jASTERIX
jASTERIX is a C++ Library for EUROCONTROL’s ASTERIX to JSON conversion. It is part of the OpenATS COMPASS project, but released as a separate library to allow easy usage in other projects.
The library allows decoding of binary ASTERIX data into JSON. The ASTERIX definitions are configuration only, so additional or customized categories or editions can be added without having to recompile. Also, using the Intel TBB library allows for reasonable performance using multi-threading.
For a detailed list of features, as well as the offered commercial services, please refer to OpenATS jASTERIX.
Supported Editions
| CAT |
Editions |
REFs |
SPFs |
| 001 |
1.1 |
|
|
| 002 |
1.0 |
|
|
| 004 |
1.4 |
|
|
| 010 |
0.24 (Sensis), 0.31 |
|
|
| 019 |
1.2, 1.3 |
|
|
| 020 |
1.5, 1.8 |
1.3 |
|
| 021 |
0.26, 2.1, 2.4 |
1.5 |
Aireon |
| 023 |
1.2 |
|
|
| 025 |
1.1, 1.5 |
|
|
| 030 |
7.0 |
|
|
| 034 |
1.26 |
|
|
| 048 |
1.15, 1.23 |
1.9 |
|
| 062 |
1.12, 1.16, 1.18, 1.21 |
1.2, 1.4 |
ARTAS TRIs |
| 063 |
1.0, 1.1 |
|
|
| 065 |
1.2, 1.3 |
|
|
| 247 |
1.2 |
|
|
| 252 |
7.0 |
|
|
Contents
- Folder “analyze”: Contains Python scripts for data analysis
- Folder “appimage”: Contains files required for AppImage building
- Folder “cmake_modules”: Contains cmake find scripts
- Folder “definitions”: Contains ASTERIX configuration definitions
- Folder “include”: Main headers
- Folder “lib”: User header-only libraries
- Folder “src”: Contains source code
- CMakeLists.txt: CMake config file
- LICENSE: GPL license
- README.md: This file
Requirements / Used Libraries
The following libraries are mandatory:
- Boost
- Intel Thread Building Blocks
- LibArchive
- libpcap
The following libaries are optional (can be deactivated in CMakeList.txt):
Also, the Nlohmann::JSON and the Catch2 libraries are used.
Client Installation without Building
A client AppImage is supplied, which can be executed (without setup effort) under all recent Linux distributions. The AppImage is built on Debian 10 (Buster), so it runs on Debian 10 and later, Ubuntu 18.04 and later, and comparable distributions.
Download the AppImage and the jASTERIX definitions from the Releases page, and extract the definitions into a local folder, e.g. ‘definitions’. Execute the following command to add the executable flag to the AppImage:
chmod +x jASTERIX_client_v0.1.6-deb10.AppImage
After this, the jASTERIX client can be run from the console (see Usage section).
Installation with Building
Download the source code from this page, then execute the following commands in the check-out folder:
mkdir build
cd build
cmake ..
make
To install the built libary in the system, execute the following command in the build folder:
sudo make install
Both steps in one go, with Ninja and install prefix /usr: scripts/make.sh (options --clean, --test).
Usage
Help
After building, the following jASTERIX client can be used (from the check-out folder):
./jASTERIX_client-x86_64.AppImage --help
INFO : Allowed options:
--help produce help message
--filename arg input file name
--definition_path arg path to jASTERIX definition files
--framing arg input framine format, as specified in the framing
definitions. raw/netto is default
--frame_limit arg number of frames to process with framing, default
-1, use -1 to disable.
--frame_chunk_size arg number of frames to process in one chunk, default
1000, use -1 to disable.
--data_block_limit arg number of data blocks to process without framing,
default -1, use -1 to disable.
--data_block_chunk_size arg number of data blocks to process in one chunk,
default 1000, use -1 to disable.
--data_write_size arg number of frame chunks to write in one file
write, default 1, use -1 to disable.
--debug print debug output (only for small files)
--debug_include_framing print debug output including framing, debug still
has to be set, disable per default
--print_cat_info print category info
--single_thread process data in single thread
--only_cats arg restricts categories to be decoded, e.g. 20,21.
--editions arg set non-default editions per category, e.g.
21:0.26,48:1.15.
