[
https://issues.apache.org/jira/browse/TEXT-88?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16048980#comment-16048980
]
Gilles commented on TEXT-88:
----------------------------
bq. reply appropriately
When a library method gets {{null}}, it cannot know whether it was intentional
or not.
I'd certainly prefer being warned that something is shaky (through NPE) rather
than the program continuing a computation that is not what I intended.
The "feature" may be fine at the _application_ level (where context in better
known).
bq. mentioning behavior in Javadoc of each method is good enough to tell users
how it will work.
It will not save them from a programming error.
Why is it not sufficient to have syntactic sugar:
{code}
WordUtils.capitalizeFully(string)
{code}
that is documented to behave like
{code}
WordUtils.capitalizeFully(string, delim)
{code}
where "delim" is e.g. whitespace?
> WordUtils.capitalizeFully behaves in a counterintuitive manner with empty
> delimiter array.
> ------------------------------------------------------------------------------------------
>
> Key: TEXT-88
> URL: https://issues.apache.org/jira/browse/TEXT-88
> Project: Commons Text
> Issue Type: Bug
> Reporter: Rob Tompkins
>
> As discussed in TEXT-85, it seems that
> {code:java}
> WordUtils.capitalizeFully("i am fine", new char[]{}) // --> i am fine
> {code}
> Both Sebb and Arun think that this is odd and that the letter "i" should have
> been capitalized resulting in the output {{I am fine}}
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
