Merge pull request #397 from chuckwagoncomputing/brokenlinks-on-ketamine

massive to changes to brokenlinks script, etc.
This commit is contained in:
mi-hol 2023-01-20 12:13:45 +01:00 committed by GitHub
commit f8a14461fc
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
4 changed files with 180 additions and 79 deletions

View File

@ -41,7 +41,7 @@
<details><summary><u>PNP Boards</u></summary>
* [PNP 48 - MRE based for Miata NA](/Hardware/pnp_microRusEFI_48na/microRusEFI48adapter_latest.pdf)
* [PNP 48 - MRE based for Miata NA](Hardware/pnp_microRusEFI_48na/microRusEFI48adapter_latest.pdf)
* [PNP 72 - MRE based for Miata NB2](Hardware/pnp_microRusEFI_nb2/hw72nb.pdf)
* PNP 88 - Proteus based for 88pin Bosch - Coming Soon
* [MRE Adapter 48](MREAdapter48)

58
check.txt Normal file
View File

@ -0,0 +1,58 @@
# Bad Link Tests
This file has some examples of good and bad link and image styles.
[good link](Idle-Control)
[bad link](Idle_Control)
should be Idle-Control
[bad link](./Idle-Control)
should be Idle-Control
[bad link](/Idle-Control)
should be Idle-Control
[bad link](./Idle-Control.md)
should be Idle-Control
[bad link](/Idle-Control.md)
should be Idle-Control
[bad link](Idle-Control.md)
should be Idle-Control
[bad link](Idle-Control.md#valve-initialization)
should be Idle-Control#valve-initialization
[bad link](Idle-Control#Valve-initialization)
should be Idle-Control#valve-initialization
[bad link](Idle-Control#valve_initialization)
should be Idle-Control#valve-initialization
![good img](Images/Red_LED.png)
![bad img](Images/Red-LED.png)
should be Images/Red_LED.png
![bad img](Red_LED.png)
should be Images/Red_LED.png
![bad img](Red_LED)
should be Images/Red_LED.png
![bad img](Images/Red_LED)
should be Images/Red_LED.png
![bad img](./Images/Red_LED.png)
should be Images/Red_LED.png
![bad img](./Images/Red_LED)
should be Images/Red_LED.png
![bad img](/Images/Red_LED.png)
should be Images/Red_LED.png
![bad img](/Images/Red_LED)
should be Images/Red_LED.png

