Re: [xml2rfc] [Rfc-markdown] [xml2rfc-dev] <br> is back, was: New xml2rfc release: v2.32.0

[Carsten Bormann <[e-mail]>]

Hi Henrik,

> On Oct 18, 2019, at 15:04, Henrik Levkowetz <[e-mail]> wrote:
> 
> Signed PGP part
> 
> On 2019-10-18 13:33, Carsten Bormann wrote:
>> On Oct 18, 2019, at 12:36, Julian Reschke <[e-mail]> wrote:
>>> 
>>> I note that there is one XML in AUTH48 using this (RFC 8668), so this
>>> really is a bit pressing.
> 
> All of the entity references used during the RPC work are converted to
> unicode code points when the prepped file is produced.  If you see an
> entity reference in a prepped file it's either a bug or it has been added
> manually.

I think that was what Julian was afraid of: Having U+2028 in an immutable published RFC.

>> This uses a &br; entity reference — we can easily change that entity into
>> “<br />” later without touching the XML file :-)
>> 
>> (I’m not finding the “rfc2629-xhtml.ent”, so maybe that already is the case.)
> 
> It’s been in the xml2rfc distribution for ages.  

I have 424 files with that name on my laptop.
These fall into the following equivalence classes (grouped by md5sum):

 316 1205eb5efbbc8d9a734ba77055388d70
  88 1aa6d2431ef0219b231913c8fb3c9253
   2 234420ff1ceb61201fa41655b841513a
   3 46cad1ba9b921fe41a9102e912073d74
   2 7a013cd802f0df7e3d9000cd85b9749f
   2 a66a1389336402406e917908aa6e3255
   8 cc77bb83d0c9c7c19afd2da03140f1b7
   3 d2faa8069ff9e2ef2c7c097d2e17cbad

Of these, I find these in xml2rfc:

  22 1205eb5efbbc8d9a734ba77055388d70
   2 234420ff1ceb61201fa41655b841513a
   2 a66a1389336402406e917908aa6e3255
   8 cc77bb83d0c9c7c19afd2da03140f1b7

I didn’t diff all these combinations against each other, but one recent change seems to be:

<!-- Typographic help characters -->
<!ENTITY zwsp   "&#8203;"><!-- U+232A RIGHT-POINTING ANGLE BRACKET        -->
<!ENTITY br     "&#8232;"><!-- U+2028 LINE SEPARATOR                      -->
<!ENTITY wj     “&#8288;"><!-- U+2060 WORD JOINER                         -->

And more recently:

<!-- Typographic help characters -->
<!ENTITY zwsp   "&#8203;"><!-- U+200B ZERO WIDTH SPACE                    -->
<!ENTITY nbhy   "&#8209;"><!-- U+2011 NON BREAKING HYPHEN                 -->
<!ENTITY br     "&#8232;"><!-- U+2028 LINE SEPARATOR                      -->
<!ENTITY wj     "&#8288;"><!-- U+2060 WORD JOINER                         -->

> If it's desired to make it
> more widely available, I’m sure we could do that.

I was looking for a statement about the definitive source (and thus the definitive version), not a 425th copy…

(It also should be put into the RFC-editor’s /authors directory, so that the authoring-RFCXMLv3 files there can be validated.)

Grüße, Carsten