%% =====================================================================
%% This library is free software; you can redistribute it and/or modify
%% it under the terms of the GNU Lesser General Public License as
%% published by the Free Software Foundation; either version 2 of the
%% License, or (at your option) any later version.
%%
%% This library is distributed in the hope that it will be useful, but
%% WITHOUT ANY WARRANTY; without even the implied warranty of
%% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
%% Lesser General Public License for more details.
%%
%% You should have received a copy of the GNU Lesser General Public
%% License along with this library; if not, write to the Free Software
%% Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
%% USA
%%
%% @copyright 2001-2003 Richard Carlsson
%% @author Richard Carlsson <[email protected]>
%% @see edoc
%% @end
%% =====================================================================
%% @doc EDoc documentation extraction.
-module(edoc_extract).
-export([source/3, source/4, source/5, header/3, header/4, header/5,
file/4, text/4]).
-import(edoc_report, [report/3, warning/3]).
%% %% @headerfile "edoc.hrl" (disabled until it can be made private)
-include("edoc.hrl").
%% @type filename() = //kernel/file:filename().
%% @type proplist() = //stdlib/proplists:property().
%% @type syntaxTree() = //syntax_tools/erl_syntax:syntaxTree().
%% @spec source(File::filename(), Env::edoc_env(), Options::proplist())
%% -> {ModuleName, edoc:edoc_module()}
%% ModuleName = atom()
%% proplist() = [term()]
%%
%% @doc Like {@link source/5}, but reads the syntax tree and the
%% comments from the specified file.
%%
%% @see edoc:read_comments/2
%% @see edoc:read_source/2
%% @see source/4
source(File, Env, Opts) ->
Forms = edoc:read_source(File, Opts),
Comments = edoc:read_comments(File, Opts),
source(Forms, Comments, File, Env, Opts).
%% @spec source(Forms, Comments::[edoc:comment()], File::filename(),
%% Env::edoc_env(), Options::proplist()) ->
%% {ModuleName, edoc:edoc_module()}
%%
%% Forms = syntaxTree() | [syntaxTree()]
%% ModuleName = atom()
%%
%% @doc Like {@link source/4}, but first inserts the given comments in
%% the syntax trees. The syntax trees must contain valid position
%% information. (Cf. {@link edoc:read_comments/2}.)
%%
%% @see edoc:read_comments/2
%% @see edoc:read_source/2
%% @see source/3
%% @see source/4
%% @see //syntax_tools/erl_recomment
source(Forms, Comments, File, Env, Opts) when is_list(Forms) ->
Forms1 = erl_syntax:form_list(Forms),
source(Forms1, Comments, File, Env, Opts);
source(Forms, Comments, File, Env, Opts) ->
Tree = erl_recomment:quick_recomment_forms(Forms, Comments),
TypeDocs = find_type_docs(Forms, Comments, Env, File),
source1(Tree, File, Env, Opts, TypeDocs).
%% @spec source(Forms, File::filename(), Env::edoc_env(),
%% Options::proplist()) ->
%% {ModuleName, edoc:edoc_module()}
%%
%% Forms = syntaxTree() | [syntaxTree()]
%% ModuleName = atom()
%% @type edoc_env() = edoc_lib:edoc_env()
%%
%% @doc Extracts EDoc documentation from commented source code syntax
%% trees. The given `Forms' must be a single syntax tree of
%% type `form_list', or a list of syntax trees representing
%% "program forms" (cf. {@link edoc:read_source/2}.
%% `Env' is an environment created by {@link
%% edoc_lib:get_doc_env/3}. The `File' argument is used for
%% error reporting and output file name generation only.
%%
%% See {@link edoc:get_doc/2} for descriptions of the `def',
%% `hidden', `private', and `todo' options.
%%
%% @see edoc:read_comments/2
%% @see edoc:read_source/2
%% @see source/5
%% @see //syntax_tools/erl_recomment
%% Note that the actual module name found in the source file will be
%% used for generating the documentation, creating relative links, etc.
