Bitcoin Forum

Bitcoin => Development & Technical Discussion => Topic started by: Gavin Andresen on October 11, 2012, 03:43:57 PM



Title: Ready for testing: 0.7.1 release candidate 1
Post by: Gavin Andresen on October 11, 2012, 03:43:57 PM
Bitcoin version 0.7.1 release candidate 1 is now available from:
  http://sourceforge.net/projects/bitcoin/files/Bitcoin/bitcoin-0.7.1/test/

This is a bug-fix minor release.

New features
------------

* Added a boolean argument to the RPC 'stop' command, if true sets
  -detachdb to create standalone database .dat files before shutting down.

* -salvagewallet command-line option, which moves any existing wallet.dat
  to wallet.{timestamp}.dat and then attempts to salvage public/private
  keys and master encryption keys (if the wallet is encrypted) into
  a new wallet.dat. This should only be used if your wallet becomes
  corrupted, and is not intended to replace regular wallet backups.

* Import $DataDir/bootstrap.dat automatically, if it exists.

Dependency changes
------------------

* Qt 4.8.2 for Windows builds

* openssl 1.0.1c

Bug fixes
---------

* Clicking on a bitcoin: URI on Windows should now launch Bitcoin-Qt properly.

* When running -testnet, use RPC port 18332 by default.

* Better detection and handling of corrupt wallet.dat and blkindex.dat files.
  Previous versions would crash with a DB_RUNRECOVERY exception, this
  version detects most problems and tells you how to recover if it
  cannot recover itself.

* Fixed an uninitialized variable bug that could cause transactions to
  be reported out of order.

* Fixed a bug that could cause occasional crashes on exit.

* Warn the user that they need to create fresh wallet backups after they
  encrypt their wallet.


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: finway on October 11, 2012, 03:49:14 PM
running for 8 hours on Windows7, seems fine.


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: Stephen Gornick on October 11, 2012, 05:24:33 PM
attempts to salvage public/private  keys and master encryption keys (if the wallet is encrypted) into a new wallet.dat.

Better detection and handling of corrupt wallet.dat and blkindex.dat files.   Previous versions would crash with a DB_RUNRECOVERY exception, this version detects most problems and tells you how to recover if it cannot recover itself.

Awesome, this will come in handy!




Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: kjj on October 12, 2012, 03:00:25 AM
In IRC, we've started working on the torrent for bootstrapping.

jgarzik and I each took our block databases and generated bootstrap.dat files from them, starting from the genesis block and going through (including) the last checkpointed block, 193,001 blocks in total.  The bootstrap.dat has ONLY the valid blocks in it (no orphans), all blocks are in sequential order, and there are is no "extra" data in the file between any blocks.

We used two totally different scripts and our own privately downloaded blockchains as the starting points.  After an initial disagreement on whether to end with the checkpointed block, or the block before the checkpoint, we now have perfect agreement of size and hash.  We expected this, because the bitcoin client verifies all of this data too, but it was still nice when it worked.

He has created a torrent file and a magnet URL, he'll be posting it and making the announcement when everything is ready.  Apparently it takes a while for the DHT system to sync up.

The file size and SHA256 hash of the file are below. 

Code:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

2491771562 a3f258e7af030165360596e4cb0b9beb24b4ce97352c22e65349b89ad5fc5d3e  bootstrap.dat
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)

iD8DBQFQd4YcXZExftMAwEgRAthpAKC0JEiOhY3/aw06BhfLDQ2P7wdUDwCgpnQr
G+GDS3h5B+gsW81I0k2S0Ew=
=ro0i
-----END PGP SIGNATURE-----


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: spiccioli on October 12, 2012, 10:19:41 AM
Bitcoin version 0.7.1 release candidate 1 is now available from:
  http://sourceforge.net/projects/bitcoin/files/Bitcoin/bitcoin-0.7.1/test/

This is a bug-fix minor release.


