boost/parser
2
0
Fork 0
mirror of https://github.com/boostorg/parser.git synced 2026-01-28 07:22:29 +00:00
Code Issues Projects Releases Wiki Activity
751 Commits 12 Branches 7 Tags
16cd1894e2bb81485fd78666735e34bfc26ef4d9
Commit Graph

326 Commits

Author SHA1 Message Date
Zach Laine
16cd1894e2 Work around GCC12 bug involving the use of declval. 2024-01-13 21:26:49 -06:00
Zach Laine
6a2a1229a4 GCC9 warning mitigation. 2024-01-13 21:16:24 -06:00
Zach Laine
1566dffe1f Fix GCC and Clang builds. 2024-01-13 21:11:43 -06:00
Zach Laine
251dcc9436 Implement the merge[] and separate[] directives that control attribute 
generation.  Needs docs.

Fixes #55.
 2024-01-13 20:57:01 -06:00
Zach Laine
cf9545fa7c Remove mooted TODOs. 2024-01-12 22:35:04 -06:00
Zach Laine
3e7812ebe5 Simplify detail::is_variant and remove enable_variant. Turns out 
detail::is_variant_v is not actually used, except in one place in the printing
code, and it's not even essential there.
 2024-01-12 21:45:39 -06:00
Zach Laine
69e4f47bff Work around the busted-ass static_asserting op<< in boost::optional by 
recognizing all enabled optionals, and printing them the same as we do
std::optional.  Applied the same logic to variants too, for consistency.
 2024-01-12 21:45:39 -06:00
Zach Laine
c8ce2cace5 Add a special seq_parser combining rule that makes a std::string ot of two 
adjacent char/char32_t values.

Partially addresses #55.
 2024-01-11 19:24:43 -06:00
Zach Laine
b6ffe5fbe3 Change detail::sequence_of and detail::container_and_value_type so that they 
generate std::strings from repeated char32_ts and combine char32_ts with
adjacent strings.  Update docs accordingly.

Partially addresses #55.
 2024-01-10 22:17:53 -06:00
Zach Laine
257dcd36aa Add a static_assert in seq_parser that indicates the most likely source of 
compilation failure there -- attribute incompatibility.

Partially addresses #56.
 2024-01-08 18:31:54 -06:00
Zach Laine
cd284572c9 Add support for parsing into attributes whose types are optionals or variants 
other than the ones from std:: (e.g. boost::optional, boost::variant, and
boost::variant2::variant).

Fixes #53.
 2024-01-07 20:44:34 -06:00
Zach Laine
4becbe34ad Fix infinite loop when handling an optional of sequence attribute type in 
repeat_parser.
 2024-01-07 19:46:43 -06:00
Zach Laine
745b1b151d Add a small example to the parser test that shows that you do not need Sprit 
2's hold[] directve.  Fix an error exercised by the example.
 2024-01-07 15:11:47 -06:00
Zach Laine
3ebbd9e08c Correct an error in repeat_parser's choice of attribute. Sometimes it is 
necessary to use the value_type of the container out-param, rather than the
attribute generated by the subparser.

Fixes #52.
 2024-01-07 12:29:23 -06:00
Zach Laine
9e374f8cc3 Flesh out support for parsing into non-sequence containers. 2024-01-07 00:44:21 -06:00
Zach Laine
6ab2a9b6a4 Fix the parsing of attributes into nested aggregate types. 
Fixes #50.
 2024-01-07 00:44:21 -06:00
Zach Laine
c7b9eb6eb5 Reimplement get() overloads as one publicly-facing function. 2024-01-06 14:21:07 -06:00
Zach Laine
e6cb3e4159 Add BOOST_PARSER_MAX_AGGREGATE_SIZE, to replace the hardcoded 50 in tuple.hpp. 
Fixes #51.
 2024-01-06 13:22:58 -06:00
Zach Laine
2e4154dcba Cruft removal. 2024-01-06 02:36:26 -06:00
Zach Laine
1e373378bd Assert when an rvalue reference is detected in the char_parser::operator() 
overlaod that takes a range, to prevent dangling references.
 2024-01-06 01:17:32 -06:00
Zach Laine
7a98202845 Address TODO about not using sizeof() to detect char32_t/single-byte input 
value type.
 2024-01-06 01:01:48 -06:00
Zach Laine
f1190f6f59 Grooming. 2024-01-06 01:01:48 -06:00
Zach Laine
2b0053bf83 Add no_case[] directive, and associated changes to char, string, and boolean 
parsers to make it work; uses the recently-added Unicode case-folding
operation.  Needs docs.