--pcap input file is a PCAP capture (libpcap); ASTERIX
payload is extracted and decoded as raw/netto (no
framing).
--log_perf enable performance log after processing
--analyze analyze data sources and contents
--analyze_csv analyze data sources and contents, print as CSV
--analyze_record_limit arg number of records to analyze. 0 (default)
disables limit.
--add_artas_md5 add ARTAS MD5 hashes
--check_artas_md5 arg add and check ARTAS MD5 hashes (with record
data), stating which categories to check, e.g.
1,20,21,48
--flat output in flat/columnar format (cat -> leaf_name
-> array)
--add_record_data add original record data in hex
--print print JSON output
--print_indent arg intendation of json print, use -1 to disable.
--write_type arg optional write type, e.g. text,zip. needs
write_filename.
--write_filename arg optional write filename, e.g. test.zip.
--encode_flat_zip arg encode flat columnar JSON from a zip (members =
chunks, as written by --flat --write_type zip)
back to raw/netto ASTERIX. needs encode_filename.
--encode_flat arg encode flat columnar JSON from a text file (one
flat chunk object per line) back to raw/netto
ASTERIX. needs encode_filename.
--encode_filename arg output binary file for flat encoding (raw/netto
ASTERIX).
--encode_cat arg restrict flat encoding to a single category. 0
(default) encodes all categories present.
editions are taken from --editions (defaults
otherwise).
To test decode and measure performance, use:
./jASTERIX_client-x86_64.AppImage --definition_path definitions/ --filename example.ff --framing ioss --log_perf
INFO : jASTERIX client: decoded 674150 frames, 1073956 records in 0h 0m 8s 936ms: 75442 fr/s, 120183 rec/s
Performance depends on the used hardware, frame_chunk_size and ASTERIX data content. On the author’s workstation (Intel i7 4790k) the performance varies between 50k-150k rec/s.
Decode & Print
To decode and print an ASTERIX file as JSON text, use:
./jASTERIX_client-x86_64.AppImage --definition_path definitions/ --filename src/test/cat020ed1.5.bin --print
{
"data_blocks": [
{
"category": 20,
"content": {
"index": 3,
"length": 98,
"records": [
{
"010": {
"SAC": 0,
"SIC": 2
},
"020": {
"CHN": 0,
"CRT": 0,
"DME": 0,
"FX": 1,
"FX2": 0,
"GBS": 0,
"HF": 0,
"MS": 1,
"OT": 0,
"RAB": 0,
"SIM": 0,
"SPI": 0,
"SSR": 0,
"TST": 0,
"UAT": 0,
"VDL4": 0
},
"041": {
"Latitude": 47.88232132925,
"Longitude": 16.32056296698
},
"042": {
"X": 173529.5,
"Y": 45109.0
},
"070": {
"G": 0,
"L": 1,
"Mode-3/A code": 7000,
"V": 0
},
"090": {
"Flight Level": 11.25,
"G": 0,
"V": 0
},
"140": {
"Time of Day": 33502.7109375
},
"161": {
"Track Number": 3528
},
"170": {
"CDM": 3,
"CNF": 0,
"CST": 0,
"FX": 0,
"MAH": 0,
"STH": 0,
"TRE": 0
},
"202": {
"Vx": -13.75,
"Vy": -9.25
},
"210": {
"Ax": 0.0,
"Ay": 0.0
},
"220": {
"Target Address": 148527
},
"230": {
"AIC": 0,
"ARC": 1,
"B1A": 0,
"B1B": 0,
"COM": 1,
"MSSC": 0,
"STAT": 0
},
"250": {
"Mode S MB Data": [
{
"BDS1": 1,
"BDS2": 0,
"MB Data": "10000000a00000"
},
{
"BDS1": 1,
"BDS2": 7,
"MB Data": "00000000000000"
}
],
"REP": 2
},
"400": {
"Contributing Receivers": [
{
"RUx": 0
},
{
"RUx": 0
},
{
"RUx": 0
},
{
"RUx": 0
},
{
"RUx": 0
},
{
"RUx": 0
},
{
"RUx": 0
},
{
"RUx": 0
},
{
"RUx": 0
},
{
"RUx": 0
},
{
"RUx": 16
},
{
"RUx": 0
},
{
"RUx": 0
},
{
"RUx": 32
},
{
"RUx": 0
},
{
"RUx": 34
}
],
"REP": 16
},
"FSPEC": [
true,
true,
true,
true,
true,
true,
true,
true,
true,
true,
true,
false,
true,
false,
false,
true,
false,
true,
false,
false,
false,
true,
true,
true,
true,
false,
false,
false,
false,
true,
false,
false
],
"REF": {
"PA": {
"DOP": {
"DOP-x": 4.5,
"DOP-xy": -3.75,
"DOP-y": 3.75
},
"SDC": {
"COV-XY (Covariance Component)": -30.5,
"SDC (X-Component)": 34.25,
"SDC (Y-Component)": 31.0
},
"SDW": {
"COV-WGS (Lat/Long Covariance Component)": -0.00033795783,
"SDW (Latitude Component)": 0.00028431372999999997,
"SDW (Longitude Component)": 0.00044524603
},
"available": [
true,
true,
false,
true,
false,
false,
false,
false
]
},
"REF_FSPEC": [
true,
false,
false,
false,
false,
false,
false,
false
]
},
"index": 3,
"length": 98
}
]
},
"length": 101
}
]
}
Decoded JSON can also be written to a text or ZIP file using the appropriate arguments.
