how can I upload code to Answers without MarkDown mangling it?

[V_at_Splunk]

Not technically a question, but pretty sure will be helpful to many. If not helpful to you, please don't upvote.

---

MarkDown (the Wiki engine used on this site) respects internal (between tokens on a line) spaces within a block delimited by pre tags; but blithely interprets underscores, asterisks and less-than signs as markup. Thus rendering your shell script or Python snippet into a dire soup. The script below takes your original code, and outputs a text file with all the troublesome chars substituted to evade MarkDown's bumbling "assistance"; just cat the result, copy, and paste here within pre tags.

Here's how you use the script (say we named it ready-code-for-SplunkAnswers.sh😞

$ ready-code-for-SplunkAnswers.sh your-script.py

Result in: readied--your-script.py.txt

---

And here's the script:

#!/bin/bash

# Vainstein K 14nov2013

­

die () {

    echo "(exit=$?) $@" >&2 && exit 42

}

failed () {

    die 'Failed to:' $@

}

­

[ $# -ne 1 ] && die 'USAGE: <fileToReady>'

readonly original=$1

[ -e $original ] || die "File '$original' not found."

readonly readied=readied--$original.txt

cp -p $original $readied || failed "overwrite $readied"

chmod u+w $readied || failed "make writable $readied"

­

# We have to distinguish #s originally there, from the #s that we'll insert as part of escape codes.

readonly fauxOctothorpe='@XXXoctothorpeXXX@'

sed -i 's/#/'$fauxOctothorpe'/g' $readied || failed 'replace: # (phase 1 of 2)'

# Same for &.

readonly fauxAmpersand='@XXXampersandXXX@'

sed -i 's/&/'$fauxAmpersand'/g' $readied || failed 'replace: & (phase 1 of 2)'

­

sed -i 's/`/\&#96;/g' $readied || failed 'replace: `'

sed -i 's/_/\&#95;/g' $readied || failed 'replace: _'

sed -i 's/\*/\&#42;/g' $readied || failed 'replace: *'

sed -i 's/</\&lt;/g' $readied || failed 'replace: <'

sed -i 's/\\/\&#92;/g' $readied || failed 'replace: \\'

­

sed -i 's/'$fauxOctothorpe'/\&#35;/g' $readied || failed 'replace: # (phase 2 of 2)'

sed -i 's/'$fauxAmpersand'/\&#38;/g' $readied || failed 'replace: & (phase 2 of 2)'

­

# MarkDown is too stupid to leave whitespace-only lines alone; so we insert a barely-visible symbol on such.

sed -i 's/^[ \t]*$/\&shy;/' $readied || failed 'replace: whitespace-only line'

­

echo "Result in: $readied"

---

Naturally enough, I prepared this script for upload using... itself.

;-)

Updated: be sure to read discussion with Ayn, in comments below.

[V_at_Splunk · Accepted solution]

hi Ayn,

OK. Let's try this silly, yet valid, shell script, each line offset by 4 leading spaces:

#!/bin/bash
two=2 three=3 four=4 # and 5 is right out
printf "Should print 24: %d\n" $((two*three*four))
echo "The hour is`date | sed -r 's/^.*( [0-9]{2}):.*/\1/'`, and all is well."
echo 'To look for "pre" preceded immediately by a less-than, try this RE: <pre\>'

And now transformed with ready-code-for-SplunkAnswers.sh, and enclosed in pre tags:

#!/bin/bash

two=2 three=3 four=4 # and 5 is right out

printf "Should print 24: %d\n" $((two*three*four))

echo "The hour is`date | sed -r 's/^.*( [0-9]{2}):.*/\1/'`, and all is well."

echo 'To look for "pre" preceded immediately by a less-than, try this RE: <pre\>'

You're right.

[Ayn]

Correct. 🙂

[V_at_Splunk]

hi Ayn,

My solution is still the only one for code snippets which are not "block-quoted": if you want a single-line code snippet to not break up your paragraph, the four-spaces approach obviously won't do.

[Ayn]

Almost want to say 'sorry' because that's a nice script after all 🙂

[Ayn]

What about simply prepending each line in a code block with 4 spaces? That's how code is entered and it hasn't failed me so far...

[jrodman]

You'd sure think the code buttton in the editor here would work. Sad that it doesn't.