I just looked at an exported .AMF, exported by OpenSCAD 2025.10.25 on Windows.
Using Notebook++ it shows as double spaced lines, ie line1/blank-line/line2/blank-line etc. (shows
encoding UTF-8)
Turning on Show Symbols, it is line1CRCRLFline2CRCRLF etc.

I went looking for specs for AMF, which points to XML, couldn't find a 'line' specification so then
asked AI (below)

[Note: Windows Notepad doesn't show blank lines, status bar says UTF-8 'Windows (CRLF)']

Per AI below, blank lines is the accurate representation per the spec. I presume we don't want to
have blanks.

tl;dr I know the Unix/DOS thing, but should it be either CRLF or LF, but never CRCRLF?

ChatGPT says:

What XML allows at the physical level

An XML document may contain line breaks encoded as:

•     LF (#xA, \n)
  

•     CR (#xD, \r)
  

•     CRLF (#xD #xA, \r\n)
  

All three are permitted in the raw document.

What XML defines logically

XML defines a single logical line break called LINE FEED (LF, #xA).

Before the XML parser processes the document:

•     Every CRLF is converted to LF
  

•     Every standalone CR is converted to LF
  

This process is called end-of-line normalization and happens before tokenization and parsing.

So, internally, an XML parser sees only LF line breaks, regardless of how the document was stored.

My prompt: So a sequence "someXMLtext#xD#xD#xAmoreXMLtext" would be a line with "someXMLtext" and a
next line of "moreXMLtext"

Almost - but there's an important subtlety.

Given the raw character sequence:

someXMLtext #xD #xD #xA moreXMLtext

XML end-of-line normalization (applied left to right)

The XML processor applies these rules:

1.  Any #xD #xA sequence ? single #xA
   

2.  Any remaining standalone #xD ? #xA
   

Apply that to your sequence:

•     First #xD (not followed by #xA) ? #xA
  

•     Remaining #xD #xA ? #xA
  

Result after normalization

someXMLtext #xA #xA moreXMLtext

That is two line feeds, not one.

Logical structure seen by the parser

This corresponds to:

•     Line 1: someXMLtext
  

•     Line 2: (empty)
  

•     Line 3: moreXMLtext
  

So there is a blank line between the two text lines.

Key takeaway

•     #xD #xA ? one line break
  

•     #xD #xD #xA ? two line breaks
  

•     XML does not collapse adjacent line breaks; it preserves their count after normalization
  

Ignore that. I just noticed the AMF deprecation warning...

From: Michael Marx (spintel) via Discuss [mailto:discuss@lists.openscad.org]
Sent: Thursday, January 01, 2026 5:56 PM
To: 'OpenSCAD general discussion Mailing-list'
Cc: Michael Marx (spintel)
Subject: [OpenSCAD] AMF export - line break?

I just looked at an exported .AMF, exported by OpenSCAD 2025.10.25 on Windows.
Using Notebook++ it shows as double spaced lines, ie line1/blank-line/line2/blank-line etc. (shows
encoding UTF-8)
Turning on Show Symbols, it is line1CRCRLFline2CRCRLF etc.

I went looking for specs for AMF, which points to XML, couldn't find a 'line' specification so then
asked AI (below)

[Note: Windows Notepad doesn't show blank lines, status bar says UTF-8 'Windows (CRLF)']

Per AI below, blank lines is the accurate representation per the spec. I presume we don't want to
have blanks.

tl;dr I know the Unix/DOS thing, but should it be either CRLF or LF, but never CRCRLF?

ChatGPT says:

What XML allows at the physical level

An XML document may contain line breaks encoded as:

•  LF (#xA, \n)
  

•  CR (#xD, \r)
  

•  CRLF (#xD #xA, \r\n)
  

All three are permitted in the raw document.

What XML defines logically

XML defines a single logical line break called LINE FEED (LF, #xA).

Before the XML parser processes the document:

•  Every CRLF is converted to LF
  

•  Every standalone CR is converted to LF
  

This process is called end-of-line normalization and happens before tokenization and parsing.

So, internally, an XML parser sees only LF line breaks, regardless of how the document was stored.

My prompt: So a sequence "someXMLtext#xD#xD#xAmoreXMLtext" would be a line with "someXMLtext" and a
next line of "moreXMLtext"

Almost - but there's an important subtlety.

Given the raw character sequence:

someXMLtext #xD #xD #xA moreXMLtext

XML end-of-line normalization (applied left to right)

The XML processor applies these rules:

1. Any #xD #xA sequence ? single #xA

2. Any remaining standalone #xD ? #xA

Apply that to your sequence:

•  First #xD (not followed by #xA) ? #xA
  

•  Remaining #xD #xA ? #xA
  

Result after normalization

someXMLtext #xA #xA moreXMLtext

That is two line feeds, not one.

Logical structure seen by the parser

This corresponds to:

•  Line 1: someXMLtext
  

•  Line 2: (empty)
  

•  Line 3: moreXMLtext
  

So there is a blank line between the two text lines.

Key takeaway

•  #xD #xA ? one line break
  

•  #xD #xD #xA ? two line breaks
  

•  XML does not collapse adjacent line breaks; it preserves their count after normalization