Suggestions

Avoid flaky tests

Avoid flaky tests (tests that sometimes pass and sometimes fail) for example when using dates or random numbers, you can avoid this by replacing the random number with a fixed number or using a regex to match the output.

Bad:

# The date will change every day, making tests flaky
date

sáb 14 oct 2023 11:26:38 -03

Good:

date | sed 's/\w\+ [0-9]\+ \w\+ [0-9]\+ [0-9]\+:[0-9]\+:[0-9]\+ -[0-9]\+/thu 00 oct 0000 00:00:00 -00/'

thu 00 oct 0000 00:00:00 -00

Avoid long outputs

This could cause some problems with Jupyter Notebook tests, depending on the settings, this will truncate very long outputs, the tests will keep working but you only will see part of the output,

Bad:

function nstrings () {
for ((i = 1; i <= $1; i++)); do
    echo -en "$2"
done
}
nstrings 50 "this is a long output\n"

This will output:

this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
this is a long output
...
this is a long output
this is a long output
this is a long output
this is a long output
Output is truncated. View as a scrollable element or open in a text editor. Adjust cell output settings...

Good:

nstrings 50 "this is a long output\n" | wc -l

this outputs:

50