improve infinite glue shrinkage log message

base/ltmarks.dtx

@@ -898,6 +898,16 @@
 %  \begin{macro}{@@_prepare_and_extract:nnn}
 %
 %    \begin{macrocode}
+\cs_set_eq:cN{Ignore~infinite~glue~shrinkage~error}\c_max_dim
+%    \end{macrocode}
+%    \begin{macrocode}
+\cs_new_protected:Npx \@@_vbox_set_split_to_maxdimen:NN #1#2
+  { \tex_setbox:D #1 \tex_vsplit:D #2 to \exp_not:N\use:n{
+      \use:c{Ignore~infinite~glue~shrinkage~error}
+     }
+  }
+%    \end{macrocode}
+%    \begin{macrocode}
 \cs_new_protected:Npn \@@_prepare_and_extract:nnn #1#2#3 {
 %    \end{macrocode}
 %
@@ -982,17 +992,17 @@
 %    \TeX{} error. Unfortunately this is something one cannot always
 %    prevent. For example if there is infinite negative glue that is
 %    canceled by infinite positive glue then we can't detect this case.
-%    however the splitting operation will balk, because it goes throw
+%    however the splitting operation will balk, because it goes through
 %    the vertical material item by item and complains the moment it
 %    finds any negative infinite glue, even though it is a harmless
-%    one if the material whole is considered as a whole.
+%    one if the material is considered as a whole.
 %    We therefore, change the interaction mode, so that such errors
 %    appear in the log, but do not stop the processing.
 %    \begin{macrocode}
           \tl_set:Ne \l_@@_interaction_tl
              { \interactionmode \the\interactionmode\relax }
-          \interactionmode 0
-          \vbox_set_split_to_ht:NNn \l_@@_ii_box \l_@@_box \c_max_dim
+          \interactionmode 0\scan_stop:
+          \@@_vbox_set_split_to_maxdimen:NN \l_@@_ii_box \l_@@_box 
           \l_@@_interaction_tl   
 %    \end{macrocode}
 %    After splitting we check if there is anything left in