Pages:
Author

Topic: [ANN][SILKCOIN][DARKSILK PREANN][TESTING STORMNODES/DarkSIlk MARKET/I2P] - page 90. (Read 289507 times)

newbie
Activity: 28
Merit: 0
Oh yea...without commenting out the debug assertions for the hashGenesisBlock, the wallet does not launch on my fresh build.

DarkSilk: ../DarkSilk/src/chainparams.cpp:86: CMainParams::CMainParams(): Assertion `hashGenesisBlock == uint256("")' failed.
Aborted

I'm running 3.13.0-24-generic #47-Ubuntu (Linux Mint 17) and building with Qt 5.3.1 (GCC 4.6.1, 64 bit).
I'll check for updates and re-build to see if that makes a difference.  It's not a show stopper by any means.  

That is strange.

I created a clean Ubuntu (15.04) VM and cloned the repo. Followed my ReadMe to the letter (to make sure I hadn't missed anything) and it built and ran perfectly first time. Both daemon and Qt.

Check you have these deps: libpng-dev qrencode libqrencode-dev

You should not have to add qrencode.h to the source.
newbie
Activity: 25
Merit: 0
Oh yea...without commenting out the debug assertions for the hashGenesisBlock, the wallet does not launch on my fresh build.

DarkSilk: ../DarkSilk/src/chainparams.cpp:86: CMainParams::CMainParams(): Assertion `hashGenesisBlock == uint256("")' failed.
Aborted

I'm running 3.13.0-24-generic #47-Ubuntu (Linux Mint 17) and building with Qt 5.3.1 (GCC 4.6.1, 64 bit).
I'll check for updates and re-build to see if that makes a difference.  It's not a show stopper by any means. 
newbie
Activity: 25
Merit: 0
Did you build from a fresh clone?  I had to add "qrencode.h" to ./src/qt/ to build successfully. 
Maybe you have a local reference that the repo is missing. 

If I find the issue, I'll fork your repo and do a pull request.
newbie
Activity: 28
Merit: 0
Since this is not live (i.e. no hashGenesisBlock defined), the wallet crashes on assert(hashGenesisBlock == uint256("")); when trying to launch.  Do you have a block chain that you're using for dev purposes?  I'd like to connect to the same BC that you're using so we can see the same transactions et al.

It shouldn't crash on assert(hashGenesisBlock ==uint256("")); as it is present:

https://github.com/SCDeveloper/DarkSilk/blob/master/src/chainparams.cpp#L95

I have just cloned and built from the same repo.

https://github.com/SCDeveloper/DarkSilk



If you find the issue could you pull request and I will merge. Believe there is a 5/6 hour difference between us and it's getting late here now.
newbie
Activity: 28
Merit: 0
Since this is not live (i.e. no hashGenesisBlock defined), the wallet crashes on assert(hashGenesisBlock == uint256("")); when trying to launch.  Do you have a block chain that you're using for dev purposes?  I'd like to connect to the same BC that you're using so we can see the same transactions et al.

It shouldn't crash on assert(hashGenesisBlock ==uint256("")); as it is present:

https://github.com/SCDeveloper/DarkSilk/blob/master/src/chainparams.cpp#L95

I have just cloned and built from the same repo.

https://github.com/SCDeveloper/DarkSilk

newbie
Activity: 25
Merit: 0
Since this is not live (i.e. no hashGenesisBlock defined), the wallet crashes on assert(hashGenesisBlock == uint256("")); when trying to launch.  Do you have a block chain that you're using for dev purposes?  I'd like to connect to the same BC that you're using so we can see the same transactions et al.
newbie
Activity: 28
Merit: 0
Yep...looks like it either does not need to be free'd because it was not explicitly not malloc'd, or it has already has been free'd earlier (maybe at some other method if it's a global object).  
When does this error occur?  Same place in the code or just at random times?

I'll clone your repo and take a look.

Same time.

Either when you connect via RPC to mine and send your first hashes at the wallet, or when it finds a block.

I have a feeling it is definitely RPC related.
newbie
Activity: 25
Merit: 0
Yep...looks like it either does not need to be free'd because it was not explicitly not malloc'd, or it has already has been free'd earlier (maybe at some other method if it's a global object). 
When does this error occur?  Same place in the code or just at random times?

I'll clone your repo and take a look.
newbie
Activity: 28
Merit: 0
An invalid pointer is not a memory leak;it's an issue with memory management.  

Chances are you're trying to free() malloc'd memory that has already been free'd.
You could also be trying to free() memory that has not actually been malloc'd.  If the ptr is null, free() won't do anything, but if the ptr points to memory that has not been explicitly malloc'd, it will cause 'undetermined' behavior because you're trying to release memory that belongs to a different handle.  

Have you tried to put a try catch around the problem areas and do a stack dump to log what's going on?  You should be able to narrow down where in the code the issue is happening fairly easily.

I am trying to locate the problem area.

*** Error in `./DarkSilk': free(): invalid pointer: 0x00007f422c004630 ***
newbie
Activity: 25
Merit: 0
An invalid pointer is not a memory leak;it's an issue with memory management. 