View File

@ -4,6 +4,7 @@
# 02/18/2021 #
# Written By David Holdeman #
# Searches for broken links in a Github Wiki repo, and suggests and applies corrections. #
# Usage: brokenlinks.sh [-s non-interactive] [-d debug] <optional file(s)> ... #
##################################################################################################
# These two functions are used to escape variables for use in a sed command
@ -19,105 +20,82 @@ export -f escapeReplace
# return status:
# 0: file is good .md
# 1: file is bad or not .md
# 1: file is bad
# 2: file is not .md
checkurl() {
LINK="$2"
HASH="$3"
# If it's an internet link, ignore it.
# That's beyond the scope of this tool.
if echo "$2" | grep -E '^[http|\/]' >/dev/null; then
return 1
if echo "$LINK" | grep -E '^http' >/dev/null; then
return 2
fi
# At some point in this scripts development, fixed links to files/images were given the './' prefix.
# This didn't really hurt anything, but it's not idiomatic.
# I added this to fix the problems I caused, and decided it was worth keeping around.
if echo "$2" | grep -E '^\./' >/dev/null; then
# NEWLINK is the corrected link
NEWLINK=$(echo "$2" | sed 's/^\.\///')
# Check for links that begin in ./ or /, as they won't function as expected everywhere.
if echo "$LINK" | grep -E '^[.]?/' >/dev/null; then
# Save the link for replacement
OLDLINK="$LINK"
# Correct the link.
# We save this to $LINK because the next check in this function need the corrected version.
LINK=$(echo "$LINK" | sed 's/^.\{0,1\}\///')
# Lock user-facing input/output so that the user is presented with one fix at a time.
(
flock -x 200
# Print the file and the old link
echo "In $1:" >&2
echo "$2" >&2
echo "$OLDLINK" >&2
# Print the options as though they are a list in order to have the same UI as other types of correction
echo "$NEWLINK" | cat --number >&2
echo "$LINK" | cat --number >&2
# Make sure we aren't in non-interactive mode.
if [ "$SCRIPT" -lt 1 ]; then
echo "Type a number, then hit return to select an alternative, or just hit return to skip fixing:" >&2
# Read the user input
read -r PICK
if [ "$PICK" -eq 1 ]; then
# Replace the old link with the new one.
# Parentheses are placed around both the old link and new one in order to ensure we replace the link,
# and not some other place in the file that happens to use the same words.
REPLACE=$(escape '('"$2"'#'"$3"')')
REPLACEWITH=$(escapeReplace "$NEWLINK"'#'"$3")
REPLACE=$(escape '('"$OLDLINK""$HASH"')')
REPLACEWITH=$(escapeReplace "$LINK""$HASH")
sed -i "s/$REPLACE/\($REPLACEWITH\)/" "$1"
# print the new URL for use in checkhash
echo "$NEWLINK"
fi
# We don't continue here because the link we fixed might be broken.
fi
# File descriptor for the lock.
) 200>brokenlinks.lock
fi
# Skip links that are to an .md file and aren't broken.
if [ "$(find . -name "$2"".md" 2>/dev/null | wc -l)" -gt 0 ]; then
# print the URL for use in checkhash
echo "$2"
if [ "$(echo "$LIST" | grep "$LINK"".md" 2>/dev/null | wc -l)" -gt 0 ]; then
# print the URL for use in `checkhash`
echo "$LINK"
return 0
fi
# Process links that are not to an .md file.
# We need to do this separately because Github/gollum behave differently with different kinds of links.
# .md files need the name of the file, without the .md extension.
# Everything else needs the path relative to the repo root.
if echo "$2" | grep -Ev ".md$" >/dev/null; then
# Skip the link if it's not broken.
if ls "$2" 2>/dev/null >/dev/null; then
return 1
fi
# Print the filename and the broken link.
echo "In $1:" >&2
echo "$2" >&2
# Build the search term we will look for.
# All hyphens and underscores are replaced with asterisks, so we
# can find files with mismatched hyphens or underscores.
SEARCH='*'$(basename "$2" | sed 's/[-_ ]/*/g')'*'
# Search for matching files.
FILES=$(find . -iname "$SEARCH")
# If there are no files, skip to next link.
if [ "$(echo -n "$FILES" | wc -c)" -lt 1 ]; then
echo "Could not find" >&2
return 1
fi
# List the potential files, with numbers.
echo "$FILES" | cat --number >&2
if [ "$SCRIPT" -lt 1 ]; then
# Read the user input
read -r PICK
# If the selection isn't a number, skip to the next link.
if ! [[ $PICK =~ ^[0-9]+$ ]]; then
return 1
fi
# Get the selected file path, without the preceding ./
FILE=$(echo "$FILES" | head -n "$PICK" | tail -n 1 | sed 's/^\.\///')
# Replace the old link with the new one.
# Parentheses are placed around both the old link and new one in order to ensure we replace the link,
# and not some other place in the file that happens to use the same words.
REPLACE=$(escape '('"$2"')')
REPLACEWITH=$(escapeReplace "$(basename "$FILE" .md)")
sed -i "s/$REPLACE/\($REPLACEWITH\)/" "$1"
fi
return 1
# Skip non-md links if they're not broken.
if echo "$LINK" | grep -v '.md$' >/dev/null && ls "$LINK" 2>/dev/null >/dev/null; then
return 2
fi
echo "In $1:" >&2
echo "$2" >&2
# Build the search term we will look for.
# All hyphens and underscores are replaced with asterisks, so we
# can find files with mismatched hyphens or underscores.
SEARCH='*'$(basename "$2" | sed 's/[-_ ]/*/g')'*'
SEARCH='*'$(basename "$LINK" | sed 's/[-_ ]/*/g')'*'
