Bitcoin Forum

Bitcoin => Bitcoin Discussion => Topic started by: Gavin Andresen on April 20, 2011, 07:12:20 PM



Title: Please help test: Bitcoin version 0.3.21 release candidate
Post by: Gavin Andresen on April 20, 2011, 07:12:20 PM
Linux and Windows binary releases are at:
  https://sourceforge.net/projects/bitcoin/files/Bitcoin/bitcoin-0.3.21/

Changes and new features include:
  • Support for Universal Plug and Play to open a port for incoming connections (off by default, turn it on using the -upnp=1 command-line switch or the checkbox on the Options dialog box).
  • Sending and displaying arbitrary precision amounts of bitcoins (you can send exactly 1.0001 bitcoins if you like).  Sending less than 0.01 bitcoins requires a 0.01 bitcoin fee, however.
  • New rpc command "sendmany" to send bitcoins to more than one person in a single transaction (already being used by some of the mining pools for payouts).
  • Several bug fixes, including a serious intermittent bug that would sometimes cause bitcoind to stop accepting rpc requests.

If you find bugs, report them at: https://github.com/bitcoin/bitcoin/issues


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: Raulo on April 20, 2011, 08:01:10 PM
I have been using the git HEAD version for some time on desktop and for mining and have not encountered any problems but I'm pretty sure this issue would cause a lot of confusion and a deluge of questions from new users:
https://github.com/bitcoin/bitcoin/issues/170

I understand the reason for the change (although these transactions will become confirmed after some time or if the block is less than 3 kB which happens from time to time) but the error dialog is very confusing.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: trentzb on April 20, 2011, 09:10:44 PM
Is this thread the right place to discuss issues with RC?

What version of MiniUPnP is intended to be supported? There seem to be some changes in most recent MiniUPnP (18/04/2011) that will need some massaging in RC to get it to build.

I will be happy to post an issue at github if needed. I just wanted to get an idea whether this should be an issue at all or should RC just support the 1.5 MiniUPnP.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: xf2_org on April 20, 2011, 09:14:38 PM
"jgarzik's favorite features list" for 0.3.21 includes,

  • A new method of seeding addresses, via DNS A records.  -dnsseed option will enable this, if you'd like to avoid IRC seeding for whatever reason.
  • -logtimestamps option, to add a timestamp to each line in debug.log.  Note: satoshi worried about the privacy implications of this option, so it defaults to 'off'.
  • Immature blocks (newly generated, under 120 confirmations) are now shown in listtransactions.  My own xlisttransactions (http://bitcointalk.org/index.php?topic=611.0) patch should be obsolete at this point, unless I'm missing something.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: xf2_org on April 20, 2011, 09:15:57 PM
Is this thread the right place to discuss issues with RC?

Absolutely!

Quote
What version of MiniUPnP is intended to be supported? There seem to be some changes in most recent MiniUPnP (18/04/2011) that will need some massaging in RC to get it to build.

I will be happy to post an issue at github if needed. I just wanted to get an idea whether this should be an issue at all or should RC just support the 1.5 MiniUPnP.

1.5 is the supported and tested version, but I'm curious what massaging is needed for more recent versions?



Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: eMansipater on April 20, 2011, 09:27:37 PM
I understand the reason for the change (although these transactions will become confirmed after some time or if the block is less than 3 kB which happens from time to time) but the error dialog is very confusing.
Agreed.  Can we change this from "This is an oversized transaction that requires a transaction fee" to something like "Because of its amount, complexity, or rapid re-use of recently received coins, this transaction requires a transaction fee (0.0x recommended)" ?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: Gavin Andresen on April 20, 2011, 09:39:50 PM
RE: changing the confusing transaction fee message:  good idea, and yes, "we" can.

How about:

This transaction requires a transaction fee of at least 0.0N because of its amount, complexity, or use of recently received funds


I don't want to use the word "recommend", because the GUI doesn't let you try to send them without a fee.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: dust on April 20, 2011, 09:40:40 PM
What are the disadvantages (security, etc) of having UPnP turned on?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: trentzb on April 20, 2011, 09:47:22 PM
1.5 is the supported and tested version, but I'm curious what massaging is needed for more recent versions?
Yea, 1.5 builds fine.

