Commit Graph

1067 Commits

Author SHA1 Message Date
Eric Hawicz
b01d50b8dc Merge pull request #639 from smcv/symbol-versions
build: Add a symbol version to all exported symbols
2020-07-04 14:26:20 -04:00
Simon McVittie
c2c94024f5 build: Add symbol versions to all exported symbols
With this version script, newly-linked binaries that depend on the
json-c shared library will refer to its symbols in a versioned form,
preventing their references from being resolved to a symbol of the same
name exported by json-glib or libjansson if those libraries appear in
dependency search order before json-c, which will usually result in
a crash. This is necessary because ELF symbol resolution normally uses
a single flat namespace, not a tree like Windows symbol resolution.
At least one symbol (json_object_iter_next()) is exported by all three
JSON libraries.

Linking with -Bsymbolic is not enough to have this effect in all cases,
because -Bsymbolic only affects symbol lookup within a shared object,
for example when json_object_set_serializer() calls
json_object_set_userdata(). It does not affect calls from external
code into json-c, unless json-c was statically linked into the
external caller.

This change will also not prevent code that depends on json-glib or
libjansson from finding json-c's symbols and crashing; to prevent
that, a corresponding change in json-glib or libjansson would be needed.

Adding a symbol-version is a backwards-compatible change, but once
added, removing or changing the symbol-version on a symbol would be an
incompatible change that requires a SONAME bump.

Resolves: https://github.com/json-c/json-c/issues/621
(when combined with an equivalent change to libjansson).

Signed-off-by: Simon McVittie <smcv@collabora.com>
2020-07-01 18:24:26 +01:00
Eric Haszlakiewicz
6465e74020 Use constants referring to the signed integer types when setting SSIZE_T_MAX.
In practice, the sizes of the signed and unsigned integer types will
almost cetainly be the same, but this is more correct.
Pointed out in issue #638.
2020-07-01 00:34:46 +00:00
Eric Haszlakiewicz
34334e5d3f Replace one call to json_object_new_array() with json_object_new_array_ext() to ensure it at least minimally works. 2020-06-29 02:35:26 +00:00
Eric Haszlakiewicz
0b67caec1a Add doc comment for json_object_new_array_ext(). 2020-06-29 02:31:32 +00:00
Eric Haszlakiewicz
a4e3700972 Fix code formatting 2020-06-29 02:31:18 +00:00
Eric Haszlakiewicz
5f3bf70f03 Make sure TEST_PARSE_CHUNKSIZE is valid if it's set. 2020-06-29 02:18:34 +00:00
Eric Haszlakiewicz
f23486a321 In the json_tokener_state_number case, explicitly adjust what "number" characters are allowed based on the exact micro-state that we're in, and check for invalid following characters in a different way, to allow a valid json_type_number object to be returned at the top level.
This causes previously failing strings like "123-456" to return a valid json_object with the appropriate value.  If you care about the trailing content, call json_tokener_parse_ex() and check the parse end point with json_tokener_get_parse_end().
