I have a problem with a complex replacement algorithm. In the end I was able to reduce the problem to this minimal code:
const input="test hello test world"
let start = 0
let output = [...input]
const replacements = []
for (let end = 0; end <= input.length; end++) {
const c = input[end]
if (c == ' ') {
if (start !== end) {
const word = input.substring(start, end).toLowerCase()
if (word == 'test') {
replacements.push({start, length:(end - start), text:'REPLACEMENT'})
}
}
start = end + 1
}
}
for(let i=replacements.length-1;i>=0;i--) {
output.splice(replacements[i].start, replacements[i].length, replacements[i].text)
}
console.log(output.join(''))
My input is "test hello test world"
and the expected output would be "REPLACEMENT hello REPLACEMENT world"
, but it is actually "REPLACEMENT hello tREPLACEMENTworld"
. I can remember from the Twitter API that JavaScript has a strange way to handle byte positions and char indices. So the issue is caused oblicious by the smiley.
How can I fix my code, so that the replacement works as expected? Bonus question why is that happening?
Well that was quick:
When I use output array as input the indices work as expected and my replacement works again. However I will give anyone the accepted state who can explain why that change is required.