%% INHERIT-OPTIONS: add_macro_defs/3
%% INHERIT-OPTIONS: edoc_data:module/4
source(Forms, File, Env, Opts) when is_list(Forms) ->
source(erl_syntax:form_list(Forms), File, Env, Opts);
source(Tree, File0, Env, Opts) ->
TypeDocs = find_type_docs(Tree, [], Env, File0),
source1(Tree, File0, Env, Opts, TypeDocs).
%% Forms0 and Comments is used for extracting Erlang type documentation.
source1(Tree, File0, Env, Opts, TypeDocs) ->
Forms = preprocess_forms(Tree),
File = edoc_lib:filename(File0),
Module = get_module_info(Tree, File),
{Header, Footer, Entries} = collect(Forms, Module),
Name = Module#module.name,
Env1 = Env#env{module = Name,
root = ""},
Env2 = add_macro_defs(module_macros(Env1), Opts, Env1),
Entries1 = get_tags([Header, Footer | Entries], Env2, File, TypeDocs),
Entries2 = edoc_specs:add_data(Entries1, Opts, File, Module),
edoc_tags:check_types(Entries2, Opts, File),
Data = edoc_data:module(Module, Entries2, Env2, Opts),
{Name, Data}.
%% @spec header(File::filename(), Env::edoc_env(), Options::proplist())
%% -> {ok, Tags} | {error, Reason}
%% Tags = [term()]
%% Reason = term()
%%
%% @doc Similar to {@link header/5}, but reads the syntax tree and the
%% comments from the specified file.
%%
%% @see edoc:read_comments/2
%% @see edoc:read_source/2
%% @see header/4
header(File, Env, Opts) ->
Forms = edoc:read_source(File),
Comments = edoc:read_comments(File),
header(Forms, Comments, File, Env, Opts).
%% @spec header(Forms, Comments::[edoc:comment()], File::filename(),
%% Env::edoc_env(), Options::proplist()) ->
%% {ok, Tags} | {error, Reason}
%% Forms = syntaxTree() | [syntaxTree()]
%% Tags = [term()]
%% Reason = term()
%%
%% @doc Similar to {@link header/4}, but first inserts the given
%% comments in the syntax trees. The syntax trees must contain valid
%% position information. (Cf. {@link edoc:read_comments/2}.)
%%
%% @see header/3
%% @see header/4
%% @see //syntax_tools/erl_recomment
header(Forms, Comments, File, Env, Opts) when is_list(Forms) ->
Forms1 = erl_syntax:form_list(Forms),
header(Forms1, Comments, File, Env, Opts);
header(Forms, Comments, File, Env, Opts) ->
Tree = erl_recomment:quick_recomment_forms(Forms, Comments),
header(Tree, File, Env, Opts).
%% @spec header(Forms, File::filename(), Env::edoc_env(),
%% Options::proplist()) ->
%% {ok, Tags} | {error, Reason}
%% Forms = syntaxTree() | [syntaxTree()]
%% Tags = [term()]
%% Reason = term()
%%
%% @doc Extracts EDoc documentation from commented header file syntax
%% trees. Similar to {@link source/5}, but ignores any documentation
%% that occurs before a module declaration or a function definition.
%% (Warning messages are printed if content may be ignored.) `Env' is
%% assumed to already be set up with a suitable module context.
%%
%% @see header/5
%% @see //syntax_tools/erl_recomment
header(Forms, File, Env, Opts) when is_list(Forms) ->
header(erl_syntax:form_list(Forms), File, Env, Opts);
header(Tree, File0, Env, _Opts) ->
Forms = preprocess_forms(Tree),
File = edoc_lib:filename(File0),
Module = #module{name = Env#env.module}, % a dummy module record
%% We take only "footer" tags, i.e., any kind of definition will
%% kill all the information above it up to that point. Then we call
%% this the 'header' to make error reports make better sense.
{Header, Footer, Entries} = collect(Forms, Module),
if Header#entry.data /= {[],[],[]} ->
warning(File, "documentation before module declaration is ignored by @headerfile", []);
true -> ok
end,
if Entries /= [] ->
warning(File, "documentation before function definitions is ignored by @headerfile", []);
true -> ok
end,
[Entry] = get_tags([Footer#entry{name = header}], Env, File),
Entry#entry.data.