2020-06-29 02:14:26 +00:00
Eric Haszlakiewicz
6eac6986c9 Fix incremental parsing of invalid numbers with exponents, such as "0e+-" and "12.3E12E12", while still allowing "0e+" in non-strict mode.
Deprecate the json_parse_double() function from json_util.h
2020-06-27 15:35:04 +00:00
Eric Haszlakiewicz
84e6032883 Issue #635: Fix "expression has no effect" warning in json_tokener.c by casting to void. 2020-06-23 02:51:46 +00:00
Eric Haszlakiewicz
7a72805e34 Fix memory leak in test_parse's single_incremental_parse(). 2020-06-22 01:12:03 +00:00
Eric Haszlakiewicz
7d3c2d9fad Drop extra blank lines from arraylist.h 2020-06-21 18:30:40 +00:00
Eric Haszlakiewicz
a68566bf6a Issue #616: Change the parsing of surrogate pairs in unicode escapes so it uses a couple of additional states instead of assuming the low surrogate is already present, to ensure that we correctly handle various cases of incremental parsing. 2020-06-21 18:29:57 +00:00
Eric Haszlakiewicz
197e372464 In test_parse, fix lengths passed during a couple of incremental tests.
Add an optional way to use an incremental chunk size ($TEST_PARSE_CHUNKSIZE)
for all of the single_basic_parse tests, in additional to the regular way.
2020-06-21 17:36:38 +00:00
Eric Haszlakiewicz
36118b681e Rearrange the json_tokener_state_escape_unicode case in json_tokener to simplify the code slightly and make it a bit easier to understand.
While here, drop the utf8_replacement_char that is unnecesarily added if we run out of input in the middle of a unicode escape.  No other functional changes (yet).
2020-06-21 03:10:55 +00:00
Eric Haszlakiewicz
50179fb09f Update the json_tokener_parse_ex() docs to clarify that the final '\0' character is to be included in length passed in. 2020-06-21 02:35:42 +00:00
Eric Hawicz
da76ee26e7 Merge pull request #633 from dota17/issue616
fix issue 616: support the surrogate pair in split file.
2020-06-20 22:33:17 -04:00
Eric Haszlakiewicz
e26a1195f4 Add json_object_array_shrink() (and array_list_shrink()) and use it in json_tokener to minimize the amount of memory used. This results in a 39%-50% reduction in memory use (peak RSS, peak heap usage) on the jc-bench benchmark and 9% shorter runtime.
Also add the json_object_new_array_ext, array_list_new2, and array_list_shrink functions.
2020-06-20 18:03:04 +00:00
Eric Hawicz
99bb2121c6 Merge pull request #632 from json-c/json_object-split
Json object split
2020-06-20 13:04:32 -04:00
Eric Haszlakiewicz
0710c835a1 Reformat the json_object-split branch with clang-format 2020-06-16 13:17:58 +00:00
Eric Hawicz
401b29c021 Merge pull request #634 from micahsnyder/json-c-fPIC
Issue #508: `-fPIC` to link libjson-c.a with libs
2020-06-14 17:56:35 -04:00
Micah Snyder
60494684a1 Issue #508: -fPIC to link libjson-c.a with libs
json-c has symbol collisions with other popular C JSON libraries.
To prevent random crashes in downstream applications that may use a
library which depends on json-c, the solution is to statically link
libjson-c.a into those libraries.

