On Sat, 21 Feb 2015, Wilfred van Velzen wrote to mark lewis:

 MvdV>> I request that the following option be added to Makenl:

 MvdV>> RemoveBOM 1

 MvdV>> What should it do?

 MvdV>> Remove any UTF-8 Byte Order Mark (BOM).

 WvV>> Why should it be an option? Is there a situation where you want to
 WvV>> keep the BOM?

 ml> what would happen on big endian machines if they tried to compile 
 ml> a UTF-8 nodelist without a BOM in it? perhaps something is being 
 ml> missed in this thinking and request to remove the BOM? granted, 
 ml> we're talking about UTF-8 and not UTF-16 or UTF-32 where it might 
 ml> matter... after some bit of time, yeah, i think i can agree with 
 ml> the request, too...

 WvV> The proposed option is about removing BOMs from the imput files.

yes, i understand this...

 WvV> So the generated nodelist, doesn't contain a number of BOMS in 
 WvV> the middle of it. 

this i didn't quite understand... i'm not sure how one can assemble a
segment by simply concatenating files together... the CRC would not be
right if it exists at all...

 WvV> Whether or not the generated nodelist, should contain a BOM on 
 WvV> the front of the file is a different matter, and this could be a 
 WvV> usefull option in makenl.

i view it as all the same... individual segments are made in the same
manner as the full nodelist... as i said, after thinking about it, removing
the BOMs from the input files is ok by me... there definitely should not be
any in the rest of the document and certainly not in the middle of some
line(s)... granted, the UTF-8 specs state that BOMs in the middle of the
lines are to be treated in a specific manner which would basically render
them invisible but still they shouldn't be anywhere except at the beginning
of a document :shrug:

)\/(ark