1
0
Fork 0
mirror of https://git.sr.ht/~seirdy/seirdy.one synced 2024-11-23 12:52:10 +00:00

Refactor shell scripts.

- Avoid relative paths, so I can run them from any working directory.
- Make a dedicated curl-wrapping shell script instead of re-defining the
  same alias everywhere.
- Support extended offline periods: allow get-webmentions.sh to fall
  back to the cached copy of my webmentions for up to a day, and don't
  accidentally destroy it; keep changes atomic.
- Verify that the fetched webmentions are legit before replacing the
  cached ones.
- Make shellcheck happy about quoting in vnu.sh by passing the list of
  files with xargs instead of a shell variable.
This commit is contained in:
Rohan Kumar 2023-11-23 18:45:03 -08:00
parent a3e4729b6c
commit 2b016aa973
No known key found for this signature in database
GPG key ID: 1E892DB2A5F84479
5 changed files with 66 additions and 27 deletions

View file

@ -1,7 +1,7 @@
#!/bin/sh
# Runs axe on every page of my sitemap
# first arg is output directory, after that comes urls from a sitemap. We replace the urls with equivalent local file paths.
PATH="scripts/bin:$PATH"
PATH="$(dirname "$0"):$PATH"
output_dir="$1"
shift
urls="$(echo "$*" | tr ' ' '\n' | sd '^https?://[^/]*' "file://$PWD/$output_dir" | sd '/$' '/index.html' | tr '\n' ' ')"

View file

@ -1,14 +1,14 @@
#!/bin/sh
set -e -u
alias curl_cmd='curl --proto "=https" --proto-default https --http2 -siSL --tlsv1.3 --cert-status'
ipv6_success=1
ipv4_success=1
curl_cmd -6 'seirdy.one/ip' || ipv6_success=0
curl_wrapper="$(dirname "$0")/curl-wrapper.sh"
"$curl_wrapper" -6 'https://seirdy.one/ip' || ipv6_success=0
echo
curl_cmd -4 'seirdy.one/ip' || ipv4_success=0
"$curl_wrapper" -4 'https://seirdy.one/ip' || ipv4_success=0
echo
if [ "$ipv6_success" = 0 ] && [ "$ipv4_success" = 0 ]; then

3
scripts/curl-wrapper.sh Executable file
View file

@ -0,0 +1,3 @@
#!/bin/sh
# a curl wrapper for seirdy.one
curl --proto "=https" --tlsv1.3 --cert-status -sS -m10 "$@"

View file

@ -2,18 +2,48 @@
#
# Script to fetch all approved webmentions from webmentiond as a big json response.
# Uses POSIX and cURL in CI, also uses any pass/pash-compatible pwmngr otherwise
# The response is cached for 90 minutes.
# The response is cached for 90 minutes. If fetching fresh webmentions fails, fall back to the cached version for up to one day (1440 minutes). If cached webmentions are older than a day, error out.
set -e -u
dirname="$(dirname "$0")"
curl_wrapper="$dirname/curl-wrapper.sh"
auth_url='https://seirdy.one/webmentions/authenticate/access-key'
webmentions_url='https://seirdy.one/webmentions/manage/mentions?limit=9999&status=approved'
webmentions_file="$(realpath data/webmentions.json)"
webmentions_file="$dirname/../data/webmentions.json"
# just a little curl wrapper I use on seirdy.one
alias ccurl='curl --proto "=https" --proto-default https --tlsv1.3 --cert-status'
stale_after_minutes=90
fallback_stale_after_minutes=1440
# use a long-lived key (password) to fetch a short-lived bearer token.
skip_check=0
check_cached_webmentions() {
if [ "$skip_check" = 0 ]; then
expires_in="$fallback_stale_after_minutes"
if [ $# -gt 0 ]; then
expires_in="$1"
else
echo "Failed to fetch webmentions. Using fallback cache expiry."
fi
exit_status=1
if [ -f "$webmentions_file" ]; then
old_webmention_file="$(find "$webmentions_file" -mmin +"$expires_in")"
if [ "$old_webmention_file" = "" ]; then
echo 'Using cached webmentions'
skip_check=1
exit_status=0
fi
fi
if [ "$exit_status" = 1 ]; then
echo "webmentions are outdated. failed to fetch for over a day."
fi
exit "$exit_status"
fi
}
trap check_cached_webmentions EXIT
# Grab my long-lived key (password). We will use this to authenticate.
key() {
set +u
if [ -n "$BUILD_SUBMITTER" ]; then
@ -24,23 +54,32 @@ key() {
set -u
}
# Fetch a short-lived access token from my webmention receiver.
token() {
key_response="$(key)"
ccurl -sX POST "$auth_url" -d "key=$key_response"
"$curl_wrapper" -sX POST "$auth_url" -d "key=$key_response"
}
# use that token to fetch all webmentions
# Verify that the webmentions file has downloaded succesfully by ensuring that it starts and ends with a given substring.
# The total number of webmentions should be at least a 3-digit number.
verify_webmentions() {
grep -E '^\{"items":\[\{"id":".*,"total":[0-9]{3}([0-9]*)?\}$' "$webmentions_file.tmp" >/dev/null
}
# use the token to fetch all webmentions.
fetch_webmentions() {
echo 'Fetching webmentions'
token_response="$(token)"
ccurl --compressed -H "Authorization: Bearer $token_response" "$webmentions_url" -o "$webmentions_file"
"$curl_wrapper" --compressed -H "Authorization: Bearer $token_response" "$webmentions_url" -o "$webmentions_file.tmp" || return 1
if verify_webmentions; then
mv "$webmentions_file.tmp" "$webmentions_file"
else
echo 'Error: webmentions failed to verify'
exit 1
fi
}
# fetch webmentions if we don't have a fresh copy already.
if [ -f "$webmentions_file" ] \
&& [ "$(find "$webmentions_file" -mmin +90)" = "" ]; then
echo 'Using cached webmentions'
else
echo 'Fetching webmentions'
if ! check_cached_webmentions "$stale_after_minutes"; then
fetch_webmentions
fi

View file

@ -1,6 +1,6 @@
#!/bin/sh
set -e -u
pwd="$(dirname "$0")"
dirname="$(dirname "$0")"
output_dir="$1"
find_files_to_analyze() {
@ -8,12 +8,9 @@ find_files_to_analyze() {
| grep -Ev '(bimi\.svg|search/index\.x?html)$'
}
files_to_analyze="$(find_files_to_analyze)"
# files_to_analyze="$(find_files_to_analyze)"
# we skip the BIMI icon (VNU can't handle SVG 1.2) and the search page (it has raw templates).
vnu \
--stdout \
--format json \
--also-check-svg \
$files_to_analyze \
| sh "$pwd/filter-vnu.sh"
find_files_to_analyze \
| xargs vnu --stdout --format json --also-check-svg \
| sh "$dirname/filter-vnu.sh"