`-fPIC`/`-fPIE` is required when building a `.a` so it can be linked
into a `.so`.
2020-06-14 12:35:49 -04:00
Eric Haszlakiewicz
9128ec49b1 Include unistd.h to fix the build on OSX 2020-06-14 03:11:44 +00:00
Eric Haszlakiewicz
5ebfeaedc5 Drop the _delete field from struct json_object and call the type-specific delete functions directly from json_object_put. (Thanks @dota17 for the suggestion in PR #632!) 2020-06-13 18:34:35 +00:00
Eric Haszlakiewicz
4c10712114 Drop the useless "char data[1]" from struct json_object. Fix a typo in a comment. 2020-06-13 18:25:32 +00:00
dota17
c1b872d817 fix issue 616: support the surrogate pair in split file. 2020-06-08 17:19:32 +08:00
Eric Haszlakiewicz
85c244f048 Eliminate unnecessary cast that was added to test_double_serializer. 2020-06-07 18:50:10 +00:00
Eric Haszlakiewicz
02fe2e0ccd Summarize the changes from the json_object-split branch in the ChangeLog. 2020-06-07 18:45:17 +00:00
Eric Haszlakiewicz
ecdfeb18cf Move the ssize_t typedef from json_inttypes.h to json_object_private.h so as not to affect publically exposed symbols. 2020-06-07 18:29:56 +00:00
Eric Haszlakiewicz
66d91fdf86 The split of json_object into type-specific sub-structures is now functionally complete.
Remove all of the temporary defines, structures, old compat fuctions, extra fields, etc... that were needed during the conversion to a split set of json_object_* structures.
2020-06-07 17:23:56 +00:00
Eric Haszlakiewicz
c4cc673071 More fixes for old MSVC builds. 2020-06-07 15:25:59 +00:00
Eric Haszlakiewicz
0a16b23adf Fix typo in previous commit to check for SSIZE_T on MSVC. 2020-06-07 15:19:29 +00:00
Eric Haszlakiewicz
eab1375123 Change CMakeLists.txt to look for SSIZE_T on MSVC too. 2020-06-07 03:36:59 +00:00
Eric Haszlakiewicz
b0466b626b On MSVC, add a ssize_t typedef using SSIZE_T from BaseTsd.h 2020-06-07 03:27:13 +00:00
Eric Haszlakiewicz
0fc9d91277 Kick json_type_string out of struct json_object.
The default is now that string data is stored inline at the end of json_object, though to allow for json_object_set_string() to set a _longer_ string, we still need to allow for the possibility of a separate char * pointer.
All json types have been split out now, next step it cleanup.
2020-06-07 02:56:59 +00:00
Eric Hawicz
bc06f82f85 Merge pull request #628 from clamwin/compat-fixes-master
get_cryptgenrandom_seed: compat with old windows + fallback
2020-06-03 23:53:46 -04:00
Gianluigi Tiesi
481d0a8ede get_cryptgenrandom_seed: compat with old windows + fallback 2020-06-03 07:51:42 +02:00
Eric Haszlakiewicz
1c6086a86a Apply the fix from @pointbre in issue #626 to skip "inline" on AIX, but invert the test to make it a little easier to understand. 2020-05-31 03:22:14 +00:00
Eric Haszlakiewicz
fe308b8862 Issue #626: Restore compatibility with cmake 2.8 by adjusting quoting and explicitly defining the PROJECT_VERSION* variables. 2020-05-30 19:37:46 +00:00
Eric Haszlakiewicz
9ecb1222bd Kick json_type_int and json_type_double out of struct json_object.
Clean up the code in json_object_new_* a bit by dropping unnecesary json_object_base variables.
2020-05-26 02:31:35 +00:00
Eric Haszlakiewicz
0351bb55c8 Declare variables earlier, to appease Visual Studio 2010. 2020-05-25 04:10:11 +00:00
Eric Haszlakiewicz
d1f83bf5ea Kick json_type_boolean out of struct json_object. 2020-05-25 04:05:32 +00:00
Eric Haszlakiewicz
5d89fc8a9d Add some backwards compat for Visual Studio 2013. 2020-05-25 03:52:36 +00:00
Eric Haszlakiewicz
02b687b9a6 Kick json_type_array out of struct json_object; re-enable the test_deep_copy test. 2020-05-25 03:44:56 +00:00
Eric Haszlakiewicz
853b4b5dee Start splitting struct json_object into multiple sub-types, as descibed at https://github.com/json-c/json-c/wiki/Proposal:-struct-json_object-split
The current changes split out _only_ json_type_object, and thus have a number of hacks
 to allow the code to continue to build and work.

Originally  mentioned in issue #535.
When complete, this will probably invalidate #552.
This is likely to cause notable conflicts in any other significant un-merged
changes, such as PR#620.
2020-05-25 03:18:36 +00:00
Eric Haszlakiewicz
4a546e7b2f In arraylist, use malloc instead of calloc, avoid clearing with memeset until we really need to, and micro-optimize array_list_add(). 2020-05-24 03:54:04 +00:00
Eric Hawicz
fbe1543644 Merge pull request #622 from besser82/topic/besser82/doc_subdir
doc: Move Doxyfile into doc subdir.
2020-05-18 15:30:21 -04:00
Björn Esser
1e94da779a CMake: Fix grammar: written -> wrote. 2020-05-18 20:36:16 +02:00
Björn Esser
61e2bae511 doc: Move Doxyfile into doc subdir 2020-05-18 20:36:16 +02:00
Eric Haszlakiewicz
fa6a7dccb9 With the change in cc80203, Doxyfile no longer needs to be updated for a release. 2020-05-18 17:31:22 +00:00