First of all, this says right in the next sentence
"the result of that evaluation, together with the id-expression, determines the result of the entire postfix expression."
So the "template" obviously doesn't affect the value of the entire postfix expression (it's not mentioned here).

[temp.names] p5 says (essentially) that "template" is only a parser guide.

There are also some changes in this area by the pending omnibus name lookup paper
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2020/p1787r5.html

I'm not seeing a defect here.

First of all, this says right in the next sentence
"the result of that evaluation, together with the id-expression, determines the result of the entire postfix expression."

It only claims that the result is determined. How does this answer «how the result of «postfix-expression . template id-expression» is determined?»

So the "template" obviously doesn't affect the value of the entire postfix expression

It could be obvious, but this doesn't seem to be specified by the standard.

[temp.names] p5 says (essentially) that "template" is only a parser guide.

Is [temp.names]/5 the paragraph you wanted to mention? I don't see it there. Could you cite the wording please?

I would just propose

--- a/source/expressions.tex
+++ b/source/expressions.tex
@@ -3224,7 +3224,7 @@ converted to the equivalent form \tcode{(*(E1)).E2}; the remainder of
 
 \pnum
 Abbreviating
-\grammarterm{postfix-expression}\tcode{.}\grammarterm{id-expression}
+\grammarterm{postfix-expression}\tcode{.}\opt{\terminal{template}}\grammarterm{id-expression}
 as \tcode{E1.E2}, \tcode{E1} is called the \defn{object expression}.
 If the object expression is of scalar type,
 \tcode{E2} shall name the pseudo-destructor

but I wasn't sure the form «postfix-expression . template id-expression» is not handler somewhere else.

"determined": Yes, that's exactly what it says: It says that the result is determined from the postfix-expression before the dot and the id-expression. Note the absence of "template", thus the presence or absence of "template" does not contribute to the determination of the result. Which is exactly what we want.

I think your suggested change is not helpful, because it possibly equates E2 with "template id-expression", which falls apart when we dissect E2 further down.

So, you want to say something like: paragraph 1 says that the result is determined by the expression in front of . and the id-expression, but not template. The rest of the subclause speaks about postfix-expression . id-expression and how the result is determined for it. And it should be deduced that for postfix-expression . template id-expression, the result is the same as for postfix-expression . id-expression because the paragraph 1 didn't say that template is involved in result determination? Looks very intricate.

I think your suggested change is not helpful, because it possibly equates E2 with "template id-expression", which falls apart when we dissect E2 further down.

template_opt can be moved out of E2, abbreviating as E1. template[opt] E2.

Editorial meeting consensus: Yes, this does need a bit of clarification, but this should be handled by CWG.