Fixes #44.
 2024-01-06 01:01:48 -06:00
Zach Laine
3b5ea8717c Lower the max number of supported struct elements from 100 to 50 so as not to 
choke MSVC.
 2024-01-04 19:26:47 -06:00
Zach Laine
f34f096c66 Add logic to repeat_parser analogous to the logic in seq_parser, to enable 
implicit tuple <-> struct conversions.

Fixes #47.
 2024-01-04 18:38:44 -06:00
Zach Laine
e10cc0619c Refactor detail::move_back() overloads to eliminate code duplication. 2024-01-04 17:58:57 -06:00
Zach Laine
e81035615b Fix MSVC builds -- std::array::begin() is not necessarily a pointer! 2024-01-04 08:53:06 -06:00
Zach Laine
c37274c9a5 Fix broken build. 2024-01-04 02:40:33 -06:00
Zach Laine
d0fe18b61d Add single-code point Unicode case folding function; data and tests are 
generated from CaseFolding.txt.

Partially addressed #44.
 2024-01-04 02:26:36 -06:00
Zach Laine
fcd3a9fc7d Gross hack to fix GCC < 12 builds. 
Fixes #28.
 2024-01-03 23:44:51 -06:00
Zach Laine
d60d33e40b Finish implementing support for parsing into structs as if they were tuples. 
This makes it possible to map from sequence parsers onto structs with no
semantic actions.  Also, add a doc section describing how parsing structs
works (including a couple of short example programs).

Fixes #45.
Partially addresses #28.
 2024-01-03 23:24:37 -06:00
Zach Laine
8bad1f2963 detail::aggregate_to_tuple() -> detail::tie_aggregate(), and change 
detail::tie_aggregate() to return a tuple of references.

Partially addresses #28.
 2024-01-03 02:28:10 -06:00
Zach Laine
0e78ea2d26 Remove the T >> T -> std::vector<T> attribute combining rule from seq_parser. 2024-01-03 02:04:04 -06:00
Zach Laine
0454d4dad0 Fix MSVC build. 2024-01-03 00:55:42 -06:00
Zach Laine
f08255cf94 Use a tuple instead of a pair in seq_parser::{make_temp_result,combine}. Add a 
third tuple to the output of make_temp_result; this tuple is true/false values
indicating whether this parser's attribute was merged with an adjacent
parser's container of attributes of the same type.

Fixes #36.
 2024-01-02 17:58:03 -06:00
Zach Laine
e7daff5b26 Use std::move consistently when assigning/appending values. 2024-01-02 17:58:03 -06:00
Zach Laine
2020e44fda Radically change the way that repeat_parser generates a sequence of 
attributes.  It now produces std::string when Attribute is char, and
std::vector<Attribute> otherwise (except for nope special-casing of course).

Partial fix for #36.
 2024-01-02 17:57:46 -06:00
Zach Laine
d34d4b7e6a Remove the seq_parser C<T> >> C<T> -> C<T> collapsing rule. Add special logic 
for recursive rules to make accumulating recursive parsers like
recursive_string_rule in the parser_rule test easier to write.
 2024-01-01 15:29:45 -06:00
Zach Laine
dced620290 Remove mooted TODO. 2024-01-01 15:29:45 -06:00
Zach Laine
b02e9e32a6 Add preprocessor code from Boost.Describe instead of conditionally using 
Boost.Preprocessor.  Thanks, Peter!

