We can describe a symtrivial module overLet $R$ be a Dedekind domain $R$ with field of fractions $K$ as follows:
(1) Torsion-free symtrivial modules are submodules of $K$.
(2) The torsion part of a symtrivial module is symtrivial
(3) A torsion module $M$ is symtrivial if and only if itits maximal torsion-free quotient $F$ is a directsubmodule of $K$ and its torsion submodule $T$ is a sum over each prime $p$ such that $F$ is $p$-divisible of symtriviala $p$-powerdivisible $p$-torsion modules
(4) Amodule and a $p$-power-torsion cyclic module $M$ is symtrivial if$R/p^n$.
If $F$ and only if it satisfies either $M/pM=0$ or$T$ are modules as described, then we can classify extnesions $M/pM=R/pR$$0 \to T \to M \to F \to 0$ by the group $Ext^1(F,T)$. If $F\neq 0$, this group is equal to $\prod_p R/p^n / \sum_p R/p^n$.
(5) If Proof: First we prove that such a $p$-power-torsion module is symtrivial. To do this, let $M$ satisfies$a$ and $M/pM=R/pR$, then it is$b$ be two elements and let $I$ be the direct sumsubmodule of a $p$-divisible$F$ generated by $p$-torsion module$a$ and a cyclic module$b$ modulo $T$. (Obviously if$I$ is a finitely generated submodule of $M/pM=0$ then$M$, hence it is projective, so we can find an injection $I \to M$. Hence $a$ and $b$ can be written as $x_1+t_1$, $x_2+t_2$ with $x_1,x_2 \in I$, $t_1,t_2 \in T$. We need to check:
a. $p$-divisible$x_1 \otimes x_2 = x_2\otimes x_1$ for $p$-torsion module$x_1,x_2 \in I$.)
(6) If b. $0 \longrightarrow T \longrightarrow M \longrightarrow F \longrightarrow 0$ is an exact sequence, with$x \otimes t= t\otimes x$ for $T$ torsion$x\in I$, $M$ symtrivial, and$t\in T$
c. $F$ torsion-free, then$t_1\otimes t_2 = t_2\otimes t_2$ for each prime $p$, either$t_1,t_2, \in T$.
Proof of a: $F$$I$ is either $p$-divisible$0$ or a fractional ideal. The case $T$ has no$I=0$ is trivial, so we may assume that $p$-power-torsion$I$ is a fractional ideal. $I\otimes I$ is $I^2$, another fractional ideal, and the multiplication is clearly commutative.
(7) If Proof of b: $T$$t$ is torsion and symtrivial, so assume $F$ is torsion-free and symtrivial, and$n t =0$ for some $n \in R$. For each prime $p$ dividing $n$, eitherwe know that $F$ is $p$-divisble ordivisible, so $T$ has no$x$ is $p$-power$n$-torsiondivisible up to torsion. In other words, thenwe can write $T \oplus F$ is symtrivial$x=ny+t'$ for $t'$ torsion. Then
If none$$x\otimes t= (ny+t') \otimes t = n (y\otimes t)+(t'\otimes t)= (t'\otimes t)$$
This reduces this case to case c.
Proof of my proofs are mistakenc: Write $T= C+D$ for $C=\sum_p R/p^n$ and $D$ a divisible torsion module. Since $D$ is divisible, the remaining questions aretensor product of $D$ with any torsion module is $0$. Hence $T\otimes T = (C+D)\otimes (C+D)= C\otimes C$. But $C\otimes C = \sum_p R/p^n \otimes R/p^n = \sum_p R/p^n = C$, where the isomorphism is given by the multiplicative structure of $R/p^n$, which is clearly commutative, so $t_1\otimes t_2 = t_2 \otimes t_1$.
(Q1) For $T$ torsion Next we prove that any symtrivial module has this form. Given a symtrivial $M$, we write it as an exact sequence $0 \to T \to M \to F \to 0$, with $T$ torsion and $F$ torsion-free. First we show that $F \subseteq K$. This just follows Todd Trimble's argument from the comments - tensoring with $K$ is symmetric monoidal, so preserves symtriviality, so it takes $M$ to a symtrivial vector space, such that for each primewhich must be $p$$0$ or $1$-dimensional. The image of the map $M \to M \otimes_R K$ is just $F$, eitherso $F \subseteq K$.
Next we must show that $T_p=0$ if $F$ is not $p$-divisble ordivisible and is a sum of a cyclic module and a $T$ has no$p$-divisible $p$-torsionmodule otherwise. To do this we may consider the localization $M_p$, what iswhich fits into an exact sequence $\mathrm{Ext}^1(F,T)$?