Flat / Columnar Output
By default the JSON output is structured (nested per record, mirroring the ASTERIX data block / record hierarchy as shown above). Using the --flat argument, the output is instead given in a flat, columnar layout: keyed by category number, each leaf field becomes a top-level array with one entry per record (indexed by record order). All arrays of a category have the same length; fields not present in a record (not selected by its FSPEC) are null at that position.
./jASTERIX_client-x86_64.AppImage --definition_path definitions/ --filename src/test/cat048ed1.15.bin --flat --print
{
"48": {
"010.SAC": [0, 0, 5],
"010.SIC": [1, 1, 2],
"040.RHO": [73.92, null, 55.10],
"040.THETA": [89.67, null, 120.33],
"070.Mode-3/A reply": [470, 471, null],
"090.Flight Level": [370.0, 365.0, null],
"140.Time-of-Day": [33499.84, 33500.12, 33500.50],
"220.AIRCRAFT ADDRESS": [11226301, 11226301, null]
}
}
This format is convenient for direct import into columnar/data-frame tooling (e.g. pandas, Apache Arrow). It is available both via the CLI --flat flag and the library API (the do_flat parameter of decodeFile() / decodeData() / decodePCAPFile()).
Repetitive Items in Flat Output
Repetitive items (a REP count octet followed by N identical subitems) are flattened down to the leaf. Each leaf gets its own column, keyed by its full path. The cell of a record holds an array of scalars, aligned by repetition index. A <prefix>.REP column holds the count. The structured output keeps its array-of-objects form for the same item.
Structured:
"SPF": { "REP": 2, "Target Report Identifiers": [ { "TRI": "76427f0a" }, { "TRI": "10c4d792" } ] }
Flat:
"SPF.REP": [2, ...],
"SPF.Target Report Identifiers.TRI": [["76427f0a", "10c4d792"], null, ...]
An item with several fields per repetition produces one such column per field. The columns stay aligned by repetition index. Extendable items are not flattened this way, they keep their whole array-of-objects in one column.
Note that this representation changed in v0.1.5. Before, the whole repetitive item was a single column holding an array of objects per record. Consumers of flat output must be adapted.
Re-Encoding to Binary ASTERIX
Flat JSON output can be encoded back to binary ASTERIX. Decode with --flat --write_type text or --flat --write_type zip, then encode the result with --encode_flat or --encode_flat_zip. The output file is set with --encode_filename. Round trips are byte-exact.
./jASTERIX_client-x86_64.AppImage --definition_path definitions/ --filename recording.rec --flat --write_type zip --write_filename flat.zip
./jASTERIX_client-x86_64.AppImage --definition_path definitions/ --encode_flat_zip flat.zip --encode_filename reencoded.rec
Use --encode_cat to restrict the output to one category. Editions come from --editions, or from the defaults. The encoder always writes raw/netto output, framings are not generated again.
