Not using inheritance and preferring composition to avoid including the
detail/json.h header do an indirection via a unique_ptr and then just
re-use the Double() implementation.
threading/formatters/JSON.h currently includes rapidjson headers for declaring
the NullDoubleWriter. This appears mostly an internal detail, but
results in the situation that 1) we need to ship rapidjson headers with
the Zeek install tree and 2) taking care that external plugins are able
to find these headers should they include formatters/JSON.h.
There are currently no other Zeek headers that include rapidjson, so this
seems very unfortunate and self-inflicted given it's not actually required.
Attempt to hide this implementation detail with the goal to remove the
rapidjson includes with v7.1 and then also stop bundling and exposing
the include path to external plugins.
The NullDoubleWriter implementation moves into a new formatters/detail/json.h
header which is not installed.
Closes#3128
* origin/topic/johanna/1095-just-get-rid-of-it:
Re-add TYPE_COUNTER without function and deprecation marker.
Completely remove all traces of the COUNTER type.
This unfortunately cuases a ton of flow-down changes because a lot of other
code was depending on that definition existing. This has a fairly large chance
to break builds of external plugins, considering how many internal ones it broke.
This also now installs the full rapidjson include/ tree in order to
allow external plugins to properly build if they include
threading/formatters/JSON.h
The Zeek code base has very inconsistent #includes. Many sources
included a few headers, and those headers included other headers, and
in the end, nearly everything is included everywhere, so missing
#includes were never noticed. Another side effect was a lot of header
bloat which slows down the build.
First step to fix it: in each source file, its own header should be
included first to verify that each header's includes are correct, and
none is missing.
After adding the missing #includes, I replaced lots of #includes
inside headers with class forward declarations. In most headers,
object pointers are never referenced, so declaring the function
prototypes with forward-declared classes is just fine.
This patch speeds up the build by 19%, because each compilation unit
gets smaller. Here are the "time" numbers for a fresh build (with a
warm page cache but without ccache):
Before this patch:
3144.94user 161.63system 3:02.87elapsed 1808%CPU (0avgtext+0avgdata 2168608maxresident)k
760inputs+12008400outputs (1511major+57747204minor)pagefaults 0swaps
After this patch:
2565.17user 141.83system 2:25.46elapsed 1860%CPU (0avgtext+0avgdata 1489076maxresident)k
72576inputs+9130920outputs (1667major+49400430minor)pagefaults 0swaps
Tweaks:
- Small change to the logic for removing quotes around strings.
- Updated NEWS & COPYING.3rdparty
- Use of intrusive_ptr for stack-allocated StringVals
- Little bit of refactoring (I would love to merge the two BuildJSON() functions, too, but that's a larger task)
* origin/topic/timw/595-rapidjson:
Use the list of files from clang-tidy when searching for unit tests
Optimize json_escape_utf8 a bit by removing repeated calls to string methods
Expand unit test for json_escape_utf8 to include all of the strings from the ascii-json-utf8 btest
GHI-595: Convert from nlohmann/json to rapidjson for performance reasons
Convert type-checking macros to actual functions
* origin/topic/timw/595-json-perf:
Update COPYING.3rdparty
Use json::emplace to avoid some extra calls to operator[]
Use tessil/unordered-map instead of nlohmann/fifo-map to mitigate performance issues when logging JSON
* origin/topic/timw/150-to-json:
Update submodules for JSON work
Update unit tests for JSON logger to match new output
Modify JSON log writer to use the external JSON library
Update unit test output to match json.zeek being deprecated and slight format changes to JSON output
Add proper JSON serialization via C++, deprecate json.zeek
Add new method for escaping UTF8 strings for JSON output
Move do_sub method from zeek.bif to StringVal class method
Move record_fields method from zeek.bif to Val class method
Add ToStdString method for StringVal
This also installs symlinks from "zeek" and "bro-config" to a wrapper
script that prints a deprecation warning.
The btests pass, but this is still WIP. broctl renaming is still
missing.
#239
It now produces numbers as large as is required to match the data it
needs to represent. It's up to the consumer to decide how to
appropriately handle values that may be outside their supported
range/precision.
Fixes GH-282
This was causing occasional problems with the time on processes
running lots of threads. The use of gmtime in the json
formatter is the likely culprit due to the fact that the
json formatter runs in threads. More evidence for this is that
the problem only appears to exhibit when logs are being written
as JSON.
Cleaned up the surrounding code a bit and also added '[' as another
case (not sure that can happen, but doesn't hurt eihter).
* 'master' of https://github.com/aeppert/bro:
Whitespace
Remove
Remove.
Fix for JSON formatter
A fatal error, especially in DEBUG, should result in a core.
Seems to fix a case where an entry in the table may be null on insert.
In the event that the first entry in a record is optional AND not present, the serializer will incorrectly add a leading comma. This leading common is invalid JSON and will, more often than not, cause parser failures downstream.
* origin/topic/seth/json-formatter:
Updating a couple of tests.
Expanded support for modifying the timestamp format in the JSON formatter.
Ascii input reader now supports all config options per-input stream.
Added an option to the JSON formatter to use ISO 8601 for timestamps.
Refactored formatters and updated the the writers a bit.
Includes some minor bugfixes and cleanup at various places, including
in old code.
- It's not *exactly* ISO 8601 which doesn't seem to support
subseconds, but subseconds are very important to us and
most things that support ISO8601 seem to also support subseconds
in the way I'm implemented it.
- Formatters have been abstracted similarly to readers and writers now.
- The Ascii writer has a new option for writing out logs as JSON.
- The Ascii writer now has all options availble as per-filter
options as well as global.