%% NEW-OPTIONS: def
%% DEFER-OPTIONS: source/4
add_macro_defs(Defs0, Opts, Env) ->
Defs = proplists:append_values(def, Opts),
edoc_macros:check_defs(Defs),
Env#env{macros = Defs ++ Defs0 ++ Env#env.macros}.
%% @spec file(File::filename(), Context, Env::edoc_env(),
%% Options::proplist()) -> {ok, Tags} | {error, Reason}
%% Context = overview
%% Tags = [term()]
%% Reason = term()
%%
%% @doc Reads a text file and returns the list of tags in the file. Any
%% lines of text before the first tag are ignored. `Env' is an
%% environment created by {@link edoc_lib:get_doc_env/3}. Upon error,
%% `Reason' is an atom returned from the call to {@link
%% //kernel/file:read_file/1} or the atom 'invalid_unicode'.
%%
%% See {@link text/4} for options.
%% INHERIT-OPTIONS: text/4
file(File, Context, Env, Opts) ->
case file:read_file(File) of
{ok, Bin} ->
Enc = edoc_lib:read_encoding(File,[{in_comment_only, false}]),
case catch unicode:characters_to_list(Bin, Enc) of
String when is_list(String) ->
{ok, text(String, Context, Env, Opts, File)};
_ ->
{error, invalid_unicode}
end;
{error, _} = Error ->
Error
end.
%% @spec (Text::string(), Context, Env::edoc_env(),
%% Options::proplist()) -> Tags
%% Context = overview
%% Tags = [term()]
%%
%% @doc Returns the list of tags in the text. Any lines of text before
%% the first tag are ignored. `Env' is an environment created by {@link
%% edoc_lib:get_doc_env/3}.
%%
%% See {@link source/4} for a description of the `def' option.
%% INHERIT-OPTIONS: add_macro_defs/3
%% DEFER-OPTIONS: source/4
text(Text, Context, Env, Opts) ->
text(Text, Context, Env, Opts, "").
text(Text, Context, Env, Opts, Where) ->
Env1 = add_macro_defs(file_macros(Context, Env), Opts, Env),
Cs = edoc_lib:lines(Text),
Ts0 = edoc_tags:scan_lines(Cs, 1),
Tags = sets:from_list(edoc_tags:tag_names()),
Ts1 = edoc_tags:filter_tags(Ts0, Tags, Where),
Single = sets:from_list(edoc_tags:tags(single)),
Allow = sets:from_list(edoc_tags:tags(Context)),
case edoc_tags:check_tags(Ts1, Allow, Single, Where) of
true ->
exit(error);
false ->
Ts2 = edoc_macros:expand_tags(Ts1, Env1, Where),
How = dict:from_list(edoc_tags:tag_parsers()),
edoc_tags:parse_tags(Ts2, How, Env1, Where)
end.
%% @spec (Forms::[syntaxTree()], File::filename()) -> module()
%% @doc Initialises a module-info record with data about the module
%% represented by the list of forms. Exports are guaranteed to exist in
%% the set of defined names.
get_module_info(Forms, File) ->
L = case catch {ok, erl_syntax_lib:analyze_forms(Forms)} of
{ok, L1} ->
L1;
syntax_error ->
report(File, "syntax error in input.", []),
exit(error);
{'EXIT', R} ->
exit(R);
R ->
throw(R)
end,
{Name, Vars} = case lists:keyfind(module, 1, L) of
{module, N} when is_atom(N) ->
{N, none};
{module, {N, _}=Mod} when is_atom(N) ->
Mod;
_ ->
report(File, "module name missing.", []),
exit(error)
end,
Functions = ordsets:from_list(get_list_keyval(functions, L)),
Exports = ordsets:from_list(get_list_keyval(exports, L)),
Attributes = ordsets:from_list(get_list_keyval(attributes, L)),
Records = get_list_keyval(records, L),
Encoding = edoc_lib:read_encoding(File, []),
#module{name = Name,
parameters = Vars,
functions = Functions,
exports = ordsets:intersection(Exports, Functions),
attributes = Attributes,
records = Records,
encoding = Encoding}.
get_list_keyval(Key, L) ->
case lists:keyfind(Key, 1, L) of
{Key, As} ->
ordsets:from_list(As);
_ ->
[]
end.
