Optional positional params in shortcodes

[paulcmal]

Hi,

Following a discussion in the forum, I'm opening a bug here.

Currently in Hugo v0.38, when dealing with shortcodes, you can .Get a missing named parameter, but not a missing positional parameter.

The former will return an empty string, which integrates perfectly with with, if or cond. The latter will produce errors like so:

ERROR 2018/03/28 14:51:23 No shortcode param at .Get 1 in page posts/2018/01/projection-de-conquerir-notre-9346.fr.md, have params: [BXv4Txr9CYU]

Like i said on the forum:

I think the problem is that named parameters when queried by .Get return "" if they don’t exist, while positional parameters raise an error instantly.

Adressing this issue by returning an empty string when a positional parameter is missing woud allow us to write code like so to handle parameters:

{{ $width := cond (gt (len .Params) 1) (.Get 1) "512" }}

I'll try to submit a PR soon :)

[paulcmal]

I was investigating the tests, and i stumbled upon TestNestedNamedMissingParam. I assumed {{% div2 %}}d2{{% /div2 %}} would produce an error because of a missing param, as the shortcode looks like this:

<div {{with .Get 0}} class="{{ . }}"{{ end }}>

However, .Get always returns nil when no params are passed (params length 0). But when it comes to named parameters, "" is returned when a key is missing

Shouldn't these behaviours be harmonized, as they both indicate a missing param?

[bep]

Shouldn't these behaviours be harmonized, as they both indicate a missing param?

I'm no sure I follow. Which test case fails that you mean should pass?

[paulcmal]

Which test case fails that you mean should pass?

The test passes. That's normal. I'm just wondering if attempting to get a missing param (whether it's because params is empty, or because the param was not found) should return the same value (either nil or ""), as opposed to now where it can either return nil or "" depending on the cases.

This was just a side-issue not directly related to the ticket, but it stood out as i was wondering whether to return nil or "" in case of a missing positional param, and I couldn't figure the reason these two cases had different return values.

[paulcmal]

I'm working on a PR at the moment.

I stumbled upon this discussion in the forums:

However, when the id parameter is explicitly set to an empty string in the shortcode call (id="") the else clause is executed (the same as when id is not passed at all), unlike what you would expect for an isset condition.

So i would be inclined to return nil specifically, but from what i read of the isset implementation, this would not address the point that person made. Am i correct to think that both nil and ""'s reflect.IsValid return false and therefore cannot be distinguished on the template side? At the moment I'll go with returning "" like with named parameters.

[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.