api2021 part I #4

tomasz.sowa · 2021-05-27T10:36:57+02:00

tomasz.sowa commented

2021-05-27 10:36:57 +02:00

merge some work from api2021 branch:

new Space struct (full json support)
PascalCase to snake_case in some places
CSVParser
MainOptionsParser
some tests

merge some work from api2021 branch: - new Space struct (full json support) - PascalCase to snake_case in some places - CSVParser - MainOptionsParser - some tests

tomasz.sowa added 67 commits 2021-05-27 10:36:59 +02:00

f771f12027 renamed: std::wstring Toa(...) -> std::wstring to_str(...)

added: std::string to_str(...)

df88d4c0fc temporarily blocked for compilation: mainspaceparser, jsontospaceparser, spaceparser, spacetojson

effe9be0a3 added: typedef with char_type to TextStreamBase

fac3a7eb71 reorganization in utf8

- utf8 auxiliary functions moved to utf8_private.h file
- in utf8.h are shown only functions available for consumers
- template functions has been moved to utf8_template.h (in utf8.h are only declarations)
  utf8_template.h is included at the end of utf8.h
- functions which take std::ostream changed to template (the stream is a template argument now)

7ccd435e2c make depend

f65d934e8c start working on a new version of Space struct - better support for JSON format

now we have a correct model, some methods for setting/getting values and serialization to json
(no serialization to Space yet)

4ef8f5a884 template<size_t stack_size, size_t heap_block_size> changed to template<typename StreamType>

bc9e4a3844 use PT::WideStreamToUTF8() instead of PT::WideToUTF8() for a moment

ba7fa1c195 fixed: in serialize_json_double(): the buffer length was calculated incorrectly: sizeof(char) changed to sizeof(wchar_t)

added to Space struct:
 void set_null();
 void set_empty_string();
 void set_empty_wstring();
 void set_empty_table();
 void set_empty_object();
 void clear();

0c0f15ab8a changed: JSONToSpaceParser can parse json string/file to the new Space format now

(some minor chars escaping not ready yet)

61291cf7ad make depend

5b5a1dfbb6 changed the way how the library is built

- now there is only one Makefile in the root directory and we have only one pikotools.a lib file
- removed: mainparser

31f7bdb857 added to Space: child_spaces and a name of a Space

6e169f7650 added to JSONToSpaceParser: possibility to parse the Space format

renamed: ParseFile() -> ParseJSONFile()
added: ParseSpaceFile(), ParseJSON(), ParseSpace()

961a02ab39 removed: old SpaceParser

ba82ac7cbe renamed: JSONToSpaceParser -> SpaceParser

1baf5042fe removed: SpaceToJson

db5e516564 added to Space: serialization do Space format

4119461f8e fixed: when reading a non-escaped space token we should check whether a delimiter was escaped or not

also a '#' was not tested

679b9e6173 added some indentation when printing 'pretty Space format'

539faa4976 changed the indentation algorithm when pretty printing the space format (no visual changes)

1c643a08b0 added a test whether we have reach the end of the input string

without this such an input json (incorrect json string as there is no {} characters):
"key": "value"
would be parsed correctly: "key" would be parsed and the rest would be skipped

0aeac12fa0 make depend

a2339eed34 fixed: in Space: pointers 'name' and 'child_spaces' were not correctly initialized in cctors

added in Space:
- some methods for adding values to an object, such as:
  Space & Space::add(const std::wstring & field, bool val) (bool, short, int, long, long long etc.)
- methods for creating lists:
  void Space::to_list(std::list<std::string> & output_list, bool clear_list) const
  bool Space::to_list(const wchar_t * field, std::list<std::string> & output_list, bool clear_list) const
- methods for converting a value from an object field:
  bool Space::to_bool(const wchar_t * field, bool default_value) const
- methods for testing strings:
  bool Space::is_equal(const char * val) const
  bool Space::is_equal(const std::string & val) const
  bool Space::is_equal(const wchar_t * val) const
  bool Space::is_equal(const std::wstring & val) const
- methods to get the raw pointer to a value from an object, such as:
  bool * Space::get_bool(const wchar_t * field)
  float * Space::get_float(const wchar_t * field)
- methods for finding a child space (used in Space format only)
  Space * Space::find_child_space(const wchar_t * name)
  Space & Space::find_add_child_space(const wchar_t * name)

a1e8f13f46 fixed: calculating buffer length in Space: sizeof(char) -> sizeof(wchar_t)

5d34db5fcf added to Space:

- methods for testing a string value if a Space is a string or a table:
  bool has_value(const char * val) const;
  bool has_value(const std::string & val) const;
  bool has_value(const wchar_t * val) const;
  bool has_value(const std::wstring & val) const;
- methods for testing a string value in an object (testing a string or a table):
  bool has_value(const wchar_t * field, const char * val) const;
  bool has_value(const wchar_t * field, const std::string & val) const;
  bool has_value(const wchar_t * field, const wchar_t * val) const;
  bool has_value(const wchar_t * field, const std::wstring & val) const;
- methods for removing a child space:
  void remove_child_space(const wchar_t * name);
  void remove_child_space(const std::wstring & name);
  void remove_child_space(size_t index);

297940ff7c fix in SpaceParser when parsing json format: floating point values such as 1.123e+01 where not correctly parsed ('+' character was not correctly parsed)

added to Space: some methods for searching a field with ignoring case:
Space * get_object_field_nc(const wchar_t * field)
Space * get_object_field_nc(const std::wstring & field);
const Space * get_object_field_nc(const wchar_t * field) const;
const Space * get_object_field_nc(const std::wstring & field) const;

16f6bdb358 fixed: Space::has_value(...) incorrectly compared strings

f572250572 make depend

96eedd9be9 added support for morm::Model to Log:

Log & operator<<(morm::Model & model);
but we need some kind of a macro to allow this

463cec3283 fixed #2 : Procedures for reading an utf8 string incorrectly read some utf-8 characters.