(Q2) Do all elements of$0 \to T_p \to F_p \to M_p$. $\mathrm{Ext}^1(F,T)$ give$M_p$ must be symtrivial modules?
Proofs:
(1) This. $M_p \subset R_p$. $F_p \subseteq K_p$, so $F_p=0,R_p$, or $K_p$. The middle case is immediatethe non-$p$-divisible case, and it clearly forces $T_p=0$, since $R_p$ is projective so the exact sequence splits, and we have $(R_p + T_p) \otimes (R_p + T_p) = R_p+T_p+T_p+ T_p\otimes T_p$, with the two $T_p$s switched by Todd Trimble's argumentthe symmetry, so they must both vanish.
(2) Indeed If $F_p=0$, take a module $M$ with torsion submodule$T_p$ is symtrivial. We will demonstrate that if $T$$F_p=K_p$, $T_p$ is symtrivial still. Suppose not. Find an $a,b$ in $T_p$ such that $a \otimes b \neq b \otimes a$$a\otimes b \neq b \otimes a$ in $T\otimes T$$T_p \otimes T_p$, but $a\otimes b = b \otimes a$$a \otimes b = b \otimes a$ in $M\otimes M$$M_p \otimes M_p$. The equalityproof that $a \otimes b = b\otimes a$ must be the consequence of finitely many relations, involvinginvolve finitely many elements of $R_p$,. Consider the submodule $M'$ of $M_p$ generated by the whole torsion module$T_p$ and those finitely many elements. The torsion-free quotient of $M'$ is a finitely generated submodule of $K$$K_p$, thus is a fractional ideal$R_p$ or $0$, thus is projective, so the submodule splits into a direct sum of torsion and torsion-free parts. But than $T\otimes T$$T_p\otimes T_p$ is a direct summand of $M'\otimes M'$, so if $a\otimes b \neq b\otimes a$ in $T\otimes T$$T_p\otimes T_p$, they do not equal each other in $M'\otimes M'$ - but a complete set of relations implying that they do are relations of $M' \otimes M'$, a contradiction.
(3) This So $T_p$ is immediate from things noted in the original question.
(4) Only if follows from the fact that quotients of symtrivial modules are symtrivial. For if, first note that if $M/pM=0$, then $M$We must show it is a sum of a $p$-divisible and $p$-torsion, so its tensor product with itself is trivial. If $M/pM=R/pR$, let $x$ be module and a lift of $1 \in R/pR = M/pM$ to $M$$p$-power cyclic module. For anySince tensoring with $a \otimes b$$R/p$ is a symmetric monodical functor, choose $n \geq 0$ such that$T_p/pT_p$ is symtrivial, so $p^n a = p^n b = p^n x = 0$. Choose$T_p/pT_p = 0$ or $u,v \in \mathbb{Z}$ with$R/p$. In the first case, $a \equiv ux \pmod {p^n}$ and$T_p$ is a $b \equiv vx \pmod {p^n}$. Then$p$-divisible $a \otimes b = ux \otimes b = ux \otimes vx = vx \otimes ux = b \otimes ux = b \otimes a$$p$-torsion module, so we are done.
(5) Let So the $\pi$ be a uniformizer$p$-adic completion of $R_p$. This defines$T_p$ is, by Nakayama's Lemma, generated by a surjectionsingle element. The $p^{n-1}M/p^nM \to p^n M/p^{n+1}M$, so either$p$-adic completion has the form $p^nM/p^{n+1}M$ is eventually$\hat{R}_p/p^n$ for some $0$ or we get$n$, and $n$ must be finite because otherwise the completion morphism would be a nontrivial morphism to a $M \to \lim_n M/p^n M = \lim_n R/p^nR=\hat{R}_p$ which is$\hat{R}_p$, a torsion-free module, so it's eventually $0$which is impossible. Then we have a surjection $M \to M/p^n M= R/p^nR$ whoseThe kernel of $T_p \to \hat{T}_p$ is $p$-divisible because its elements are exactly the elements of $T_p$ that are in $p^n T_p$ for each $n$. Since it is $p$-divisible $p$-torsion groups are divisible, they areit is injective, so the exact sequence splits and we get a direct sum.
(6) Localize the exact sequence at $p$$0 \to \cap_n p^n T_p \to T_p \to R/p^n \to 0$ splits. Then $F_p$(The last map is a submodule of $K_p$, so either $R_p$ or $K_p$. In the second case $F$ issurjection because every map to $p$-divisible so take the first case.$R/p^n$ that projects nontrivially onto $R_p$$R/p$ is projective so the exact sequence splits into a direct sum, so $R_p \otimes T_p = 0$ by one of the notes in the question, but $R_p \otimes T_p=T_p$ so $T_p=0$surjective.)
