doc

docs/source/libraries.rst

@@ -435,7 +435,7 @@ Decompressors are available for RLE, TSCrunch and ZX0 (Salvador).
         The TSCrunch in-place format is a bit different than regular memory decompression.
         It works with PRG files (so with a 2 byte load-address header) for both the *source* and *compressed* data files.
         So if you want to compress and decompress a block of data from $a000-$c000 your source file has to start with
-        the bytes $00 $0a, then followed by the 8192 data byes, for a total of 8194 bytes.
+        the bytes $00 $0a, then followed by the 8192 data bytes, for a total of 8194 bytes.
         Then you need to call the compressor program with the '-i' argument to tell it to create an in-place compressed data file.
         The data file will *not* be loaded at $a000 but have its own load address closer to the end of the memory buffer.
         If all is well, you can then load and decompress it like so::

docs/source/todo.rst

@@ -40,7 +40,7 @@ Future Things and Ideas
 - Make all constants long by default? or not? (remove type name altogether), reduce to target type implictly if the actual value fits.  Experiment is in branch 'long-consts'
   This will break some existing programs that depend on value wrap arounds, but gives more intuitive constant number handling.
   Can give descriptive error message for old syntax that still includes the type name?
-- improve ANTLR grammar with better error handling
+- improve ANTLR grammar with better error handling (see grammar-changes.md)
 - add documentation for more library modules instead of just linking to the source code
 - add an Index to the documentation
 - can we just use index numbers instead of object hashes for struct instance labels (and other places?)  see labelnameForStructInstance