Fixed race conditions

[riquito]

assert.sh fail in a couple cases

1. using assert, if your tested command has as parameter a string with a space separated asterisk then the asterisk will be evaluated by the shell.

2. using assert, if your tested command has as parameter a string with characters that have a special meaning in printf, printf will output something unexpected

I suspect that in assert_raises
(eval $1 <<< ${3:-}) > /dev/null 2>&1
should really be written as
(eval "$1" <<< ${3:-}) > /dev/null 2>&1

but I can't produce a test case similar to 1)

p.s.
I couldn't figure out the meaning of "x required to overwrite older results": is it still needed once printf is gone?

[riquito]

It still fails with
assert "echo "a\nb"" "a\nb"

result is

test #10 "echo "a\nb"" failed:
expected "a\nb"
got "a\nb"

This happens with or without my pull request.

It's compairing the literal "a\nb" (4 letters) with "a
b" (3 letters) (the error message doesn't help here)

[riquito]

The latest commit ( df02b94 ) should solve the newline problem

[lehmannro]

I don't see commit df02b94 in this pull request or your repository. Could you recheck if you've pushed it?

[riquito]

It's two days that I wait for github to update the page (via git I see it pushed). Alas, it has finally appeared, you should be able to see it.

[mnothic]

hi I have a problem with this assert
bash test.sh
test #24 "get_color red" failed:
expected "\032[0;31m"
got "\033[0;31m"
1 of 25 tests failed in 0.171s.

[lehmannro]

It looks like indeed your expected output and the result are different —
one has \032, the other has \033.

For future problems, please also open a new ticket instead of amending an
existing, unrelated issue.

[mnothic]

Sorry, my fault, but with the same string still failing

test #24 "get_color red" failed:
expected "\033[0;31m"
got "\033[0;31m"
1 of 25 tests failed in 0.086s.