1.5.20110418 does not build but it looks like it may be partially a miniupnp install issue. It fails to install portlistingparse.h and miniupnpctypes.h.

upnpDiscover():53 is in miniupnpc.h as
Code:
upnpDiscover(int delay, const char * multicastif,
             const char * minissdpdsock, int sameport,
             int ipv6,
             int * error);

net.cpp calls it as:
Code:
devlist = upnpDiscover(2000, multicastif, minissdpdpath, 0);

There may be more, that was the first one I hit though.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: dishwara on April 20, 2011, 09:47:32 PM
WOW, is it true?
I see 8 decimal points, that completely solves the microbitcent problem.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: eMansipater on April 20, 2011, 09:49:28 PM
This transaction requires a transaction fee of at least 0.0N because of its amount, complexity, or use of recently received funds
Much more succinct!  And you're right--"we" is just a polite way of some random person on the internet trying to lend themselves credibility.  It's the people who actually wrangle code that do the work   :-[


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: ByteCoin on April 20, 2011, 09:50:01 PM
This transaction requires a transaction fee of at least 0.0N because of its ... complexity,

Since the prohibition on non-standard transactions, the only reason for a transaction to be overly "complex" is due to its large number of inputs and outputs.

In this case the solution is for the client software to offer the option of  recasting the transaction into two or more less complex ones, possibly using staging addresses, which would no longer incur fees.

ByteCoin


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: dishwara on April 20, 2011, 09:52:37 PM
Using for the past 10-20 minutes. So far no glitch or anything, besides, instead of 2 digits , now i see 8 digits after decimal point.

Quote
Support for Universal Plug and Play to open a port for incoming connections

Is that mean can any port be used to connect bitcoin client to internet besides 8332.
I have a fear that suddenly ISP's will block 8332 port, at least my ISP.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: trentzb on April 20, 2011, 09:58:52 PM
Quote
Support for Universal Plug and Play to open a port for incoming connections

Is that mean can any port be used to connect bitcoin client to internet besides 8332.
I have a fear that suddenly ISP's will block 8332 port, at least my ISP.
I don't know anything about UPnP but I would have to say no. It would only open port 8333 (for incoming). Someone correct me if I am wrong.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: khal on April 20, 2011, 11:39:14 PM
I suspect some fee problem when sending with a CIn like 0.01234567.
I've sent 0.01234567 to myself, waited for 10 confirmations and try to send 0.01234567 to myself again.
=> bitcoin GUI suggest me a 0.01 fee
This transaction is over the size limit.  You can still send it for a fee of 0.01, which goes to the nodes that process your transaction and helps to support the network.  Do you want to pay the fee?

I tried 7/8 times, same message appears. There shouldn't be a fee here ?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: mewantsbitcoins on April 20, 2011, 11:49:09 PM
I don't have an account with GitHub, so here goes:

UPnP doesn't seem to work. I checked "Map port using UPnP" and restarted the client but still stuck at 8 connections.
I'm running Windows 7 Ultimate SP1 32bit in VMware Fusion 3.1. utorrent UPnP works fine in the same environment.

Also, might be just me, but I think "Send" and "Cancel" buttons are too small in "Send Coins"(in Windows 7 default theme)


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: xf2_org on April 21, 2011, 12:06:07 AM
I suspect some fee problem when sending with a CIn like 0.01234567.
I've sent 0.01234567 to myself, waited for 10 confirmations and try to send 0.01234567 to myself again.
=> bitcoin GUI suggest me a 0.01 fee
This transaction is over the size limit.  You can still send it for a fee of 0.01, which goes to the nodes that process your transaction and helps to support the network.  Do you want to pay the fee?

I tried 7/8 times, same message appears. There shouldn't be a fee here ?

To bitcoin, this type of transaction appears to be a low priority, spam transaction.  Sending tiny amounts of bitcoins to yourself is a waste of network resources -- remember, you are asking every bitcoin node, all over the world, to validate and store your transaction.



Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: Matt Corallo on April 21, 2011, 12:17:05 AM
I don't have an account with GitHub, so here goes:

UPnP doesn't seem to work. I checked "Map port using UPnP" and restarted the client but still stuck at 8 connections.
I'm running Windows 7 Ultimate SP1 32bit in VMware Fusion 3.1. utorrent UPnP works fine in the same environment.

Also, might be just me, but I think "Send" and "Cancel" buttons are too small in "Send Coins"
Would you mind grabbing a copy of http://miniupnp.tuxfamily.org/files/download.php?file=upnpc-exe-win32-20110418.zip (http://miniupnp.tuxfamily.org/files/download.php?file=upnpc-exe-win32-20110418.zip) and running upnpc-static.exe -a YourInternalIPv4 8333 8333 tcp and posting the output?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: mewantsbitcoins on April 21, 2011, 01:47:21 AM
Would you mind grabbing a copy of http://miniupnp.tuxfamily.org/files/download.php?file=upnpc-exe-win32-20110418.zip (http://miniupnp.tuxfamily.org/files/download.php?file=upnpc-exe-win32-20110418.zip) and running upnpc-static.exe -a YourInternalIPv4 8333 8333 tcp and posting the output?

Sorry about the delay. After seeing the error message I realized what happened. I had manually forwarded 8333 on my router to another computer running bitcoin client. I now deleted the rule and it works ok.

Code:
C:\Users\peter>upnpc-static.exe -a 192.168.1.5 8333 8333 tcp
upnpc : miniupnpc library test client. (c) 2006-2011 Thomas Bernard
Go to http://miniupnp.free.fr/ or http://miniupnp.tuxfamily.org/
for more information.
List of UPNP devices found on the network :
 desc: http://192.168.1.1:5000/Public_UPNP_gatedesc.xml
 st: urn:schemas-upnp-org:service:WANPPPConnection:1

Found valid IGD : http://192.168.1.1:5000/Public_UPNP_C3
Local LAN ip address : 192.168.1.5
ExternalIPAddress = 86.5.50.90
AddPortMapping(8333, 8333, 192.168.1.5) failed with code 718 (ConflictInMappingE
ntry)
GetSpecificPortMappingEntry() failed with code -1 (Miniupnpc Unknown Error)

After:
Code:
C:\Users\peter>upnpc-static.exe -a 192.168.1.5 8333 8333 tcp
upnpc : miniupnpc library test client. (c) 2006-2011 Thomas Bernard
Go to http://miniupnp.free.fr/ or http://miniupnp.tuxfamily.org/
for more information.
List of UPNP devices found on the network :
 desc: http://192.168.1.1:5000/Public_UPNP_gatedesc.xml
 st: urn:schemas-upnp-org:service:WANIPConnection:1

Found valid IGD : http://192.168.1.1:5000/Public_UPNP_C3
Local LAN ip address : 192.168.1.5
ExternalIPAddress = *
InternalIP:Port = 192.168.1.5:8333
external *:8333 TCP is redirected to internal 192.168.1.5:8333 (duratio
n=0)

After this I installed the new client on the other machine. When I turned "Map port using UPnP" the first machine's port forwarding was dropped by the router. How would I go about setting up multiple machines running bitcoin client behind 1 router?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: dishwara on April 21, 2011, 01:54:30 AM
Also, might be just me, but I think "Send" and "Cancel" buttons are too small in "Send Coins"
+1, the buttons are very small.

I suspect some fee problem when sending with a CIn like 0.01234567.
I've sent 0.01234567 to myself, waited for 10 confirmations and try to send 0.01234567 to myself again.
=> bitcoin GUI suggest me a 0.01 fee
This transaction is over the size limit.  You can still send it for a fee of 0.01, which goes to the nodes that process your transaction and helps to support the network.  Do you want to pay the fee?

I tried 7/8 times, same message appears. There shouldn't be a fee here ?

No, its not error or bug. Gavin mentioned it in first post itself. You can send coins less than 0.01, but you have to pay 0.01 as fees
Quote
Sending and displaying arbitrary precision amounts of bitcoins (you can send exactly 1.0001 bitcoins if you like).  Sending less than 0.01 bitcoins requires a 0.01 bitcoin fee, however.


