{{substr "BatMan" -3}} generating error

[jeremielp]

What version of Hugo are you using (hugo version)?

$ hugo version
Hugo Static Site Generator v0.75.0-FEF924BA windows/amd64 BuildDate: 2020-09-14T09:19:16Z

Does this issue reproduce with the latest release?

Not tested.

Looking at this page of the documentation:
https://gohugo.io/functions/substr/
It says: "To extract characters from the end of the string, use a negative start number."

I was expecting this to happen, like the substr PHP function:

{{substr "BatMan" -3}} → "Man"
{{substr "BatMan" -2}} → "an"

Yet, I got errors

{{substr "BatMan" -3}}
execute of template failed: template: partials/test.html:1:3: executing "partials/test.html" at <substr "BatMan" -3>: error calling substr: runtime error: slice bounds out of range **[-2:]**

{{substr "BatMan" -2}}
execute of template failed: template: partials/test.html:1:3: executing "partials/test.html" at <substr "BatMan" -3>: error calling substr: runtime error: slice bounds out of range **[-1:]**

This one is working and returning the full string
{{substr "BatMan" -1}} → "BatMan"

[jmooring]

{{ substr "BatMan" -1 3 }} → "Man"
{{ substr "BatMan" -2 3 }} → "tMa"
{{ substr "BatMan" -3 3 }} → "atM"

With a negative start number you're saying, "Start this many characters from the end, and go backwards this many characters."

[jeremielp]

This is not coherent with any other programming languages.

How one can know that -1 means the end of the string and -2 means 1 char from the end. Because if I follow your logic, then -1 should mean start 1 character from the end.

At least an example should be added to the doc

[jmooring]

I agree. I expected {{ substr "BatMan" -2 }} to yield an.

[jmooring]

OK, so the documentation now shows the current behavior, but I'm not sure if the current behavior is correct for this example:

{{ substr "BatMan" -3 3 }} → "atM"

Shouldn't this be "Man" instead? There's a test for this in strings_test.go, but I think the test is wrong:

{"abcdef", -3, 3, "bcd"},

Shouldn't this be "def" instead?

We're extracting backward instead of forward when the start number is negative, but I think we should always be extracting forward (left to right).

EDIT: I think Python goes right to left with a negative start number. Perhaps we're modeling that behavior.

[jmooring]

@moorereason,

Thank you for addressing this. There is one remaining problem.

Statement	Expected Return	Actual Return
substr "abcdef" 0 0	empty string	abcdef
substr "abcdef" 1 0	empty string	bcdef
substr "abcdef" -1 0	empty string	f

If length is zero, you should always get an empty string. This matches PHP substr behavior.

[moorereason]

Thanks, @jmooring. I've submitted a new PR to fix that case.

[github-actions]

This issue has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.