# Search for matching files.
# We are using `find` here because we need to search for all files, while $LIST has only .md files
FILES=$(find . -iname "$SEARCH")
# Lock user-facing input/output so that the user is presented with one fix at a time.
(
flock -x 200
# Print the filename and the broken link.
echo "In $1:" >&2
echo "$LINK" >&2
# If there are no files, skip to next link.
if [ "$(echo -n "$FILES" | wc -c)" -lt 1 ]; then
if [ -z "$FILES" ]; then
echo "Could not find" >&2
return 1
fi
# List the potential files, with numbers.
echo "$FILES" | cat --number >&2
# Make sure we aren't in non-interactive mode.
if [ "$SCRIPT" -lt 1 ]; then
echo "Type a number, then hit return to select an alternative, or just hit return to skip fixing:" >&2
# Read the user input
read -r PICK
# If the selection isn't a number, skip to the next link.
@ -125,21 +103,40 @@ checkurl() {
return 1
fi
# Get the selected file path, without the preceding ./
FILE=$(basename "$(echo "$FILES" | head -n "$PICK" | tail -n 1)" .md)
FILE=$(echo "$FILES" | head -n "$PICK" | tail -n 1 | sed 's/^\.\///')
# Track if the linked file is a .md file
MD=0
if echo "$FILE" | grep ".md$" >/dev/null; then
MD=1
# Drop the .md from the link
FILE=$(basename "$FILE" .md)
fi
# Replace the old link with the new one.
# Parentheses are placed around both the old link and new one in order to ensure we replace the link,
# and not some other place in the file that happens to use the same words.
REPLACE=$(escape '('"$2"'#'"$3"')')
REPLACEWITH=$(escapeReplace "$FILE"'#'"$3")
sed -i "s/$REPLACE/\($REPLACEWITH\)/" "$1" >&2
echo "$FILE"
return 0
REPLACE=$(escape '('"$LINK""$HASH"')')
REPLACEWITH=$(escapeReplace "$FILE""$HASH")
sed -i "s/$REPLACE/\($REPLACEWITH\)/" "$1"
# print the URL for use in `checkhash`
echo "$LINK"
if [ "$MD" -eq 1 ]; then
return 0
else
return 2
fi
fi
return 1
# File descriptor for the lock.
) 200>brokenlinks.lock
# The returns within the lock closure don't return from the function, only from the closure.
return $?
}
export -f checkurl
checkhash() {
# $1: file
# $2: hash
# $3: url - won't always be present
# TODO check hash fragment validity
return 0
}
@ -148,43 +145,72 @@ export -f checkhash
# Main processing function
# Passed the path to a .md file
searchfile() {
STATUS=0
# This loops for every link in the file.
# See the end of the function for the grep that finds the links in the file.
# We use file descriptor 3, because if we used stdin, the read calls inside this loop would read from that instead of
# reading the user's input.
while IFS= read -r -u 3 LINK; do
URL=$(echo "$LINK" | cut -d '#' -f 1)
HASH=$(echo "$LINK" | cut -d '#' -f 2)
# Break the link into URL and hash fragment, if one is present
if echo "$LINK" | grep '#' >/dev/null; then
URL=$(echo "$LINK" | cut -d '#' -f 1)
HASH="#"$(echo "$LINK" | cut -d '#' -f 2)
else
URL="$LINK"
HASH=""
fi
# We need to store the return status of `checkurl` to know whether we need to check the hash
URLSTATUS=0
if [ -n "$URL" ]; then
# `checkurl` returns the URL if it is good,
# so that if it was fixed, we have the update version to use in `checkhash`.
URL=$(checkurl "$1" "$URL" "$HASH")
URLSTATUS=$?
if [ "$URLSTATUS" -eq 1 ]; then
STATUS=1
fi
fi
# Only check the hash if it exists and the URL was good.
if [ -n "$HASH" ] && [ "$URLSTATUS" -eq 0 ]; then
# Parameters are reversed because we won't always have a URL - we might only have a hash fragment.
checkhash "$1" "$HASH" "$URL"
if [ "$?" -gt 0 ]; then
STATUS=1
fi
fi
if [ "$DEBUG" -eq 1 ]; then
echo "$(date +%T.%N) $1 $URL $HASH"
fi
# This regex finds links in the file that is passed to searchfile
# Results are fed to file descriptor 3 for the reasons previously explained.
done 3< <(grep -oP '(?<=\]\().*?(?=[\)])' "$1" | sed -e "s/^<//g" -e "s/>$//g" | cut -d '"' -f1 | sed -e 's/^[[:space:]]*//' -e 's/[[:space:]]*$//')
return $STATUS
}
export -f searchfile
FILES=()
export SCRIPT=0
export DEBUG=0
for i in $@; do
if [ "$i" == "-s" ]; then
export SCRIPT=1
elif [ "$i" == "-d" ]; then
export DEBUG=1
else
FILES+=("${i}")
fi
done
export LIST=$(find . -iname "*.md" ! -name '_*')
if [ "${#FILES[@]}" -gt 0 ]; then
# Only run `searchfile` on passed-in file names.
for f in "${FILES[@]}"; do
searchfile "$f"
exit $?
done
else
# run searchfile on every .md file in the repo
find . -iname "*.md" -exec bash -c 'searchfile "$0"' {} \;
# run `searchfile` on every .md file in the repo
xargs -0 -P $(nproc --all) -a <(echo "$LIST" | tr '\n' '\0') -I {} bash -c 'searchfile "$@"' _ {}
fi

View File

@ -1,6 +1,23 @@
#!/usr/bin/env bash
CHANGED=$(git diff --name-only HEAD HEAD~1 | grep -v '^_' | grep '.md$')
if [ "$CHANGED" == "" ]; then
exit 0
STATUS=0
CHANGED=$(git diff --name-only HEAD~1 HEAD | grep -v '^_' | grep '.md$')
if [ -n "$CHANGED" ]; then
markdownlint -i '_*' --disable MD033 MD034 MD013 MD024 MD036 -- $CHANGED
fi
markdownlint -i '_*' --disable MD033 MD034 MD013 MD024 MD036 -- $CHANGED
if [ "$?" -gt 0 ]; then
STATUS=1
fi
if git diff --compact-summary HEAD~1 HEAD | grep -E "=>|\(new\)" >/dev/null; then
bash wiki-tools/brokenlinks.sh -s
if [ "$?" -gt 0 ]; then
STATUS=1
fi
else
bash wiki-tools/brokenlinks.sh -s $CHANGED
if [ "$?" -gt 0 ]; then
STATUS=1
fi
fi
exit "$STATUS"