87 Commits

Author SHA1 Message Date
Christopher Dunn
6416350438 fix ,/. problem in reader
fixes #293
2015-06-18 22:45:36 -05:00
Christopher Dunn
2250b3c29d use Json::RuntimeError 2015-03-08 12:44:55 -05:00
Christopher Dunn
717b08695e clarify errors
* use macros for logic errors, not input errors
* throw on parsing failure in `operator>>()`, not assert
* throw on malloc, not assert
2015-03-08 12:06:22 -05:00
Christopher Dunn
62ad140d18 rejectDupKeys 2015-03-06 12:39:05 -06:00
Christopher Dunn
527332d5d5 add rejectDupKeys feature - not yet impld 2015-03-06 12:38:58 -06:00
Christopher Dunn
c312dd5ef7 Builder::operator[] plus tests 2015-03-05 09:18:01 -06:00
Christopher Dunn
19c49a459d fix Builders::validate()
(cherry picked from commit 626cfcdbb83bdc93243d7b64144e88f08c510085)
2015-03-04 21:14:53 -06:00
Christopher Dunn
2d653bd15d fix security hole for string-key-lengths > 2^30 2015-03-03 00:14:54 -06:00
Christopher Dunn
0c66e698fb allowSingleQuotes
issue #182
2015-02-24 15:49:45 -06:00
Christopher Dunn
7b3683ccd1 apply fix to old Reader 2015-02-19 11:37:17 -06:00
Christopher Dunn
58499031a4 fix all cases from issue -- all pass! 2015-02-19 11:35:28 -06:00
Christopher Dunn
c58e93b014 fix failing object case 2015-02-19 11:34:35 -06:00
Christopher Dunn
f4be815c86 failIfExtra
1. failing regression tests, from #164 and #107
2. implemented; tests pass
3. allow trailing comments
2015-02-13 09:39:08 -06:00
Christopher Dunn
3ebba5cea8 stop calling validate() in newReader/Writer()
By not calling validate(), we can add
non-invasive features which will be simply ignored when user-code
is compiled against an old version. That way, we can often
avoid a minor version-bump.

The user can call validate() himself if he prefers that behavior.
2015-02-11 11:15:32 -06:00
Christopher Dunn
56df206847 limit stackDepth for old (deprecated) Json::Reader too
This is an improper solution. If multiple Readers exist,
then the effect stackLimit is reduced because of side-effects.
But our options are limited. We need to address the security
hole without breaking binary-compatibility.

However, this is not likely to cause any practical problems because:

* Anyone using `operator>>(istream, Json::Value)` will be using the
new code already
* Multiple Readers are uncommon.
* The stackLimit is quite high.
* Deeply nested JSON probably would have hit the system limits anyway.
2015-02-11 10:20:53 -06:00
Christopher Dunn
4dca80da49 limit stackDepth 2015-02-11 10:20:47 -06:00
Christopher Dunn
249ad9f47f stackLimit 2015-02-11 10:01:58 -06:00
Christopher Dunn
99b8e856f6 stackLimit_ 2015-02-11 10:01:58 -06:00
Christopher Dunn
2474989f24 Old -> Our 2015-02-11 09:48:24 -06:00
Christopher Dunn
052050df07 copy Features to OldFeatures 2015-02-10 17:01:08 -06:00
Christopher Dunn
435d2a2f8d passes 2015-02-10 17:01:08 -06:00
Christopher Dunn
6123bd1505 copy Reader impl to OldReader 2015-02-10 17:01:08 -06:00
Christopher Dunn
7477bcfa3a renames for OldReader 2015-02-10 17:01:08 -06:00
Christopher Dunn
5e3e68af2e OldReader copied from Reader 2015-02-10 17:01:08 -06:00
Christopher Dunn
b56381a636 <stdexcept> 2015-02-09 18:29:11 -06:00
Christopher Dunn
f757c18ca0 add all features 2015-02-09 18:24:56 -06:00
Christopher Dunn
3cf9175bde remark defaults via doxygen snippet 2015-02-09 18:16:24 -06:00
Christopher Dunn
a9e1ab302d Builder::settings_
We use Json::Value to configure the builders so we can maintain
binary-compatibility easily.
2015-02-09 17:30:11 -06:00
Christopher Dunn
66a8ba255f clarify Builders 2015-02-09 01:29:43 -06:00
Christopher Dunn
2c1197c2c8 CharReader/Builder
* CharReaderBuilder is similar to StreamWriterBuilder.
* use rdbuf(), since getline(string) is not required to handle EOF as delimiter
2015-02-08 13:22:09 -06:00
Christopher Dunn
1e3149ab75 rm trailing newlines for *all* comments
This will make it easier to fix newlines consistently.
2015-01-25 14:32:13 -06:00
Mark Zeren
e39fb0083c Normalize comment EOLs while reading instead of while writing
Tests are currently failing when git cloning on Windows with autocrlf = true. In
that setup multiline comments contain \r\n EOLs. The test code assumes that
comments contain \n EOLs and opens the .actual files (etc.) with "wt" which
converts \n to \r\n. Thus we end up with \r\r\n EOLs in the output, which
triggers a test failure.

