta tag for
-the exact build time.
+Repository contains scripts collecting language statistics for
+[mozilla/gecko-dev] repository. The data is saved as JSON in
+[gh-pages] branch along with a pie chart, which is rendered from
+`index.mustache` template. The GitHub Pages is updated by weekly
+CI job, with the build time recorded in `date` meta tag.
 
 # Build
 
@@ -25,4 +24,5 @@ mustache data.json index.mustache > index.html
 
 
 [mozilla/gecko-dev]: https://github.com/mozilla/gecko-dev
-[gh-pages]: https://4e6.github.io/firefox-lang-stats/
+[the chart]: https://4e6.github.io/firefox-lang-stats/
+[gh-pages]: https://github.com/4e6/firefox-lang-stats/tree/gh-pages

dev/build-data

@@ -34,7 +34,7 @@ echo_data_footer()
 # print a row for a diagram
 echo_data_line()
 {
-  echo "{\"name\":\"$1\", \"loc\":$2}$3"
+  echo "  {\"name\":\"$1\", \"loc\":$2}$3"
 }
 
 echo_meta_date()

dev/git-file-stats

@@ -10,7 +10,8 @@
 #  54 xml
 # 106 xsd
 
-git ls-files | # list files
-  xargs -n 1 -i sh -c 'v="{}"; echo "${v##*.}"' | # get extensions
-  sort | uniq -c | # group
-  sort -n -k 1 # sort by first column
+git ls-files |           # list files
+  grep -Po "(?<=\.)\w+$" |   # match strings and show only matched lines
+  sort |                 # sort before counting to avoid group splits
+  uniq -c |              # count unique
+  sort -n -k 1           # sort naturally by count (first column)