./jASTERIX_client-x86_64.AppImage --definition_path definitions/ --encode_flat flat.json --encode_cat 62 --editions 62:1.21 --encode_filename cat062.rec
The library API encodes structured JSON directly, via encodeRecord() and encodeDataBlock(). Both return a complete data block (CAT, LEN, records).
REF and SPF Fields
A Reserved Expansion Field (REF) and a Special Purpose Field (SPF) carry a leading length octet. That length is authoritative for the framing of the record.
If the content of such a field does not match the selected definition, jASTERIX does not fail the data block. The partial decode is discarded. The field content is kept as a raw hex string, the same as when no REF or SPF edition is selected. A ref_error or spf_error flag is added to the record, and a warning is logged. Parsing resumes after the announced length, so the rest of the record and the data block decode normally. This happens for example when equipment writes its own content into an SPF.
Only an announced length that overruns the data block stays a hard decode error, because the stream is then out of sync.
Affected records are counted. The library reports them through numREFErrors() and numSPFErrors(). The analysis result holds them as num_ref_errors and num_spf_errors, next to num_errors, which stays 0 for these records. In flat output the leaf columns of the field are null for such records.
PCAP Captures
ASTERIX data captured in a PCAP file (e.g. recorded with tcpdump/Wireshark) can be decoded directly using the --pcap argument. The ASTERIX payload is extracted from the network streams and decoded as raw/netto (no framing). Per-packet capture timestamps are tracked during decoding.
./jASTERIX_client-x86_64.AppImage --definition_path definitions/ --filename capture.pcap --pcap --print
--pcap combines with the other arguments, e.g. --flat, --only_cats, --write_type/--write_filename, and --analyze/--analyze_csv.
In the library API the corresponding entry points are:
decodePCAPFile(filename, callback, do_flat) - decode a PCAP capture, delivering JSON chunks to the callbackanalyzePCAPFile(filename, record_limit) - return a JSON summary of the data sources/contents in a captureanalyzePCAPFileCSV(filename, record_limit) - same analysis, returned as CSV text
Analysis
In the ‘analyze’ folder, example Python scripts are given which can be used to analyse decoded JSON in a simple but powerful manner. Using the decoder and the print function, JSON data is piped into a Python script, where it can be analysed. A user could simply write the own scripts for deep data analysis.
Data Item Analysis
The Python script ‘analyze/data_items.py’ allows for the following options:
python3 analyze/data_items.py --help
usage: data_items.py [-h] --framing FRAMING [--cats CATS]
ASTERIX data item analysis
optional arguments:
-h, --help show this help message and exit
--framing FRAMING Framing True or False
--cats CATS ASTERIX categories to be analyzed as CSV
To analyze the data items in a recording, start the jASTERIX decoder and pipe the output in the the script with the appropriate arguments.
The following example analyzes the CAT048 content in the first 100000 frames:
./jASTERIX_Client-x86_64.AppImage --definition_path definitions/ --filename example.ff --framing ioss --print --print_indent -1 --only_cats 48 --frame_limit 100000 | python3 analyze/data_items.py --framing True
framing True
cats [48]
blocks 1 time 0.28s records 73 filtered 652 rate 2601 rec/s
...