Instead we should cannonicalize comments while reading so that they contain only
\n EOLs. This approach simplifies other parts of the reader and writer logic,
and requires no changes to the test. It is a breaking change, but probably the
Right Thing going forward.

This change also fixes dereferencing past the end of the comment string in
StyledWriter::writeCommentBeforeValue.

Tests should be added with appropriate .gitattributes for the input files to
ensure that we run tests for DOS, Mac, and Unix EOL files on all platforms. For
now this change is enough to unblock Windows builds.

issue #116
2015-01-20 13:45:44 -06:00
Christopher Dunn
836f0fb863 fix comments before several types
tests pass
2015-01-20 12:23:44 -06:00
Christopher Dunn
66eb72f121 use SwapPayload() to retain comments
All tests pass, but we might be missing coverage.

issue #47
2015-01-20 12:07:01 -06:00
datadiode
09d352ac13 Remove unused functions 2015-01-17 13:26:23 +01:00
datadiode
50753bb808 Simplify Reader::decodeNumber() 2015-01-17 13:21:42 +01:00
datadiode
c1441ef5e0 stricter float parsing
fixes `test/jsonchecker/fail31.json`
(issue #113)
2015-01-16 15:05:12 -06:00
Gaurav
767713be2b Remove gcc compilation warning in json_reader.cpp
Submitting Patch for Issue : https://github.com/open-source-parsers/jsoncpp/issues/77
It will fix warnings in json_reader.cpp
2014-11-17 14:04:03 +05:30
Mara Kim
b84a39cae5 Add public semantic error reporting
Closes open-source-parsers/jsoncpp#57
2014-10-23 02:18:14 -05:00
Cory Quammen
fd06bfca79 Removed unneeded newlines from parsed comments
Newlines from comments separated by lines are retained when comments
are appended, so adding a newline between separate comments for a
node is not needed.
2014-10-09 16:33:29 -04:00
Chuck Atkins
9dc9026e0b Workaround for missing C99 functions in older versions of Visual Studio 2014-09-19 13:16:09 -04:00
Christopher Dunn
4002f8a4be Revert "Revert "Removed vim mode lines.""
This reverts commit af77b5b59456f6a9c7a9c6a5464001ef5665f897.

See discussion at
  32009b17e4 (commitcomment-7827708)
2014-09-18 16:46:40 -07:00
Christopher Dunn
af77b5b594 Revert "Removed vim mode lines."
This reverts commit 32009b17e4d35a812575cfcf6bae9ec49fc67446.
2014-09-16 12:42:32 -07:00
Aaron Jacobs
11086dd6a7 Enabled PointerBindsToType in clang-format options. 2014-09-15 10:15:29 +10:00
Aaron Jacobs
30b07c0275 Ran clang-format over all .h and .cpp files.
clang-format -i $(find . -name '*.h' -or -name '*.cpp')
2014-09-15 10:14:48 +10:00
Aaron Jacobs
32009b17e4 Removed vim mode lines.
Users can set their own preferences in their personal vimrc.
2014-09-15 08:23:41 +10:00
pffang
27e3263894 WinCE Compatibility Fix
Note: str.imbue and std::locale::classic() are not supported on WINCE
2014-07-10 20:27:52 -07:00
Christopher Dunn
8582876c5c vim modelines 2014-07-10 20:24:23 -07:00
Aaron Jacobs
445328ace6 Fixed some clang-format weirdness. 2014-07-01 09:15:11 +10:00
Aaron Jacobs
9fa4e849a1 Ran clang-format over all .h and .cpp files.
clang-format -i $(find . -name '*.h' -or -name '*.cpp')
2014-07-01 08:48:54 +10:00