%% @spec (Forms::[syntaxTree()]) -> [syntaxTree()]
%% @doc Preprocessing: copies any precomments on forms to standalone
%% comments, and removes "invisible" forms from the list.
preprocess_forms(Tree) ->
preprocess_forms_1(erl_syntax:form_list_elements(
erl_syntax:flatten_form_list(Tree))).
preprocess_forms_1([F | Fs]) ->
case erl_syntax:get_precomments(F) of
[] ->
preprocess_forms_2(F, Fs);
Cs ->
Cs ++ preprocess_forms_2(F, Fs)
end;
preprocess_forms_1([]) ->
[].
preprocess_forms_2(F, Fs) ->
case erl_syntax_lib:analyze_form(F) of
comment ->
[F | preprocess_forms_1(Fs)];
{function, _} ->
[F | preprocess_forms_1(Fs)];
{attribute, {module, _}} ->
[F | preprocess_forms_1(Fs)];
text ->
[F | preprocess_forms_1(Fs)];
{attribute, {record, _}} ->
[F | preprocess_forms_1(Fs)];
{attribute, {N, _}} ->
case edoc_specs:is_tag(N) of
true ->
[F | preprocess_forms_1(Fs)];
false ->
preprocess_forms_1(Fs)
end;
_ ->
preprocess_forms_1(Fs)
end.
%% This collects the data for the header and the functions of the
%% module. Note that the list of forms is assumed to have been
%% preprocessed first, so that all "invisible" forms are removed, and
%% the only interesting comments are those that are standalone comments
%% in the list.
collect(Fs, Mod) ->
collect(Fs, [], [], [], [], [], undefined, Mod).
collect([F | Fs], Cs, Ss, Ts, Rs, As, Header, Mod) ->
case erl_syntax_lib:analyze_form(F) of
comment ->
collect(Fs, [F | Cs], Ss, Ts, Rs, As, Header, Mod);
{function, Name} ->
L = erl_syntax:get_pos(F),
Export = ordsets:is_element(Name, Mod#module.exports),
Args = parameters(erl_syntax:function_clauses(F)),
collect(Fs, [], [], [], [],
[#entry{name = Name, args = Args, line = L,
export = Export,
data = {comment_text(Cs),Ss,Ts,Rs}} | As],
Header, Mod);
{attribute, {module, _}} when Header =:= undefined ->
L = erl_syntax:get_pos(F),
collect(Fs, [], [], [], [], As,
#entry{name = module, line = L,
data = {comment_text(Cs),Ss,Ts,Rs}},
Mod);
{attribute, {record, {_Name, Fields}}} ->
case is_typed_record(Fields) of
true ->
collect(Fs, Cs, Ss, Ts, [F | Rs], As, Header, Mod);
false ->
collect(Fs, Cs, Ss, Ts, Rs, As, Header, Mod)
end;
{attribute, {N, _}} ->
case edoc_specs:tag(N) of
spec ->
collect(Fs, Cs, [F | Ss], Ts, Rs, As, Header, Mod);
type ->
collect(Fs, Cs, Ss, [F | Ts], Rs, As, Header, Mod);
unknown ->
%% Drop current seen comments.
collect(Fs, [], [], [], Rs, As, Header, Mod)
end;
_ ->
%% Drop current seen comments.
collect(Fs, [], [], [], [], As, Header, Mod)
end;
collect([], Cs, Ss, Ts, Rs, As, Header, _Mod) ->
Footer = #entry{name = footer, data = {comment_text(Cs),Ss,Ts,Rs}},
As1 = lists:reverse(As),
if Header =:= undefined ->
{#entry{name = module, data = {[],[],[],[]}}, Footer, As1};
true ->
{Header, Footer, As1}
end.
is_typed_record([]) ->
false;
is_typed_record([{_, {_, Type}} | Fs]) ->
Type =/= none orelse is_typed_record(Fs).
%% Returns a list of simplified comment information (position and text)
%% for a list of abstract comments. The order of elements is reversed.
comment_text(Cs) ->
comment_text(Cs, []).
comment_text([C | Cs], Ss) ->
L = erl_syntax:get_pos(C),
comment_text(Cs, [#comment{line = L,
text = [remove_percent_chars(S)
|| S <- erl_syntax:comment_text(C)]}
| Ss]);
comment_text([], Ss) ->
Ss.
%% @spec (string()) -> string()
%%
%% @doc Replaces leading `%' characters by spaces. For example, `"%%%
%% foo" -> "\s\s\s foo"', but `"% % foo" -> "\s % foo"', since the
%% second `%' is preceded by whitespace.
remove_percent_chars([$% | Cs]) -> [$\s | remove_percent_chars(Cs)];
remove_percent_chars(Cs) -> Cs.
%% Extracting possible parameter names from Erlang clause patterns. The
%% atom '_' is used when no name can be found. (Better names are made up
%% later, when we also may have typespecs available; see edoc_data.)
parameters(Clauses) ->
select_names([find_names(Ps) || Ps <- patterns(Clauses)]).
patterns(Cs) ->
edoc_lib:transpose([erl_syntax:clause_patterns(C) || C <- Cs]).
find_names(Ps) ->
find_names(Ps, []).
find_names([P | Ps], Ns) ->
case erl_syntax:type(P) of
variable ->
find_names(Ps, [tidy_name(erl_syntax:variable_name(P)) | Ns]);
match_expr ->
%% Right-hand side gets priority over left-hand side!
%% Note that the list is reversed afterwards.
P1 = erl_syntax:match_expr_pattern(P),
P2 = erl_syntax:match_expr_body(P),
find_names([P1, P2 | Ps], Ns);
list ->
P1 = erl_syntax:list_tail(P),
find_names([P1 | Ps], Ns);
record_expr ->
A = erl_syntax:record_expr_type(P),
N = list_to_atom(capitalize(erl_syntax:atom_name(A))),
find_names(Ps, [N | Ns]);
infix_expr ->
%% this can only be a '++' operation
P1 = erl_syntax:infix_expr_right(P),
find_names([P1 | Ps], Ns);
_ ->
find_names(Ps, Ns)
end;
find_names([], Ns) ->
lists:reverse(Ns).
select_names(Ls) ->
select_names(Ls, [], sets:new()).
select_names([Ns | Ls], As, S) ->
A = select_name(Ns, S),
select_names(Ls, [A | As], sets:add_element(A, S));
select_names([], As, _) ->
lists:reverse(As).
select_name([A | Ns], S) ->
case sets:is_element(A, S) of
true ->
select_name(Ns, S);
false ->
A
end;
select_name([], _S) ->
'_'.
%% Strip leading underscore characters from parameter names. If the
%% result does not begin with an uppercase character, we add a single
%% leading underscore. If the result would be empty, the atom '_' is
%% returned.
tidy_name(A) ->
case atom_to_list(A) of
[$_ | Cs] ->
list_to_atom(tidy_name_1(Cs));
_ ->
A
end.
tidy_name_1([$_ | Cs]) -> tidy_name_1(Cs);
tidy_name_1([C | _]=Cs) when C >= $A, C =< $Z -> Cs;
tidy_name_1([C | _]=Cs) when C >= $\300, C =< $\336, C =/= $\327-> Cs;
tidy_name_1(Cs) -> [$_ | Cs].
%% Change initial character from lowercase to uppercase.
capitalize([C | Cs]) when C >= $a, C =< $z -> [C - 32 | Cs];
capitalize(Cs) -> Cs.
%% Collects the tags belonging to each entry, checks them, expands
%% macros and parses the content.
%% %This is commented out until it can be made private
%% %@type tags() = #tags{names = set(atom()),
%% % single = set(atom()),
%% % module = set(atom()),
%% % footer = set(atom()),
%% % function = set(atom())}
%% % set(T) = sets:set(T)
-record(tags, {names,single,module,function,footer}).
get_tags(Es, Env, File) ->
get_tags(Es, Env, File, dict:new()).
get_tags(Es, Env, File, TypeDocs) ->
%% Cache this stuff for quick lookups.
