Bootsrapped

README.md

@@ -5,19 +5,16 @@ This is a rather straightforward port/implementation of the grammar defined for
 
 ## Current status
 
-- This is a **beta** (at least, until bootstrapping is implemented)
 - As far as I can tell, implements everything from the PEG.js grammar
 - Generates complete useable parsers
-- The project has not been bootstrapped yet (that is the parser is not yet 
-  generated from the grammar by the parser itself)
-- For now uses [Neotoma](https://github.com/seancribbs/neotoma) to generate the parser
-- Is based on an earlier definition of the grammar (probably [this](https://github.com/dmajda/pegjs/blob/f0a6bc92cc24b623689c7811bebc1ce2921442f0/src/parser.pegjs))
+- The project is bootstrapped (see `priv/pegjs_parse.pegjs`). Original grammar for 
+  [Neotoma](https://github.com/seancribbs/neotoma) is also available in `priv/pegjs_parse.peg`
+- It's based on an earlier definition of the grammar (probably [this](https://github.com/dmajda/pegjs/blob/f0a6bc92cc24b623689c7811bebc1ce2921442f0/src/parser.pegjs))
   than the one that currently [exists](https://github.com/dmajda/pegjs/blob/master/src/parser.pegjs) 
   for PEG.js. Currenlty there's no timeframe or even a decision on when/whether to update the erlang version
 
 ## Further work
 
-- Bootstrap
 - Support @append (see, e.g. [core-pegjs](https://github.com/for-GET/core-pegjs) 
   in the for-GET project)
 - Add more options:
@@ -59,7 +56,8 @@ There are several options you can pass along to `pegjs:file(File, Options::optio
 -type option()  :: {ignore_unused, boolean()}         %% ignore unused rules. Default: true
                  | {ignore_duplicates, boolean()}     %% ignore duplicate rules. Default: false
                  | {ignore_unparsed, boolean()}       %% ignore incomplete parses. Default: false
-                 | {ignore_missing_rules, boolean()}. %% Default: false
+                 | {ignore_missing_rules, boolean()}, %% Default: false
+                 | {parser, atom()}.                  %% use a different module to parse grammars. Default: pegjs_parse
 
 ```
 
@@ -75,10 +73,45 @@ branch of neotoma.
 
 Non-generated parser combinators can be found in `priv/pegjs.template`.
 
-To generate the `pegjs_parse` module, all you need is:
+Safe working parser is always available at `src/pegjs_parse.erl.safe`.
+
+### pegjs grammar
+
+The current grammar from which the project is now bootstrapped lives in 
+`priv/pegjs_parser.pegjs`. When you've tweaked it and you want to try your changes, 
+generate a different module and tell `pegjs` to use your new module instead:
+
+```erlang
+> pegjs:file("priv/pegjs_parse.pegjs", [{output, "src"}, {module, modified_parser}]).
+ok
+> c(modified_parser).
+{ok, modified_parser)
+> pegjs:file("extra/json.pegjs", [{parser, modified_parser}]).
+ok
+... etc. ...
+```
+
+Once you're satisfied with your changes, overwrite pegjs_parser (which is used by default):
+
+```erlang
+> pegjs:file("priv/pegjs_parse.pegjs", [{output, "src"}]).
+ok
+> c(pegjs_parse).
+{ok, pegjs_parse)
+> pegjs:file("extra/json.pegjs").
+ok
+... etc. ...
+```
+
+### Original neotoma
+
+The original parser for pegjs was derived from a grammar defined for [Neotoma](https://github.com/seancribbs/neotoma).
+You can also start your work from there:
 
 ```erlang
 > neotoma:file("priv/pegjs_parse.peg", [{output, "src/"}]).
 ok
 > pegjs:file(.... etc ... )
 ```
+
+However, the original grammar will get increasingly outdated as time goes on, so it's there for reference only.

priv/pegjs.template

@@ -1,6 +1,8 @@
 %% -type index() :: {{line, pos_integer()}, {column, pos_integer()}}.
 -type input() :: binary().
+-ifndef(index).
 -type index() :: {{line, integer()}, {column, integer()}}.
+-endif.
 -type parse_failure() :: {error, term()}.
 -type parse_success() :: {term(), input(), index()}.
 -type parse_result() :: parse_failure() | parse_success().