[ISmokezU]

Quote:

Originally Posted by vasyok28 Hello! After omission of MySQL is disconnected from the server, and restores the connection is not re-encoding. What solution ?

It provides a good logging system (IMO) Check it, and post it.

[Cypress]

pawn Код:

"SELECT `password`, `password_salt`, UNIX_TIMESTAMP(`last_active`) FROM `players` WHERE `username` = '%e' LIMIT 1"

pawn Код:

[12:10:16] [plugins/mysql] cache_get_value_name: field 'last_active' not found

Don't know what's the issue here. Also how would I fetch this data into a variable?

Edit: fixed it.

[f0Re3t]

hi
I would like more information on working with ORM scenarios
for example, I have a table with the accounts of the players and need to do so you can't register with the same email address
for usual queries it is easy to do, but what about ORM ?

[liquor]

Anyone know if there's a problem with this and easy-mysql? Or Windows 10? I get error 19 but I've 10,000-times-checked files, folder, includes and recompiled it and tried earlier versions of this plugin...

[IstuntmanI]

I decided to upgrade to the latest MySQL version (after a problem with MySQL public calls in sampGDK). I converted hundreds of lines without any of those regex codes and it took me 2 days (a total of few hours) on a pretty big gamemode (also writing a sampGDK include for MySQL plugin). I see that R40+ is much more organised, passing return values by reference isn't that bad, it actually looks way better when there is a lot of fetch of strings, ints and floats one after one. Multiple result sets, cleanups and other things make R40 worthy of all these conversions.

By the way, in GitHub "a_mysql.inc" is named "a_mysql.inc.in" and it contains this:

Quote:

/** * MySQL plugin @MYSQL_PLUGIN_VERSION@ */

you should name it "a_mysql.inc" and try to fix that line. I guess that this is a problem from compiling and that you are aware of it ?

-----------------------------

I have a problem with MySQL plugin in sampGDK:
(I also tried to run an empty gamemode just with the MySQL connection to a MySQL server and it connected fine, without any crash or error, with the regular server files structure.)

I tried to invoke MySQL's natives in a sampGDK gamemode but the whole server crashes at mysql_connect. Here's how I tried to "convert" them using invokeNative:

Код:

int mysql_init_options( )
{
	static AMX_NATIVE Native = sampgdk::FindNative( "mysql_init_options" );
	return sampgdk::InvokeNative( Native, "" );
}

template < class Args >
//not sure if the params parameter passing is correct
bool mysql_set_option( int option_id, E_MYSQL_OPTION type, Args * params )
{
	static AMX_NATIVE Native = sampgdk::FindNative( "mysql_set_option" );
	return !!sampgdk::InvokeNative( Native, "iir", option_id, ( int ) type, params );
}

int mysql_connect( const char host[ ], const char user[ ], const char password[ ], const char database[ ], int option_id = 0 )
{
	static AMX_NATIVE Native = sampgdk::FindNative( "mysql_connect" );
	return sampgdk::InvokeNative( Native, "ssssi", host, user, password, database, option_id );
}

and here's how I call them:

Код:

mysql_log( ALL );

int lConnectOptions = mysql_init_options( ), lPort = 3306, lPoolSize = 4;
bool lAutoReconnect = true;

sampgdk::logprintf( "OGMI - 1" );

Plugins::MySQL::set_option( lConnectOptions, SERVER_PORT, &lPort );
Plugins::MySQL::set_option( lConnectOptions, AUTO_RECONNECT, &lAutoReconnect );
Plugins::MySQL::set_option( lConnectOptions, POOL_SIZE, &lPoolSize );

sampgdk::logprintf( "OGMI - 2 - Options ID %d", lConnectOptions );

mysql_connect( MYSQL_HOST, MYSQL_USER, MYSQL_PASSWORD, MYSQL_DATABASE, lConnectOptions );

sampgdk::logprintf( "OGMI - 3" );

This is what it prints:

Quote:

Originally Posted by server_log.txt OGMI - 1 OGMI - 2 - Options ID 1

Which means that it crashes on the mysql_connect line. (I even tried without any option_id parameter set and it is still crashing) The invokeNative call looks fine to me.

After which it crashes (info by crashdetect plugin) (I don't know if this helps too much):

Код:

[debug] Server crashed due to an unknown error
[debug] Native backtrace:
[debug] #0 5108f48e in ?? () from log-core.dll
[debug] #1 5108f607 in ?? () from log-core.dll
[debug] #2 5108a975 in ?? () from log-core.dll
[debug] #3 77bd7cd0 in ?? () from plugins\mysql.DLL
[debug] #4 77beec64 in ?? () from plugins\mysql.DLL
[debug] #5 77658a40 in ?? () from plugins\MYGAMEMODE.DLL
[debug] #6 776586a8 in ?? () from plugins\MYGAMEMODE.DLL
[debug] #7 776507d4 in ?? () from plugins\MYGAMEMODE.DLL
[debug] #8 77703492 in ?? () from plugins\MYGAMEMODE.DLL
[debug] #9 7773ff5b in ?? () from plugins\MYGAMEMODE.DLL
[debug] #10 77834cdd in ?? () from plugins\MYGAMEMODE.DLL
[debug] #11 77648cd3 in ?? () from plugins\MYGAMEMODE.DLL
[debug] #12 7765784f in ?? () from plugins\MYGAMEMODE.DLL
[debug] #13 7764ae43 in ?? () from plugins\MYGAMEMODE.DLL
[debug] #14 0046f679 in ?? () from samp-server.exe
[debug] #15 65646f6d in ?? () from samp-server.exe
[debug] #16 5a472f73 in ?? () from samp-server.exe
[debug] #17 61625f53 in ?? () from samp-server.exe
[debug] #18 612e6573 in ?? () from samp-server.exe
[debug] Registers:
[debug] EAX: 00000000 EBX: 00000001 ECX: 00000001 EDX: 00000000
[debug] ESI: 029bae90 EDI: 00000001 EBP: 61625f53 ESP: 00515c0c
[debug] EIP: 612e6573 EFLAGS: 00010246
[debug] Stack:
[debug] ESP+00000000: 0000786d 00000000 00000000 00000000
[debug] ESP+00000020: 00000000 00000000 00000000 00000000
[debug] ESP+00000040: 00000000 00000000 00000000 00000000
[debug] ESP+00000060: 00000000 00000000 00000000 00000000
[debug] ESP+00000080: 00000000 00000000 00000000 00000000
[debug] ESP+000000a0: 00000000 00000000 00000000 00000000
[debug] ESP+000000c0: 00000000 00000000 00000000 00000000
[debug] ESP+000000e0: 00000000 00000000 00000000 00000000
[debug] ESP+00000100: 00000000 00000000 00000000 00000000
[debug] ESP+00000120: 00000065 00000000 00000000 00000000
[debug] ESP+00000140: 00000000 00000000 00000000 00000000
[debug] ESP+00000160: 00000006 00000002 00000001 023c1850
[debug] ESP+00000180: 00000000 00516020 00000000 00000000
[debug] ESP+000001a0: 00000000 ffffffff ffffffff 00000000
[debug] ESP+000001c0: ffffffff 00000000 00000000 00000000
[debug] ESP+000001e0: 00000000 00000000 00000fa0 ffffffff
[debug] ESP+00000200: 00000fa0 ffffffff ffffffff 00000000
[debug] ESP+00000220: ffffffff 00000000 00000000 00000000
[debug] ESP+00000240: 00000000 00000000 00000fa0 ffffffff
[debug] ESP+00000260: 00000fa0 ffffffff ffffffff 00000000
[debug] ESP+00000280: ffffffff 00000000 00000000 00000000
[debug] ESP+000002a0: 00000000 00000000 00000fa0 ffffffff
[debug] ESP+000002c0: 00000fa0 ffffffff ffffffff 00000000
[debug] ESP+000002e0: ffffffff 00000000 00000000 00000000
[debug] ESP+00000300: 00000000 00000000 00000000 00000000
[debug] ESP+00000320: 00000000 00000000 00000000 ffffff88
[debug] ESP+00000340: 00610064 00640072 00540020 006d0069
[debug] ESP+00000360: 00000000 00000000 00000000 00000000
[debug] ESP+00000380: 00000000 00540047 00200042 00610044
[debug] ESP+000003a0: 006d0069 00000065 00000000 00000000
[debug] ESP+000003c0: 00000000 00030000 00050000 00000003
[debug] ESP+000003e0: 00000001 00000000 00000000 00000000
[debug] Loaded modules:
[debug] 00400000 - 00519000 samp-server.exe
[debug] 77240000 - 773c3000 ntdll.dll
[debug] 75bb0000 - 75c90000 KERNEL32.DLL
[debug] 76590000 - 76731000 KERNELBASE.dll
[debug] 74680000 - 75a59000 SHELL32.dll
[debug] 76c00000 - 76cbe000 msvcrt.dll
[debug] 73fb0000 - 73fe6000 cfgmgr32.dll
[debug] 75e30000 - 7639e000 windows.storage.dll
[debug] 73d80000 - 73f91000 combase.dll
[debug] 76f90000 - 77070000 ucrtbase.dll
[debug] 75c90000 - 75d51000 RPCRT4.dll
[debug] 73d00000 - 73d1e000 SspiCli.dll
[debug] 73cf0000 - 73cfa000 CRYPTBASE.dll
[debug] 742c0000 - 7431a000 bcryptPrimitives.dll
[debug] 75a60000 - 75aa1000 sechost.dll
[debug] 76740000 - 76785000 powrprof.dll
[debug] 74240000 - 742b7000 advapi32.dll
[debug] 76d20000 - 76d66000 shlwapi.dll
[debug] 74520000 - 7454b000 GDI32.dll
[debug] 77070000 - 771cb000 gdi32full.dll
[debug] 763a0000 - 764ff000 USER32.dll
[debug] 74660000 - 74675000 win32u.dll
[debug] 75e20000 - 75e2d000 kernel.appcore.dll
[debug] 74320000 - 743a8000 shcore.dll
[debug] 76cc0000 - 76ccf000 profapi.dll
[debug] 57ab0000 - 57ab8000 WSOCK32.dll
[debug] 70ab0000 - 70ad4000 WINMM.dll
[debug] 771d0000 - 77233000 WS2_32.dll
[debug] 00670000 - 00693000 WINMMBASE.dll
[debug] 75d60000 - 75d85000 IMM32.DLL
[debug] 510a0000 - 510f0000 crashdetect.DLL
[debug] 77ce0000 - 77d49000 MSVCP100.dll
[debug] 77c20000 - 77cdf000 MSVCR100.dll
[debug] 77bc0000 - 77c15000 mysql.DLL
[debug] 77ae0000 - 77bbe000 libmariadb.dll
[debug] 51080000 - 5109e000 log-core.dll
[debug] 65100000 - 6516d000 MSVCP140.dll
[debug] 650e0000 - 650f5000 VCRUNTIME140.dll
[debug] 70070000 - 70082000 napinsp.dll
[debug] 70050000 - 70066000 pnrpnsp.dll
[debug] 70030000 - 70044000 NLAapi.dll
[debug] 73b30000 - 73b7e000 mswsock.dll
[debug] 6fba0000 - 6fc1d000 DNSAPI.dll
[debug] 74220000 - 74227000 NSI.dll
[debug] 73c00000 - 73c2f000 IPHLPAPI.DLL
[debug] 73aa0000 - 73aac000 winrnr.dll
[debug] 73010000 - 73021000 wshbth.dll
[debug] 73b80000 - 73b88000 rasadhlp.dll
[debug] 77a80000 - 77ad7000 streamer.DLL
[debug] 10000000 - 1000d000 sscanf.DLL
[debug] 77e00000 - 77e0d000 gvar.DLL
[debug] 77630000 - 77a73000 MYGAMEMODE.DLL
[debug] 73a80000 - 73a96000 sslsp105.dll
[debug] 76e10000 - 76f8d000 CRYPT32.dll
[debug] 75e10000 - 75e1e000 MSASN1.dll
[debug] 70180000 - 701c9000 fwpuclnt.dll
[debug] 71cd0000 - 71ceb000 bcrypt.dll

Here are the logs generated by log-core:

Quote:

Originally Posted by logs\log-core.log [ERROR] exception 0XC0000005 (ACCESS_VIOLATION) from Vectored Exception Handler catched; shutting log-core down

Quote:

Originally Posted by logs\errors.log [log-core] exception 0XC0000005 (ACCESS_VIOLATION) from Vectored Exception Handler catched; shutting log-core down

Quote:

Originally Posted by logs\plugins\mysql.log [INFO] changed log level from 'warning, error' to 'debug, info, warning, error' [DEBUG] mysql_init_options() [INFO] Options instance with id '1' successfully created. [DEBUG] mysql_init_options: return value: '1' [DEBUG] mysql_set_option(134265809, 5439571) [DEBUG] mysql_set_option: return value: '1' [DEBUG] mysql_set_option(134265809, 5439571) [DEBUG] mysql_set_option: return value: '1' [DEBUG] mysql_set_option(134265809, 5439571) [DEBUG] mysql_set_option: return value: '1'

MySQL-related Server files structure:

Код:

samp-server.exe
log-core.dll
libmariadb.dll
[..]
/plugins/
    mysql.dll
    MYGAMEMODE.dll

Loading order:

Quote:

Originally Posted by server.cfg plugins crashdetect mysql [..] MYGAMEMODE

EDIT: If I remove the mysql_connect line, it crashes exactly the same at the first mysql_query call. Also, I find it weird that mysql_set_option has the same values in the debug info ("[DEBUG] mysql_set_option(134265809, 5439571)" above). I tried without initing options and setting them and it still crashes.

EDIT 2: Tried to use ProcDump (I have 0 experience in it): "Exception Information: The thread tried to read from or write to a virtual address for which it does not have the appropriate access." (I also tried without "const" arguments) It points me to this disassembly:


EDIT 3: I managed to configure MySQL Plugin's project and recompile it, I added few debug prints and I got to the conclusion that it is crashing on line "logger.LogNativeCall(amx, func, params_format);" from "CLog.cpp" .

Few more debugs:

samplog_LogNativeCall, right before "for" loop:

Quote:

Debug: module 'plugins/mysql' - name 'mysql_connect' - params_format 'ss*sd' - params '00825678'

case 's' from that switch:

Quote:

Debug: case 's': params_format: 's' - current param: '134248417' - amx '02C69BC0' // if you ask me, I don't think the 'current param' looks fine

then it crashes at

Код:

fmt_msg << '"' << amx_GetCppString(amx, current_param) << '"';

in which amx_GetCppString crashes at

Код:

amx_GetString(&string[0], addr, 0, len + 1);

before which I had a debug print:

Quote:

Debug: param '134248417' - addr '00000000' - len '0'

from amxplugin2.cpp.

This means that it gets the address '00000000' from this line:

Код:

amx_GetAddr(amx, param, &addr);

which probably means that "param" is not set correctly ?

EDIT 4: Well, setting logging to NONE works without any crash (obviously). You should try to fix this, if you have any idea about why it happens and if you can reproduce it.

----------------

EDIT 5: Ok, so I got past those crashes, by disabling the logs. But there's the old problem, which I posted in the sampGDK plugin topic: no callback is called. I replaced the log from "invalid amx instance" with a logprintf, thinking that this is the problem which is stopping callbacks to execute and I was right, I receive "invalid amx instance" for every callback called from my gamemode plugin. I guess that this is because my gamemode is a plugin and not a regular filterscript/gamemode. I simply removed that check for "valid amx instances" from CCallback.cpp and it looks like OnPublicCall is called for them. (I see in AmxLoad this line: "samplog::RegisterAmx(amx);". I wouldn't be surprised if this would be the problem with the logging too. So AmxLoad isn't called for a sampGDK plugin fakeamx in other plugins, if I'm correct)

Can you please do something to fix these things in R41-3 or R42 ? Keeping logs disabled and having those checks removed aren't exactly good solutions. Maybe some function to manually call AmxLoad from MySQL Plugin to add the fakeamx to the valid instances ?

[maddinat0r]

The problem is, that the log core uses a function (GetNativeParamsPtr) to get the parameters pointer which was passed to a native function. Parameters are always pushed onto the stack, that's why that helper function correctly returns the native functions parameters array.
However, natives invoked through sampgdk's InvokeNative function don't push the passed parameters onto the stack, but on the heap. Don't ask me why, but I guess Zeex had a good reason to do this.
A proper solution would be to pass the params pointer directly to the log core instead of retrieving it implicitly. However, this would involve an API change, which in turn would force me to release updates to 3 different plugins. I'll probably commit the fix, but I won't do the releases, since I don't have the time for that.
You'd also just have to disable debug logging, as this bug only impacts native-call logging, which is only enabled when debug logging is enabled.

Regarding the callbacks: that AMX validation is there in case someone sends a threaded query, which takes very long to execute, from within a filterscript and unloads that filterscript while it's still executing. The AMX structure associated with that callback is now invalid, but it's still stored in the internal MySQL plugin callback structures to call the callback when that query finishes. To make sure the plugin doesn't use that invalid AMX instance, I keep a list of valid AMX instances passed by the AmxLoad/AmxUnload callbacks.
I won't add a function to manually add a AMX instance to that list, since this is more of a hack than a proper solution. I don't see a good solution to this problem here, I'm open for any suggestions though.

As I said in the sampgdk thread, I'd really suggest you incorporate the plugin source into your project. You only have to copy all the source files minus some unnecessary ones (main.cpp, CCallback.*, CDispatcher.*, COrm.*, natives.*, sdk.hpp, version.hpp.in), (maybe) remove some include directives, add the Boost headers to your include path (no need to compile the whole Boost library!) and link with the MySQL C connector. If you want to use the log core, you'd have to link to it too (you can download the developer package of the log core on GitHub). You could then use the logging functionality in your gamemode for your systems too. Write me a PM if you get stuck somewhere or need more detailed instructions on specific things.

[IstuntmanI]

Quote:

Originally Posted by maddinat0r Regarding the callbacks: that AMX validation is there in case someone sends a threaded query, which takes very long to execute, from within a filterscript and unloads that filterscript while it's still executing. The AMX structure associated with that callback is now invalid, but it's still stored in the internal MySQL plugin callback structures to call the callback when that query finishes. To make sure the plugin doesn't use that invalid AMX instance, I keep a list of valid AMX instances passed by the AmxLoad/AmxUnload callbacks. I won't add a function to manually add a AMX instance to that list, since this is more of a hack than a proper solution. I don't see a good solution to this problem here, I'm open for any suggestions though.

I was also thinking about adding a function to manually add the AMX, you read my mind. You could loop through all queries which are being processed at amxUnload and if the callback is in the unloaded amx you could remove the callback, so it won't execute any. In this way you could remove the valid AMX instances list checking and the whole list.

Quote:

Originally Posted by maddinat0r As I said in the sampgdk thread, I'd really suggest you incorporate the plugin source into your project. You only have to copy all the source files minus some unnecessary ones (main.cpp, CCallback.*, CDispatcher.*, COrm.*, natives.*, sdk.hpp, version.hpp.in), (maybe) remove some include directives, add the Boost headers to your include path (no need to compile the whole Boost library!) and link with the MySQL C connector. If you want to use the log core, you'd have to link to it too (you can download the developer package of the log core on GitHub). You could then use the logging functionality in your gamemode for your systems too. Write me a PM if you get stuck somewhere or need more detailed instructions on specific things.

This is a change way too big for me, for now. I'm currently migrating the whole gamemode and converting MySQL Plugin would take a lot, as I use it pretty intensive. I also have other big problems with migrating. Using invokeNative is a much better alternative for now.

Damn it, pointers.

[Kaperstone]

Quote:

Originally Posted by maddinat0r I've meant the Discord plugin when I wrote about updating "it". And Jesus Christ you don't have to rewrite anything, use any version you're comfortable with. Nobody is forcing you to use the latest version.

Lol, but if you want the newest toys, you gotta stay up to date.


Also, I ain't sure, but as far as I know/understood, R+40 has some crash fixes which R30 do not have (?)

[maddinat0r]

Quote:

Originally Posted by Kaperstone Lol, but if you want the newest toys, you gotta stay up to date. Also, I ain't sure, but as far as I know/understood, R+40 has some crash fixes which R30 do not have (?)

No, that's not true. R39-6 is as stable as R41-3.

[Kaperstone]

Quote:

Originally Posted by maddinat0r No, that's not true. R39-6 is as stable as R41-3.

So is there an article or a doc to read about the differences between the two.

[Mellnik]

maddinat0r did a great job maintaining this plugin. Not a single crash back then when I was still running a server. The new features like orm also were very helpful.
We should thank him for his work.

[Maceo7]

Very good plugins, it's a pity to have an complete query when we use orm update whereas we only update one property.

[Kaperstone]

Quote:

Originally Posted by Maceo7 Very good plugins, it's a pity to have an complete query when we use orm update whereas we only update one property.