blocks 10 time 2.24s records 813 filtered 7718 rate 3813 rec/s
num records 813
data items
'048' count 813
'048.010' count 813 (100%)
'048.010.SAC' count 813 (100%) min '0' max '1'
'048.010.SIC' count 813 (100%) min '0' max '2'
'048.020' count 813 (100%)
'048.020.FX' count 813 (100%) min '0' max '0'
'048.020.RAB' count 813 (100%) min '0' max '0'
'048.020.RDP' count 813 (100%) min '0' max '0'
'048.020.SIM' count 813 (100%) min '0' max '0'
'048.020.SPI' count 813 (100%) min '0' max '0'
'048.020.TYP' count 813 (100%) min '2' max '3'
'048.040' count 813 (100%)
'048.040.RHO' count 813 (100%) min '9.703125' max '249.6875'
'048.040.THETA' count 813 (100%) min '0.03295898436' max '359.54956038323996'
'048.042' count 60 (7%)
'048.042.X-Component' count 60 (100%) min '-159.71875' max '163.96875'
'048.042.Y-Component' count 60 (100%) min '-169.078125' max '72.125'
'048.070' count 813 (100%)
'048.070.G' count 813 (100%) min '0' max '0'
'048.070.L' count 813 (100%) min '0' max '0'
'048.070.Mode-3/A reply' count 813 (100%) min '21' max '6732'
'048.070.V' count 813 (100%) min '0' max '1'
'048.090' count 807 (99%)
'048.090.Flight Level' count 807 (100%) min '18.0' max '410.0'
'048.090.G' count 807 (100%) min '0' max '0'
'048.090.V' count 807 (100%) min '0' max '0'
'048.140' count 813 (100%)
'048.140.Time-of-Day' count 813 (100%) min '71.8046875' max '1578.484375'
'048.161' count 60 (7%)
'048.161.TRACK NUMBER' count 60 (100%) min '1802' max '1847'
'048.170' count 60 (7%)
'048.170.CDM' count 60 (100%) min '0' max '2'
'048.170.CNF' count 60 (100%) min '0' max '0'
'048.170.DOU' count 60 (100%) min '0' max '0'
'048.170.FX' count 60 (100%) min '1' max '1'
'048.170.GHO' count 60 (100%) min '0' max '0'
'048.170.MAH' count 60 (100%) min '0' max '1'
'048.170.RAD' count 60 (100%) min '0' max '2'
'048.170.SUP' count 60 (100%) min '0' max '0'
'048.170.TCC' count 60 (100%) min '1' max '1'
'048.170.TRE' count 60 (100%) min '0' max '0'
'048.200' count 60 (7%)
'048.200.CALCULATED GROUNDSPEED' count 60 (100%) min '325.41500602113604' max '567.553653310327'
'048.200.CALCULATED HEADING' count 60 (100%) min '81.48559566604' max '348.71704085692'
'048.SPF' count 60 (7%) min '400101ba00' max '44d3006e0004'
'048.index' count 813 (100%) min '34' max '3301579'
'048.length' count 813 (100%) min '13' max '37'
ADS-B Quality Indicators Analysis
The Python script ‘analyze/adsb_quality.py’ allows for the following options:
usage: adsb_quality.py [-h] --framing FRAMING
ASTERIX data item analysis
optional arguments:
-h, --help show this help message and exit
--framing FRAMING Framing True or False
The following target report counters are calculated:
- MOPS Version is supported by the GS: Counters of supported by the ADS-B ground station
- MOPS Version Number: Counters of MOPS versions
- Emitter Category: Counters of emitter category (aircraft type)
- Navigation Accuracy Category - Position: Counter of NACp values of DO-260A/B transponders
- MOPS Version for ECAT: Counters of MOPS versions, split up for different emitter category values
- NACp for ECAT: Counters of NACp values, split up for different emitter category values
To analyze the data items in a recording, start the jASTERIX decoder and pipe the output in the the script with the appropriate arguments.
The following example analyzes the CAT021 content in the first 100000 frames:
./jASTERIX_Client-x86_64.AppImage --definition_path definitions/ --filename example.ff --framing ioss --print --print_indent -1 --frame_limit 100000 | python3 analyze/adsb_quality.py --framing True
blocks 1 time 0.91s records 6461 rate 7104 rec/s
...