Those characters were treated as invalid characters.

UTF8ToInt_FirstOctet incorrectly checked if the first octed is zero (after removing first bits).
This is a case only if the utf-8 character consists of two bytes. For 3 or 4 bytes
the first part can have all bits equal zero.

b055c46ae8 fixed #3 : CompareNoCase incorrectly returned that string1 is greater than string2 for some characters

when converting from 'char' to 'int' we should first convert to 'unsigned char' and then to 'int'

127f26884e make depend

3984c29fbf moved all directories to src subdirectory

7abe4b340a changes in convert/text functions

- changed function names: PascalCase to snake_case
- templates functions moved to a seperate file (text_private.h)
- as a public api only available functions with char/wchar_t/std::string/std::wstring
- ToLower(...) changed to to_lower_emplace(...), similar ToUpper(...) to to_upper_emplace(...)
- added functions:
  std::string to_lower(const std::string & str);
  std::string to_upper(const std::string & str);
  and with std::wstring too
- functions with postfix 'NoCase' changed to 'nc'

adee7d134f added macro PT_HAS_MORM in Log class where operator<<(morm::Model & model) is used

b3cd4d5f7f removed definition of CXX and CXXFLAGS from Makefile

cd9e501f48 make depend

ce81670bb6 added 'tests' directory with tests for the pikotools library

currently only tests for convert/text functions

7bdc4d2fc7 added global Makefile

d3dd74b3d4 added tests/tests to .gitignore

1260b303bc added ./m to .gitignore

3d6c4e27c0 added .editorconfig file with tab style and tab size information

https://editorconfig.org/

77d7bb5e64 fixed in Space: set_empty_string(), set_empty_wstring(), set_empty_table() and set_empty_object() didn't clear its object

if the same kind of object already existed

ac691bccb7 updated MainSpaceParser to the new Space format, changed api to snake case

now we can:
- parse short options, those beginning with a hypnen '-'
- parse long options, those beginning with two hyphens '--'
- long options can have arguments in two forms:
  - either with an equal sign, e.g.: --opion-name=argument
  - or with a space, e.g: --option argument
    in the latter case we can have more than one argument, e.g: --option argument1 argument2
- parse non-option arguments, those after two hyphens to the end of a string, e.g: -- arg1 arg2

da6a36a205 start creating tests for MainSpaceParser

fe82f63efb changed the way of building in Makefiles

91300bb245 make depend

a5c8833452 added tests for MainSpaceParser

96e60c526f moved files: mainspaceparser/mainspaceparser.(h|cpp) -> mainoptions/mainoptionsparser.(h|cpp)

ad4e8078ae MainSpaceParser class has been renamed to MainOptionsParser

db93586c0e make depend

e0565813de removed *.a from .gitignore

ae1d25d5f3 removed old not used variable from Space

0ea5497094 added CSVParser - a csv parser

430822bad8 make depend

604b47db32 added move semantics to Space class

added methods:
Space(Space && space);
Space & operator=(Space && space);
void set(const Space & space);
void set(Space && space);
Space & add(const Space & space);
Space & add(Space && space);
Space & add(const wchar_t * field, const Space & space);
Space & add(const wchar_t * field, Space && space);
Space & add(const std::wstring & field, const Space & space);
Space & add(const std::wstring & field, Space && space);

f8d24de386 fixed memory leak in CSVParser::parse() - a Space struct was allocated but never freed

6f50626dda fixed another memory leak in CSVParser::parse() (similar like before)

b574289054 namespace PT renamed to pt

59d4c9a9c8 changed utf8 functions: PascalCase to snake_case

abeca010cc fixed: SpaceParser was using space->set_empty_object() when parsing a space

and it cleared all values if not an empty object was provided to the set_space() method

82a21f6d85 removed: SpaceParser::SetSpace(...) methods, now ParseJSON/Space(...) methods take a space as an argument

6e4a0f68b3 removed from SpaceParser: SetDefault(), SkipEmpty(bool skip) and UseEscapeChar(bool escape) methods

c11aa78335 changed: names of methods in SpaceParser: PascalCase to snake_case

5ce36ea844 changed the way how child_spaces are created in Space class

- removed child_spaces and name pointers
- now a table with child spaces is created under "child_spaces" object field
- a name of the child space is stored in "name" field of the child object

added methods for manipulating with child spaces:
TableType * find_child_space_table()
bool child_spaces_empty()
size_t child_spaces_size()

Space * find_child_space(size_t table_index)
Space & add_child_space(const wchar_t * space_name)
Space & add_child_space(const std::wstring & space_name)

std::wstring * find_child_space_name()
std::wstring get_child_space_name()
bool is_child_space_name(const wchar_t * name)

added additional methods:
size_t str_size()
size_t wstr_size()
size_t object_size()
size_t table_size()

49c2b478c0 fixed return value from Space::add_child_space()

tomasz.sowa merged commit 848cdf9c03 into master

2021-05-27 10:37:36 +02:00

tomasz.sowa referenced this issue from a commit

2021-05-27 10:37:38 +02:00

Merge pull request 'api2021 part I' (#4) from api2021 into master

tomasz.sowa referenced this issue from a commit

2021-06-18 18:54:48 +02:00

fixed in Space::find_child_space_const(...) - clang address sanitizer reports stack-use-after-scope

Sign in to join this conversation.

No reviewers