Tags = #tags{names = sets:from_list(edoc_tags:tag_names()),
single = sets:from_list(edoc_tags:tags(single)),
module = sets:from_list(edoc_tags:tags(module)),
footer = sets:from_list(edoc_tags:tags(footer)),
function = sets:from_list(edoc_tags:tags(function))},
How = dict:from_list(edoc_tags:tag_parsers()),
get_tags(Es, Tags, Env, How, File, TypeDocs).
get_tags([#entry{name = Name, data = {Cs,Specs,Types,Records}} = E | Es],
Tags, Env, How, File, TypeDocs) ->
Where = {File, Name},
Ts0 = scan_tags(Cs),
{Ts1,Specs1} = select_spec(Ts0, Where, Specs),
Ts2 = check_tags(Ts1, Tags, Where),
Ts3 = edoc_macros:expand_tags(Ts2, Env, Where),
Ts4 = edoc_tags:parse_tags(Ts3, How, Env, Where),
Ts = selected_specs(Specs1, Ts4),
ETypes = [edoc_specs:type(Type, TypeDocs) || Type <- Types ++ Records],
[E#entry{data = Ts++ETypes} | get_tags(Es, Tags, Env, How, File, TypeDocs)];
get_tags([], _, _, _, _, _) ->
[].
%% Scanning a list of separate comments for tags.
scan_tags([#comment{line = L, text = Ss} | Es]) ->
edoc_tags:scan_lines(Ss, L) ++ scan_tags(Es);
scan_tags([]) ->
[].
%% Check the set of found tags (depending on context).
%% Completely unknown tags are filtered out with a warning.
check_tags(Ts0, Tags, Where) ->
Ts = edoc_tags:filter_tags(Ts0, Tags#tags.names, Where),
case check_tags_1(Ts, Tags, Where) of
false -> Ts;
true -> exit(error)
end.
check_tags_1(Ts, Tags, {_, module} = Where) ->
Allow = Tags#tags.module,
Single = Tags#tags.single,
edoc_tags:check_tags(Ts, Allow, Single, Where);
check_tags_1(Ts, Tags, {_, footer} = Where) ->
Allow = Tags#tags.footer,
Single = Tags#tags.single,
edoc_tags:check_tags(Ts, Allow, Single, Where);
check_tags_1(Ts, Tags, Where) ->
Allow = Tags#tags.function,
Single = Tags#tags.single,
edoc_tags:check_tags(Ts, Allow, Single, Where).
select_spec(Ts, {_, {_F, _A}}, Specs) ->
case edoc_tags:filter_tags(Ts, sets:from_list([spec])) of
[] ->
%% Just a dummy to get us through check_tags()
{[edoc_specs:dummy_spec(S) || S <- Specs] ++ Ts, Specs};
_ ->
{Ts,[]}
end;
select_spec(Ts, _Where, _Specs) ->
{Ts,[]}.
selected_specs([], Ts) ->
Ts;
selected_specs([F], [_ | Ts]) ->
[edoc_specs:spec(F, _Clause=1) | Ts].
%% Macros for modules
module_macros(Env) ->
[{module, atom_to_list(Env#env.module)}]
++ edoc_macros:std_macros(Env).
%% Macros for reading auxiliary edoc-files
file_macros(_Context, Env) ->
edoc_macros:std_macros(Env).
%% @doc Extracts what will be documentation of Erlang types.
%% Returns a dict of {Name, Doc} where Name is {TypeName, Arity}.
%%
%% The idea is to mimic how the @type tag works.
%% Using @type:
%%```@type t() = t1(). Some docs of t/0;
%% Further docs of t/0.'''
%% The same thing using -type:
%%```-type t() :: t1(). % Some docs of t/0;
%% Further docs of t/0.'''
find_type_docs(Forms0, Comments, Env, File) ->
Tree = erl_recomment:recomment_forms(Forms0, Comments),
Forms = preprocess_forms(Tree),
Env1 = add_macro_defs(edoc_macros:std_macros(Env), [], Env),
F = fun(C, Line) -> find_fun(C, Line, Env1, File) end,
edoc_specs:docs(Forms, F).
find_fun(C0, Line, Env, File) ->
C1 = comment_text(C0),
Text = lists:append([C#comment.text || C <- C1]),
Comm = #comment{line = Line, text = Text},
[Tag | _] = scan_tags([Comm]),
[Tag1] = edoc_macros:expand_tags([Tag], Env, File),
Tag1.