Would you mind grabbing a copy of http://miniupnp.tuxfamily.org/files/download.php?file=upnpc-exe-win32-20110418.zip (http://miniupnp.tuxfamily.org/files/download.php?file=upnpc-exe-win32-20110418.zip) and running upnpc-static.exe -a YourInternalIPv4 8333 8333 tcp and posting the output?

i downloaded & ran it, here i got result with no bitcoin client running.
Code:


C:\Users\Administrator\Desktop\upnpc-exe-win32-20110418>upnpc-static.exe -a 192.
168.1.2 8333 8333 tcp
upnpc : miniupnpc library test client. (c) 2006-2011 Thomas Bernard
Go to http://miniupnp.free.fr/ or http://miniupnp.tuxfamily.org/
for more information.
List of UPNP devices found on the network :
 desc: http://192.168.1.1:5431/dyndev/uuid:0000e0a8-20a0-00e0-20a0-488800f808e0
 st: urn:schemas-upnp-org:device:InternetGatewayDevice:1

Found valid IGD : http://192.168.1.1:5431/uuid:0000e0a8-20a0-00e0-20a0-488802f86
048/WANPPPConnection:1
Local LAN ip address : 192.168.1.2
ExternalIPAddress = 122.164.210.213
InternalIP:Port = 192.168.1.2:8333
external 122.164.210.213:8333 TCP is redirected to internal 192.168.1.2:8333 (du
ration=0)

Then i ran bitcoin client with "map port using Upnp"
i got this
Code:

C:\Users\Administrator\Desktop\upnpc-exe-win32-20110418>upnpc-static.exe -a 192.
168.1.2 8333 8333 tcp
upnpc : miniupnpc library test client. (c) 2006-2011 Thomas Bernard
Go to http://miniupnp.free.fr/ or http://miniupnp.tuxfamily.org/
for more information.
List of UPNP devices found on the network :
 desc: http://192.168.1.1:5431/dyndev/uuid:0000e0a8-20a0-00e0-20a0-488800f808e0
 st: urn:schemas-upnp-org:device:InternetGatewayDevice:1

Found valid IGD : http://192.168.1.1:5431/uuid:0000e0a8-20a0-00e0-20a0-488802f86
048/WANPPPConnection:1
Local LAN ip address : 192.168.1.2
ExternalIPAddress = 122.164.210.213
InternalIP:Port = 192.168.1.2:8333
external 122.164.210.213:8333 TCP is redirected to internal 192.168.1.2:8333 (du
ration=0)

Actually, i see both same & i don't understand anything.
Can anyone explain why i got same thing both time or what is wrong in mine?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: elggawf on April 21, 2011, 02:43:55 AM
What are the disadvantages (security, etc) of having UPnP turned on?

In the client? Nothing.

Some routers are terrible in their support of uPNP (eg, some firmware versions of D-link I believe it was would allow any uPNP client on the LAN to set the default gateway for other hosts, etc), but if you have one of those then it makes no difference if the Bitcoin client has it on or not. If it's enabled on your router and it's insecure, it's a security liability no matter what. Some folks disable it on their router automatically as a matter of habit, simply because historically router companies are idiots in the way they implement it.

So as far as just enabling/disabling it in the client, simply put if you are fine with manually forwarding ports where they need to go, or if they're already forwarded properly, don't bother enabling it.

Now for my questions, first: Does the Bitcoin client exchange peer information with other clients, such that it'll maintain a decent connection to the swarm if it's not able to go on IRC? (I'm assuming the answer to this question is affirmative, but I want to make sure because it's an assumption for the second question).

Second: Does the Bitcoin client do any type of validation on the IRC server it connects to to seed the p2p mechanism? I don't really want to be on some strange IRC network all day every day with several clients, so I was thinking about setting up an IRC server on my router and transparently forwarding my bitcoin clients there. Assuming at least one of my clients maintains at least one active peer outside my LAN, all clients should stay in the swarm with no issues right?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: mewantsbitcoins on April 21, 2011, 03:26:07 AM
Actually, i see both same & i don't understand anything.
Can anyone explain why i got same thing both time or what is wrong in mine?