Fixes #32.
 2023-12-30 17:00:23 -06:00
sehe
b7d1d3dba5 Fix spelling of hexadecimal 
The library and documentation misspells it (consistently, which is good
:)) as hexidecimal. That makes it hard for (new) users to find out by
searching documentation.

The fifteen spots fixed:

    example/callback_json.cpp|47 col 15| "four hexidecimal digits";
    example/callback_json.cpp|49 col 18| "\\uXXXX hexidecimal escape sequence";
    example/callback_json.cpp|51 col 38| escape_double_seq = "\\uXXXX hexidecimal escape sequence";
    example/json.cpp|57 col 28| // like "expected four hexidecimal digits here:", instead of "expected
    example/json.cpp|65 col 15| "four hexidecimal digits";
    example/json.cpp|67 col 18| "\\uXXXX hexidecimal escape sequence";
    example/json.cpp|69 col 38| escape_double_seq = "\\uXXXX hexidecimal escape sequence";
    example/json.cpp|158 col 18| // declare a hexidecimal parser that matches exactly 4.
    doc/tutorial.qbk|981 col 18| [ Matches a hexidecimal unsigned integral value. ]
    doc/tutorial.qbk|986 col 28| [ Matches exactly the hexidecimal unsigned integral value `_RES_np_(arg0)`. ]
    doc/tutorial.qbk|1118 col 42| So, if you wanted to parse exactly eight hexidecimal digits in a row in order
    doc/tutorial.qbk|2729 col 18| "\\uXXXX hexidecimal escape sequence";
    doc/tutorial.qbk|2731 col 38| escape_double_seq = "\\uXXXX hexidecimal escape sequence";
    doc/tutorial.qbk|2741 col 9| put a hexidecimal escape sequence there.
    include/boost/parser/parser.hpp|5622 col 13| /** The hexidecimal unsigned integer parser.  Produces an `unsigned int`

The one spot that was inconsistent (outside the gtest subtree):

    include/boost/parser/tuple.hpp|66 col 36| // 0xDEADBEEF (hexadecimal)
 2023-12-30 15:13:04 -06:00
Zach Laine
fcb435df99 Add test for the struct <-> tuple assignment machinery; fix errors. 
Partially addresses #28.
 2023-12-30 04:12:05 -06:00
Zach Laine
95374a7878 Add initial sketch of struct -> tuple assignment machinery. 
Partially addresses #28.
 2023-12-30 02:58:50 -06:00
Zach Laine
669af7fa69 Add initial sketch of tuple -> struct assignment machinery. 
Partially addresses #28.
 2023-12-30 01:40:43 -06:00
Zach Laine
7afc5e5ac3 Remove needless inclusion of <coroutine>. 
Fixes #37.
 2023-12-29 19:55:25 -06:00
Zach Laine
2030e41bc5 Doc typo. 
Fixes #31.
 2023-12-29 14:50:57 -06:00
Zach Laine
868b5854c0 MSVC warning mitigation. 
Fixes #29.
 2023-12-29 14:00:27 -06:00
Zach Laine
b218bd78d7 Print diagnostics to cerr by default instead of cout. 
Fixes #30.
 2023-12-29 13:27:52 -06:00
Zach Laine
88a965d9c5 Rename the {callback_,}parse() overloads that take an iterator/sentinel pair 
to {callback,}prefix_parse().

Fixes #21.
 2023-12-23 04:05:54 -06:00
Zach Laine
fe23476eed Get rid of phony Doxygen-only as_utf*, since they were not showing up in the 
generated docs anyway, and let Doxygen see the real as_utf*s and utf*_views
instead.
 2023-12-23 03:01:08 -06:00