(7) It suffices to prove that $T \otimes F=0$ Finally, by one of the notes inwe compute the questionExt group. But we can divideAgain write $T$ into$T= C+D$ where $p$-power-torsion modules, and a$C= \sum_p R/p^n$ adn $p$-divisble module tensor$D$ a $p$-power-torsiondivisible torsion module. $Ext^1(F,T)= Ext^1(F,C)+Ext^1(F,D)$, but $D$ is trivialinjective so $Ext^1(F,D)=0$.
Answers to Q1 and Q2we just need to find $Ext^1(F,C)$. We will do so using the following lemma:
Q1: We disambiguateLemma: "strongly locally cyclic" means that its localization at each prime is cyclic. "weakly locally cyclic" means that each finitely generated submodule is cyclic.Let $K_p/R_p$ is in the second class but not the first.
We can write$X$ be a torsion $T$ as$R$-module that is the direct sum of divisible torsion module $D$ and a strongly locally cyclic modulefinite $C$. Divisible$p$-torsion modules are injectivefor different primes $p$, so $Ext^1(F,T)=Ext^1(F,C)$$X = \sum_p X_p$.
$Ext^1(F,C)=\left(\prod_p C_p\right)/C$
Coose any nontrivial homomorphism $R \to F$, then the morphism is injective and the kernel is weakly locally cyclic, say Let $L$. We have$Y$ be a long exact sequence
$Hom(L,C) \to Hom(F,C) \to Hom(R,C) \to Ext^1(L,C) \to Ext^1(F,C) \to Ext^1(R,C)$
$Hom(R,C)=C$, $Ext^1(R,C) = 0$
$Hom(F,C) = 0$ becausetorsion-free $F$$R$-module that is $p$-divisible for alleach $p$ such that $C_p$$X_p$ is nontrivial, so the image of any homomorphism inside $C_p$ is trivial, so the whole homomorphism is trivial. Then
$Ext^1(L,C) = \prod_p Ext^1(L_p,C_p)$ because $L$ and $C$ are torsion. Whenever $C_p$ is nonzero, $L_p=K_p/R_p$ so$$Ext^1(Y,X) = Hom(Y \otimes_R K, \prod_p X_p / \sum_p X_p )$$
Proof: We use the exact sequence $Ext^1(L_p,C_p)=Ext^1(K_p,R_p,R/p^a) = R/p^a=C_p$$0\to \sum_p X_p \to \prod_p X_p \to \prod_p X_p / \sum_p X_p$. So we haveWe obtain a shortlong exact sequence
$0 \to C \to \prod_p C_p \to Ext^1(F,C) \to 0$
If we're interested in classifying extensions abstractly, we just have to mod out by the action of the automorphism group.
Q2: Allrelevant terms are symtrivial$Hom(Y, \prod_p X_p) \to Hom(Y, \prod_p X_p / \sum_p X_p) \to Ext^1(Y,X) \to Ext^(Y,\prod_p X_p)$.
Indeed We can pull out the product, letreplacing the first term with $M$ be any module whose torsion$\prod_p Hom( Y,X_P)$ and non-torsion parts satisfy the given conditions, and choosesecond term with $a \otimes b \in M \otimes M$$Ext^1(Y, X_p)$. ChooseThese modules must be $x \in M$ such that$p$-divisible and must be boundedly $a \in \langle x,t \rangle$$p$-torsion, so they vanish, and hence we get an isomorphism $b\in \langle x,t\rangle$$Hom(Y, \prod_p X_p / \sum_p X_p) =Ext^1(Y,X)$. thenBut $a \otimes b= (k_1 x+ t_1) \otimes (k_2 x + t_2) = k_1k_2 x\otimes x + k_1 x\otimes t_2 + k_2 t_1 \otimes x + t_1 \otimes t_2$. Then$\prod_p X_p/ \sum_p X_p$ is divisible, and torsion-free, hence it is a $t_1 \otimes t_2 = t_2 \otimes t_1$$K$-vector space, so the only thing that remains to check is ifHoms uniquely factor thoruhg $t_1\otimes x=x\otimes t_1$$Y \otimes_R K$. Suppose
In our case setting $n t_1=0$$X=C$, then choose $y$ such that$Y=F$, $ny \equiv x$ mod$Y \otimes_R K$ is a one-dimensional vector space if $T$$F$ is nontrivial, so
$x\otimes t_1=(x-ny)\otimes t_1 =t_1 \otimes (x-ny) = t_1 \otimes x$. the Homs are just $\prod_p R/p^n / \sum_p R/p^n$