Fix typos [ci skip]

l3kernel/l3unicode.dtx

@@ -479,8 +479,8 @@
 %    \end{macrocode}
 % Parse the main Unicode data file and pull out the NFD and case changing
 % data. The NFD data is stored on using the hash table approach and can yield
-% a predictable number of code points: one or two. We also need the case data,
-% which will be modified further below. To allow for finding rangs, the
+% a predictable number of codepoints: one or two. We also need the case data,
+% which will be modified further below. To allow for finding ranges, the
 % description of the codepoint needs to be carried forward.
 %    \begin{macrocode}
   \cs_set_protected:Npn \@@_data_auxi:w
@@ -601,7 +601,7 @@
 %    \end{macrocode}
 %   Here we want to do three things: add to and possibly complete the current
 %   block, add complete blocks quickly, then finish up the range in a final
-%   open block. We need to avoid as far as possible avoid dealing with ever
+%   open block. We need to avoid as far as possible avoid dealing with every
 %   single codepoint, so the middle step is optimised.
 %    \begin{macrocode}
   \cs_set_protected:Npn \@@_range:nnnn #1#2#3#4
@@ -668,7 +668,7 @@
 %    \end{macrocode}
 % Close out the final block, rename the first stage table, then combine all
 % of the block comma-lists into one large second-stage table with offsets.
-% As we use an index not an offet, there is a little back-and-forward to do.
+% As we use an index not an offset, there is a little back-and-forward to do.
 %    \begin{macrocode}
   \cs_set_protected:Npn \@@_finalise_blocks:
     {