Re: [Parseperl-discuss] [PATCH] preserve newlines

[Adam K. <ad...@ph...>]

Chris Dolan wrote:
> On Oct 5, 2006, at 12:40 AM, Adam Kennedy wrote:
> 
>>>> That leaves us only with the case of mixed newlines. Personally,  
>>>> outside of binary files I am not away of ANY cases in which mixed  
>>>> newlines in a text file is allowed, even in __DATA__ segments.
>>> Well, certainly my goal is to get rid of the mixed newlines!  That's  
>>> why I was writing a Perl::Critic policy against that.  :-)
>>
>> Well, you know it's possible to detect mixed newlines from the raw 
>> source right?
>>
>> Does critic have access to the original file/string?
> 
> In theory almost always yes, in practice no.  We accept PPI::Document 
> instances, so that's one case where I can't ever get the raw source.  We 
> also accept STDIN, so I'd have to explicitly save the source.  In the 
> most common case, we build a PPI::Document::File.  I *could* of course 
> use the filename() to get back to source.

Oh right, I forgot about the STDIN stuff... so yeah, practically 
speaking you don't have access. Suck.

>> Emacs et al doing binary items inside a text file is one example of 
>> what I mean by handling it wrong.
> 
> Hmm...  Maybe I've just been using Emacs so long that its behavior feels 
> intuitive to me?  :-)

Probably :)

I mean, it's a newline, and Perl doesn't have ANY binary (ignoring 
unicode) content. The fact it isn't a newline for the current platform 
(at least with respect for Perl) should be irrelevant.

> I think the correct behavior on insert is to convert to document newline 
> by default, and possibly offer a way to preserve the mixed newlines if 
> the generator requests it.  The mechanism for that "request" can be a 
> workaround (add_before() then set_newline()), or a flag to add_before(), 
> or a sticky_newline flag on the node to be inserted.

I'm happy to go with "convert to the document default automatically" and 
ignore any customization capability until someone actually asks for it :)

>> I don't think anyone ever really WANTs to make mixed newlines. As I 
>> said, I don't know of any legitimate real-world cases of this.
> 
> ... so probably the generator case is just to adopt the newline of the 
> surrounding document.
> 
> There are plenty of real world cases of mixed newlines, just none of 
> them intentional!

Right :)

So I guess to start with, we need to take the input string, and detect

1. All Unix
2. All Win32
3. All Mac
4. Mixed Unix-dominant
5. Mixed Win32-dominant
6. Mixed Mac-dominant

:/

So... $doc->{newlines} = 'unix|win32|max' and $doc->{newlines_mixed} = 1|0

What do you think?

Adam K
Adam K