blocks 10 time 8.13s records 65751 rate 8087 rec/s
num ads-b records 65751
MOPS Version is supported by the GS:
none: 0 (0.00%)
MOPS Version is supported by the GS: 65751 (100.00%)
--------------------------------------------------------------------------------
MOPS Version Number:
none: 0 (0.00%)
ED102/DO-260: 21515 (32.72%)
DO-260A: 1481 (2.25%)
ED102A/DO-260B: 42755 (65.03%)
--------------------------------------------------------------------------------
Emitter Category:
none: 23576 (35.86%)
light aircraft <= 15500 lbs: 1624 (3.85%)
75000 lbs < medium a/c < 300000 lbs: 31672 (75.10%)
300000 lbs <= heavy aircraft: 8561 (20.30%)
rotocraft: 318 (0.75%)
--------------------------------------------------------------------------------
Navigation Accuracy Category - Position:
none: 21515 (32.72%)
EPU < 10 m, VEPU < 15 m: 13395 (30.28%)
EPU < 30 m, VEPU < 45 m: 14951 (33.80%)
EPU < 0.05 NM (93 m): 14688 (33.20%)
EPU < 0.1 NM (185 m): 682 (1.54%)
EPU < 1.0 NM (1852 m): 1 (0.00%)
EPU < 4 NM (7408 m): 1 (0.00%)
EPU > 10 NM or Unknown: 518 (1.17%)
--------------------------------------------------------------------------------
MOPS Version for ECAT:
MOPS Version for ECAT: '300000 lbs <= heavy aircraft' records 8561 (13.02%):
none: 0 (0.00%)
DO-260A: 129 (1.51%)
ED102A/DO-260B: 8432 (98.49%)
MOPS Version for ECAT: '75000 lbs < medium a/c < 300000 lbs' records 31672 (48.17%):
none: 0 (0.00%)
ED102/DO-260: 79 (0.25%)
DO-260A: 1281 (4.04%)
ED102A/DO-260B: 30312 (95.71%)
MOPS Version for ECAT: 'None' records 23576 (35.86%):
none: 0 (0.00%)
ED102/DO-260: 21433 (90.91%)
DO-260A: 66 (0.28%)
ED102A/DO-260B: 2077 (8.81%)
MOPS Version for ECAT: 'light aircraft <= 15500 lbs' records 1624 (2.47%):
none: 0 (0.00%)
ED102/DO-260: 3 (0.18%)
DO-260A: 5 (0.31%)
ED102A/DO-260B: 1616 (99.51%)
MOPS Version for ECAT: 'rotocraft' records 318 (0.48%):
none: 0 (0.00%)
ED102A/DO-260B: 318 (100.00%)
--------------------------------------------------------------------------------
NACp for ECAT:
NACp for ECAT: '300000 lbs <= heavy aircraft' records 8561 (13.02%):
none: 0 (0.00%)
EPU < 10 m, VEPU < 15 m: 2249 (26.27%)
EPU < 30 m, VEPU < 45 m: 4988 (58.26%)
EPU < 0.05 NM (93 m): 1274 (14.88%)
EPU < 0.1 NM (185 m): 15 (0.18%)
EPU > 10 NM or Unknown: 35 (0.41%)
NACp for ECAT: '75000 lbs < medium a/c < 300000 lbs' records 31672 (48.17%):
none: 79 (0.25%)
EPU < 10 m, VEPU < 15 m: 8799 (27.85%)
EPU < 30 m, VEPU < 45 m: 9115 (28.85%)
EPU < 0.05 NM (93 m): 12860 (40.71%)
EPU < 0.1 NM (185 m): 598 (1.89%)
EPU < 1.0 NM (1852 m): 1 (0.00%)
EPU < 4 NM (7408 m): 1 (0.00%)
EPU > 10 NM or Unknown: 219 (0.69%)
NACp for ECAT: 'None' records 23576 (35.86%):
none: 21433 (90.91%)
EPU < 10 m, VEPU < 15 m: 707 (32.99%)
EPU < 30 m, VEPU < 45 m: 669 (31.22%)
EPU < 0.05 NM (93 m): 536 (25.01%)
EPU < 0.1 NM (185 m): 19 (0.89%)
EPU > 10 NM or Unknown: 212 (9.89%)
NACp for ECAT: 'light aircraft <= 15500 lbs' records 1624 (2.47%):
none: 3 (0.18%)
EPU < 10 m, VEPU < 15 m: 1360 (83.90%)
EPU < 30 m, VEPU < 45 m: 179 (11.04%)
EPU < 0.05 NM (93 m): 18 (1.11%)
EPU < 0.1 NM (185 m): 50 (3.08%)
EPU > 10 NM or Unknown: 14 (0.86%)
NACp for ECAT: 'rotocraft' records 318 (0.48%):
none: 0 (0.00%)
EPU < 10 m, VEPU < 15 m: 280 (88.05%)
EPU > 10 NM or Unknown: 38 (11.95%)
--------------------------------------------------------------------------------
Questions & Request
For additional information or request for additional categories / editions / analysis scripts please contact the author.
Author
Helmut Puhr
Contact: [email protected]
Licenses
The source code is released under GNU GPLv3
The binary is released under Creative Commons Attribution 4.0 International (CC BY 4.0), Legal Text
Disclaimer
THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
评论