It dies on my 10.5.8 mac with this trap:

Code:
Process:         Bitcoin-Qt [437]
Path:            /Volumes/Bitcoin-Qt/Bitcoin-Qt.app/Contents/MacOS/Bitcoin-Qt
Identifier:      com.yourcompany.Bitcoin-Qt
Version:         ??? (???)
Code Type:       X86 (Native)
Parent Process:  launchd [76]

Interval Since Last Report:          4991 sec
Crashes Since Last Report:           1
Per-App Interval Since Last Report:  4467 sec
Per-App Crashes Since Last Report:   1

Date/Time:       2012-10-12 12:14:29.909 +0200
OS Version:      Mac OS X 10.5.8 (9L31a)
Report Version:  6
Anonymous UUID:  51EEDE10-8C88-466A-AEF6-E85F259DF368

Exception Type:  EXC_BREAKPOINT (SIGTRAP)
Exception Codes: 0x0000000000000002, 0x0000000000000000
Crashed Thread:  0

Dyld Error Message:
  Symbol not found: __ZNKSt13bad_exception4whatEv
  Referenced from: /Volumes/Bitcoin-Qt/Bitcoin-Qt.app/Contents/MacOS/../Frameworks/libboost_thread-mt.dylib
  Expected in: /usr/lib/libstdc++.6.dylib

Same problem for the 0.7.0  release.

spiccioli


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: arklan on October 12, 2012, 05:26:36 PM
running in a virtual machine windows 7 64. (took forever to get that set up cause my windows disc is old...) currently syncing the blockchain problem free.

which is excellent, cause when i upgraded to 0.7 on my actual system i launched bitcoin-qt then quickly exited it, just making sure it'd launch. after thast it wouldn't lunch without an error. having done the same thing on .7.1, it behaves properly without issues. :D



Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: P_Shep on October 12, 2012, 06:09:12 PM
Works for me on WinXP.
Received, sent, etc...


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: fornit on October 12, 2012, 07:48:55 PM
works fine under win 7 64bit for several hours with browsing through most menus, tabs and dialogs.


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: Diapolo on October 12, 2012, 07:54:53 PM
Can any Windows user, who used the 0.7.1 RC1 installer, confirm clicking URIs is working again.

Dia


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: Gavin Andresen on October 12, 2012, 07:57:10 PM
It dies on my 10.5.8 mac....

This will be a WONTFIX, unless somebody steps up to continue supporting OSX 10.5.

I'll put a big note in the release notes that we're dropping OSX 10.5 support.


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: arklan on October 13, 2012, 12:23:37 AM
Can any Windows user, who used the 0.7.1 RC1 installer, confirm clicking URIs is working again.

Dia

i need a URI to try clicking. where can i find one?


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: markm on October 13, 2012, 12:43:53 AM
We couild, however, create merged mined blockchains each using a different contender for "world's best signature algorithm", so that specuilators and hedgers can speculate and hedge regarding theoretical breakings of one or more such algorithms by picking and choosing which such chains to distribute their wealth among and in what proportions...

-MarkM-


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: Gavin Andresen on October 13, 2012, 01:05:23 AM
i need a URI to try clicking. where can i find one?
I slapped one up you can try:  http://skypaint.com/bitcoin/fauceturi.html

I think typing bitcoin:1.....?label=... etc  into the browser address bar should work for testing, too.


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: arklan on October 13, 2012, 11:37:52 PM
well, i'm still downloading the blockchain so i'm not 100% sure of this result.

but, when clicking that URI, i'm asked if i want to allow the brwoser to open a progam, and of course i allow it. then i get the windows 7 equivalent of an hourglass... then nothing happens. no error message, not attemtp to start a second instance of bitcoin-qt. nothing.

i want to try again when i've sync'd with the live blockchain, so that'll be in another 10k blocks...


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: AniceInovation on October 14, 2012, 01:28:59 AM
Hey gavin,

Just tried the -salvagewallet command, to try to recover a corrupted wallet (1 bad sector) and i got this:
http://img145.imageshack.us/img145/653/errorwh.jpg

Don't know if it was supposed to work this way


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: runeks on October 14, 2012, 04:37:51 AM
I get an error message saying:

Code:
Error initializing database environment /home/rune/.bitcoin!
To recover, BACKUP THAT DIRECTORY, then remove
everything from it except for wallet.dat.

Is this expected? I've had to download the block chain from scratch a couple of times previously when I've upgraded. Is it because it uses a newer version of BDB?