I am not sure what you're saying, what is the issue the plugin features a function to send rich queries?

[Maceo7]

Quote:

Originally Posted by Kaperstone I am not sure what you're saying, what is the issue the plugin features a function to send rich queries?

if you update only one field, ORM will generate a query and update all your fields even if 90% doesn't change.

[ISmokezU]

Code:

[06:22:44] [debug] Server crashed due to an unknown error
[06:22:44] [debug] Native backtrace:
[06:22:44] [debug] #0 68d1f1d3 in ?? () from log-core.dll
[06:22:44] [debug] #1 68d12214 in ?? () from log-core.dll
[06:22:44] [debug] #2 68d12440 in ?? () from log-core.dll
[06:22:44] [debug] #3 68d11c77 in ?? () from log-core.dll
[06:22:44] [debug] #4 68d13d61 in ?? () from log-core.dll
[06:22:44] [debug] #5 68d64532 in ?? () from plugins\mysql.DLL
[06:22:44] [debug] #6 00469ab5 in ?? () from samp-server.exe
[06:22:44] [debug] #7 65646f6d in ?? () from samp-server.exe
[06:22:44] [debug] #8 414c2f73 in ?? () from samp-server.exe
[06:22:44] [debug] #9 524e434d in ?? () from samp-server.exe
[06:22:44] [debug] #10 786d612e in ?? () from samp-server.exe

This has never happened to me before the server was working fine all the time when I tried starting it again it gave these errors.

It's clearly saying the log-core.dll is the problem but I don't see how I didn't do or change anything.

I tried downgrading and updating everything and it's still happening could someone help me out here? Thanks.

Edit: Problem was solved.

[Perdun]

Hello. I have a problem. We moved the mod to MySQL R41-2, but my hosting does not support CentOs7. Can you compile the plug-in under CentOS 6

[Jeroen52]

Quote:

Originally Posted by Perdun Hello. I have a problem. We moved the mod to MySQL R41-2, but my hosting does not support CentOs7. Can you compile the plug-in under CentOS 6

MySQL plugin R39 is the latest version supported for MySQL 6.

[maddinat0r]

Quote:

Originally Posted by Maceo7 if you update only one field, ORM will generate a query and update all your fields even if 90% doesn't change.

I could implement that, however this would require me to cache all values of the registered PAWN variables and always check if the value changed before generating the query. This likely has a bigger performance-impact than generating a longer query.

Quote:

Originally Posted by Freedom. Any ideas? I am using latest version. VPS: Centos 7 x64 Other plugins: SKY streamer sscanf Whirlpool Code: *** Error in `./samp03svr': free(): corrupted unsorted chunks: 0x0e13ee78 *** ======= Backtrace: ========= /lib/libc.so.6(+0x753bd)[0xb74e13bd] /lib/libstdc++.so.6(_ZdlPv+0x1f)[0xb76d197f] /lib/libstdc++.so.6(_ZNSs4_Rep10_M_destroyERKSaIcE+0x1b)[0xb7738edb] /lib/libstdc++.so.6(+0xaff20)[0xb7738f20] plugins/mysql.so(+0x34e7a)[0xb743ee7a] plugins/mysql.so(+0x43490)[0xb744d490] plugins/mysql.so(_ZNSt14_Function_base13_Base_managerISt5_BindIFSt7_Mem_fnIM6CQueryFvvEESt10shared_ptrIS3_EEEE10_M_managerERSt9_Any_dataRKSC_St18_Manager_operation+0x120)[0xb7427900] plugins/mysql.so(+0x1e50d)[0xb742850d] plugins/mysql.so(ProcessTick+0x24)[0xb74517d4] ./samp03svr[0x80d1ce2] ./samp03svr[0x80aef6c] ./samp03svr[0x80aa13a] /lib/libc.so.6(__libc_start_main+0xf3)[0xb74859a3] ./samp03svr[0x804b4e1]

Can you reproduce this? If yes, enable debug logging, reproduce that crash and post the log file.

[JasonRiggs]

Uhm, Link for MySQL Library installation?

[Freedom.]

This occurs again.

Code:

*** Error in `./samp03svr': malloc(): memory corruption: 0xb75f6409 ***
======= Backtrace: =========
/lib/libc.so.6(+0x76a03)[0xb74b2a03]