+\newcommand{\patchisin}{\Sqsubset}
+
+\newcommand{\set}[1]{\mathbb #1}
+\newcommand{\pa}[1]{\varmathbb #1}
+\newcommand{\pay}[1]{\pa{#1}^+}
+\newcommand{\pan}[1]{\pa{#1}^-}
+
+\newcommand{\p}{\pa{P}}
+\newcommand{\py}{\pay{P}}
+\newcommand{\pn}{\pan{P}}
+
+%\newcommand{\hasparents}{\underaccent{1}{>}}
+%\newcommand{\hasparents}{{%
+% \declareslashed{}{_{_1}}{0}{-0.8}{>}\slashed{>}}}
+\newcommand{\hasparents}{>_{\mkern-7.0mu _1}}
+\newcommand{\areparents}{<_{\mkern-14.0mu _1\mkern+5.0mu}}
+
+\renewcommand{\implies}{\Rightarrow}
+
+\newcommand{\pancs}[2]{{\mathcal A} ( #1 , #2 ) }
+\newcommand{\pends}[2]{{\mathcal E} ( #1 , #2 ) }
+
+\begin{document}
+
+\section{Notation}
+
+$ C \hasparents \set X $ The parents of commit $C$ are exactly the set
+$\set X$.
+
+$ C \ge D $ $C$ is a descendant of $D$ in the git commit
+graph. This is a partial order, namely the transitive closure of
+$ D \in \set X $ where $ C \hasparents \set X $.
+
+$ C \has D $ Informally, the tree at commit $C$ contains the change
+made in commit $D$. Does not take account of deliberate reversions by
+the user or in non-Topbloke-controlled branches; these are considered
+normal, forward, commits. For merges and Topbloke-generated
+anticommits, the ``change made'' is only to be thought of as any
+conflict resolution. This is not a partial order because it is not
+transitive.
+
+\section{Invariants}
+
+No replay: \[ C \has D \implies C \ge D \]
+
+Unique base: \[ \mathop{\forall}_{C \in \py} \pends{C}{\pn} = \{ B \} \]
+
+\section{Test more symbols}
+
+$ C \haspatch \p $
+
+$ C \nothaspatch \p $
+
+$ \{ B \} \areparents C $
+