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: CharlesPonzi on October 14, 2012, 04:40:33 AM
Hey gavin,

Just tried the -salvagewallet command, to try to recover a corrupted wallet (1 bad sector) and i got this:
http://anonymouse.org/cgi-bin/anon-www.cgi/http://img145.imageshack.us/img145/653/errorwh.jpg

Don't know if it was supposed to work this way

salvagewallet  not savagewallet


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: AniceInovation on October 14, 2012, 11:26:30 AM
Done, still getting the exact same error


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: nomnomnom on October 14, 2012, 02:31:28 PM
Works fine for me on linux 64bit, I am running it the last 2 days or so.
It even recovered a wallet that I corrupted with this script (https://gist.github.com/3812689)


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: Diapolo on October 14, 2012, 07:46:33 PM
well, i'm still downloading the blockchain so i'm not 100% sure of this result.

but, when clicking that URI, i'm asked if i want to allow the brwoser to open a progam, and of course i allow it. then i get the windows 7 equivalent of an hourglass... then nothing happens. no error message, not attemtp to start a second instance of bitcoin-qt. nothing.

i want to try again when i've sync'd with the live blockchain, so that'll be in another 10k blocks...

Can you please start regedit.exe and post the value of HKEY_CLASSES_ROOT\bitcoin\shell\open\command\(default) it needs to be something like "E:\Bitcoin\Client\bitcoin-qt.exe" "%1", and the %1 at the end is the most important part. You used the 0.7.1 RC1 installer, right?

Dia


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: arklan on October 15, 2012, 06:25:19 AM
ok, regarding the URI thing: FINALLY got a successful complete sync of the main blockchain, and it worked fine. hurray!

man, virtual machines really don't like being left idle...


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: AvL42 on October 15, 2012, 10:12:21 AM
I'm test-running 0.7.1rc1 (self-compiled with libdb5.1 and without upnp on linux)
and found it drop dead already twice after sending it commands in the debug-console.

This (2nd) time it was the "move" command.
I instantly saw the mining-activity stop (cpu-meter in the panel), and the move
command didn't return a response. When it happened, I could still type more commands
in the console (though without getting an answer), but meanwhile also the debug-widget
has dropped dead and doesn't redraw anymore. I'll leave it stuck for a while and will look
back in this thread for debugging instructions before killing&restarting it...

PS: gdb says, it's stuck in pthreads_cond_wait in boost::unique_lock
  in AddressTableModel::labelForAddress(QString)
  in TransactionTableModel::lookupAddress(string,bool)



Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: runeks on October 21, 2012, 01:07:07 AM
I get an error message saying:

Code:
Error initializing database environment /home/rune/.bitcoin!
To recover, BACKUP THAT DIRECTORY, then remove
everything from it except for wallet.dat.

Is this expected? I've had to download the block chain from scratch a couple of times previously when I've upgraded. Is it because it uses a newer version of BDB?
Here's the solution to this issue (for the record):

Quote
If you were running on Linux with a version that might have been compiled
with a different version of Berkeley DB (for example, if you were using an
Ubuntu PPA version), then run the old version again with the -detachdb
argument and shut it down; if you do not, then the new version will not
be able to read the database files and will exit with an error.
https://bitcointalk.org/index.php?topic=119277.0


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: kjj on October 21, 2012, 02:29:09 AM
I'm test-running 0.7.1rc1 (self-compiled with libdb5.1 and without upnp on linux)
and found it drop dead already twice after sending it commands in the debug-console.

This (2nd) time it was the "move" command.
I instantly saw the mining-activity stop (cpu-meter in the panel), and the move
command didn't return a response. When it happened, I could still type more commands
in the console (though without getting an answer), but meanwhile also the debug-widget
has dropped dead and doesn't redraw anymore. I'll leave it stuck for a while and will look
back in this thread for debugging instructions before killing&restarting it...

PS: gdb says, it's stuck in pthreads_cond_wait in boost::unique_lock
  in AddressTableModel::labelForAddress(QString)
  in TransactionTableModel::lookupAddress(string,bool)

Interesting.  One of my nodes has been doing the same thing in response to the RPC move command.


Title: Re: Ready for testing: 0.7.1 release candidate 1
Post by: AvL42 on October 21, 2012, 08:10:07 AM
indeed, also happens with RPC, not only with debug console. Not sure, though, if any other command is affected, too.