Chances are you're trying to free() malloc'd memory that has already been free'd.
You could also be trying to free() memory that has not actually been malloc'd.  If the ptr is null, free() won't do anything, but if the ptr points to memory that has not been explicitly malloc'd, it will cause 'undetermined' behavior because you're trying to release memory that belongs to a different handle. 

Have you tried to put a try catch around the problem areas and do a stack dump to log what's going on?  You should be able to narrow down where in the code the issue is happening fairly easily.
newbie
Activity: 28
Merit: 0
I'm doing well, thanks for asking.

Just trying to get my head around where we are currently...so, my questions still stand.  

Are there different code bases between Windows and Linux for the DarkSilk wallet?  If so, does the Windows version use Microsoft Visual C++ versus the C++98 that the SilkCoin wallet was based on (pre-fork)?

No I was going to insert my C++98 code that is forked from Cravecoin into Visual Studio and run a memory leak tool to check the pages and pages of code.

I've tried the Valgrind approach with no success.
newbie
Activity: 25
Merit: 0
I'm doing well, thanks for asking.

Just trying to get my head around where we are currently...so, my questions still stand. 

Are there different code bases between Windows and Linux for the DarkSilk wallet?  If so, does the Windows version use Microsoft Visual C++ versus the C++98 that the SilkCoin wallet was based on (pre-fork)?
newbie
Activity: 28
Merit: 0
Which version of Visual Studio are you using and when did we start using different code based for Linux and Windows?  If there is a memory leak in the Windows wallet, it's almost guaranteed to be there in the Linux wallet as well; unless you have a closed source Windows version with a different code base.  Do you have a Microsoft Visual C++ version???

When I was building the Market API and Polling functions in the wallet, everything was built on the same code base using QT Creator...even in Windows.



I haven't actually built a Windows wallet yet. I have the code for DarkSilk at https://GitHub.com/SCDeveloper/DarkSilk

When I point my GPU at the wallet it instantly crashes or it will mine and crash upon finding a block.

It did this for a while and now it is mining ok. However the crash is caused by an invalid pointer.

So my plan was to get a Windows environment built so I can install Virtual Studio and then find this memory leak.

Unless you want to assist?
newbie
Activity: 28
Merit: 0
Which version of Visual Studio are you using and when did we start using different code based for Linux and Windows?  If there is a memory leak in the Windows wallet, it's almost guaranteed to be there in the Linux wallet as well; unless you have a closed source Windows version with a different code base.  Do you have a Microsoft Visual C++ version???

When I was building the Market API and Polling functions in the wallet, everything was built on the same code base using QT Creator...even in Windows.



And breath.

I am talking about the DarkSilk source Wink
newbie
Activity: 25
Merit: 0
Which version of Visual Studio are you using and when did we start using different code based for Linux and Windows?  If there is a memory leak in the Windows wallet, it's almost guaranteed to be there in the Linux wallet as well; unless you have a closed source Windows version with a different code base.  Do you have a Microsoft Visual C++ version???

When I was building the Market API and Polling functions in the wallet, everything was built on the same code base using QT Creator...even in Windows.

newbie
Activity: 28
Merit: 0
I would like to see if it is my VM so....

Clone the repo.

Follow the readme to the letter.

Here is a conf:

rpcallowip=*
rpcuser=USERNAME
rpcpassword=PASSWORD
rpcport=31500
server=1
listen=1
daemon=1
testnet=0
addnode=74.5.128.205
addnode=81.98.168.158

When you direct a miner at the wallet if it crashes then you have experienced what I have.

This will mean to me that it isn't environment and is indeed the wallet.

Let me know your findings.
newbie
Activity: 28
Merit: 0
ATM only the Linux source is up online.

Currently I believe I have a memory leak, which until I have a Windows build environment I cannot debug easily.

I will be getting a Windows environment built as quickly as possible. At the moment I am also building a 10.8 MacOSX environment as well.

Once Windows is sorted I will install Visual Studio and locate the memory leak using Visual Leak Detector for Visual C++
https://visualstudiogallery.msdn.microsoft.com/7c40a5d8-dd35-4019-a2af-cb1403f5939c

Once these are all done I can also build Windows and Mac wallets.
full member
Activity: 232
Merit: 100
To get involved in the testnet of DarkSilk please PM me.

PM'd testnet me windows bro!
legendary
Activity: 1316
Merit: 1481
too much technical speaking around I don't understand much.
Waiting peacefully for the next realease.

Nothing more at the moment.

 Wink
hero member
Activity: 896
Merit: 501
   
dimitry did you say you messaged dobbs or dobbs messaged you? .2 seems fair


No I just stumbled across Dobbs/Bob and I dont know but I have attraction to it maybe its in the name Wink

For review, "Earlz"  does the review I dont know if he have anything to do with Dobbs, Besides the review part!


Also I contacted Earlz! I think there is alot of doubters in crypto and when someone like Earlz can give a positive
that will bring good vibes for those


Pages:
Jump to: