<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0//EN" "http://www.w3.org/TR/REC-html40/strict.dtd">
<html><head><meta name="qrichtext" content="1" /><style type="text/css">
p, li { white-space: pre-wrap; }
</style></head><body style=" font-family:'Noto Sans'; font-size:9pt; font-weight:400; font-style:normal;">
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">On Saturday, 14 March 2020 22:50:19 UTC Viktor Dukhovni wrote:</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> On Mar 14, 2020, at 4:00 AM, Mark Delany <b9w@charlie.emu.st> wrote:</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> > Methinks the safest outcome is that TXT substrings should be treated</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> > as an artifact of the limitations of TXT RRs or a side-effect of</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> > serialization and should have no semantic value whatsoever.</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> </span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> I largely share the sentiment.  At this point it looks like the above</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> interpretation would most closely conform to existing practice.</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> </span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> I know of no specifications that concatenate TXT RR strings with spaces.</span></p>
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">and yet, RFC 1035 says this:</span></p>
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">"<character-string> is expressed in one or two ways: as a contiguous set</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">of characters without interior spaces, or as a string beginning with a "</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">and ending with a ".  Inside a " delimited string any character can</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">occur, except for a " itself, which must be quoted using \ (back slash)."</span></p>
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">ask yourself, if you were parsing a zone file, and you found a TXT RR, and it had spaces in it, what would you do? since the <character-string> cannot include interior spaces unless "quoted like this", what does <space> mean?</span></p>
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">it means each word is a character string, or else, spaces are a syntax error unless they are quoted.</span></p>
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">once that's decided, the ignorant wrongheadedness of the SPF interpretation becomes pretty egregious. if it went in as "TXT foo bar" then it should be treated as "foo bar" even if the wire encoding of "TXT foo bar" is in fact ( "foo" "bar" ).</span></p>
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">i realize that SPF has a massive installed base, and that an installed base gives one great market power. but it's still wrong, and should still be fixed. (BIND 4 had a 100% market share, but got this wrong, and so, got fixed.)</span></p>
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> ...</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> </span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> So while I've found no existing practice of joining with spaces, there</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> is it seems an expectation that the component substrings can be</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> separately meaningful in some applications.</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> </span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> So my take is that applications that expect a single string per TXT</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> record should just join without inserting spaces, while applications</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> that expect multiple values can use the verbatim substrings without</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">> concatenation.</span></p>
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">if we're going to write an RFC to clarify this, it could say almost that. but i think the robustness principle calls for something slightly more subtle, and it doesn't have to do with inserting spaces.</span></p>
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">if a multi-segment string is encountered by a TXT-consuming application, and if that multi-segment string can be unambiguously interpreted by the application as some machine-form instruction by concatenating the segments, then this should be done. however, if such concatenation renders an ambiguous result (possibly meaningful but possibly erroneous) then the application should try to interpret each text segment as a separate word, that is, as if separated by whitespace characters. if the segments-as-words interpretation is less ambiguous or less erroneous, then this interpretation should prevail.</span></p>
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">here's what i'm going with, by the way:</span></p>
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">_spf                    TXT     (        v=spf1\032</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro'; font-style:italic;">                              </span><span style=" font-family:'Source Code Pro';">2001:4f8::/32\032</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">                         2001:559:8000::/48\032</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">                         149.20.56.0/24\032</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">                         24.104.150.0/24\032</span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">                         ~all )</span></p>
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">but i'd like to be able to remove those \032 workarounds in ~10 years.</span></p>
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">-- </span></p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;"><span style=" font-family:'Source Code Pro';">Paul</span></p></body></html>