parser/pageparser: enforce non-alphanumeric parameters must be quoted

[guilhas07]

Problem:

Before unquoted named parameters would be parsed until a space was
found, allowing for non-alphanumeric parameters.

Solution:

Parse until a space, or a non-alphanumeric character is found.

[jmooring]

This a breaking change.

With v0.143.0 this works fine:

{{< foo param=https://example.ork/foo/?a=b&c=d#bar >}}

But with this PR:

Error: ... unrecognized character in shortcode action: U+003A ':'. Note: Parameters with non-alphanumeric args must be quoted

[jmooring]

I suggest you create a topic in the forum to discuss the particular problem you are having. If it's a bug or opportunity for improvement, we'll create an issue. Please do not submit PRs without a corresponding issue that's been approved by the project lead.

[guilhas07]

This a breaking change.

With v0.143.0 this works fine:

{{< foo param=https://example.ork/foo/?a=b&c=d#bar >}}

But with this PR:

Error: ... unrecognized character in shortcode action: U+003A ':'. Note: Parameters with non-alphanumeric args must be quoted

As you can see in the error, which was not added by me, Parameters with non-alphanumeric args must be quoted:

hugo/parser/pageparser/pagelexer_shortcode.go

Line 339 in 377287a

return l.errorf("unrecognized character in shortcode action: %#U. Note: Parameters with non-alphanumeric args must be quoted", r)

That is why I made this change. This makes unnamed and named parameters behaviour consistent.
This is currently not valid in v0.143.0:

{{< foo https://example.ork/foo/?a=b&c=d#bar >}}

Maybe I should have been more obvious in the pull request/commit description.

Please do not submit PRs without a corresponding issue that's been approved by the project lead.

I submitted a similar fix for this part of the code which was approved #13299 and noticed multiple PRs from non-members that didn’t seem to be associated with any approved issues so I made this honest mistake. See issue #13343