There's nothing wrong with yours - it's working flawlessly. Read a little about how networks work and what port forwarding does - you'll understand  ;)


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: khal on April 21, 2011, 05:38:54 AM
To bitcoin, this type of transaction appears to be a low priority, spam transaction.  Sending tiny amounts of bitcoins to yourself is a waste of network resources -- remember, you are asking every bitcoin node, all over the world, to validate and store your transaction.
There's no fee penality to send to self. But there is for small amounts & recent transactions.

No, its not error or bug. Gavin mentioned it in first post itself. You can send coins less than 0.01, but you have to pay 0.01 as fees
It's not less than 0.01.
Coin selection choose 3 input transactions of the same amout (3 x 0.01234567) to pay 0.01234567. So, i guess the transaction is above 1k (or priority implies a fee). Not a bug so, just need to optimize coin selection maybe.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: alkor on April 21, 2011, 05:43:13 AM
Linux and Windows binary releases are at:
  https://sourceforge.net/projects/bitcoin/files/Bitcoin/bitcoin-0.3.21/

What about the Apple binaries? Is Mac OS no longer supported?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: gigabytecoin on April 21, 2011, 06:24:45 AM
Linux and Windows binary releases are at:
  https://sourceforge.net/projects/bitcoin/files/Bitcoin/bitcoin-0.3.21/

Changes and new features include:
  • Support for Universal Plug and Play to open a port for incoming connections (off by default, turn it on using the -upnp=1 command-line switch or the checkbox on the Options dialog box).
  • Sending and displaying arbitrary precision amounts of bitcoins (you can send exactly 1.0001 bitcoins if you like).  Sending less than 0.01 bitcoins requires a 0.01 bitcoin fee, however.
  • New rpc command "sendmany" to send bitcoins to more than one person in a single transaction (already being used by some of the mining pools for payouts).
  • Several bug fixes, including a serious intermittent bug that would sometimes cause bitcoind to stop accepting rpc requests.

If you find bugs, report them at: https://github.com/bitcoin/bitcoin/issues


Why have you enforced a fee of 0.01 btc for small transactions? What sense does it make to send a transaction with a fee that costs more than the actual transaction itself?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: HostFat on April 21, 2011, 06:50:43 AM
yes, it's bad.
Because the exchange value of Bitcoin ( from USD ) can go very high.
So even 0.01 bitcoin can be a valuable amount of cash.

How can I add the updated translation to the next client release?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: elggawf on April 21, 2011, 06:53:39 AM
Why have you enforced a fee of 0.01 btc for small transactions? What sense does it make to send a transaction with a fee that costs more than the actual transaction itself?

It's to stop denial of service on the network, by someone geting 1 BTC and then sending the tiniest fractions of it all over the place to fill up the block chain for no good purpose.

I would imagine that the minimum fee would be lowered at some point if the currency deflates that far - I think you want it to discourage frivilous transactions, but not be prohibitive to legitimate microtransactions. At this point though there is a ~1c USD charge to send anything that is vastly below ~1cUSD... which makes sense when you think about it.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: gigabytecoin on April 21, 2011, 07:06:45 AM
Why have you enforced a fee of 0.01 btc for small transactions? What sense does it make to send a transaction with a fee that costs more than the actual transaction itself?

It's to stop denial of service on the network, by someone geting 1 BTC and then sending the tiniest fractions of it all over the place to fill up the block chain for no good purpose.

I would imagine that the minimum fee would be lowered at some point if the currency deflates that far - I think you want it to discourage frivilous transactions, but not be prohibitive to legitimate microtransactions. At this point though there is a ~1c USD charge to send anything that is vastly below ~1cUSD... which makes sense when you think about it.

Couldn't somebody interested in a DOS attack on bitcoin in this manner simply re-build their own customized client of bitcoin?

I am sure whoever wanted to accomplish a DOS attack could do that. So, why inhibit regular users because of a possible attacker really is the question?

Captchas for example... some websites choose not to use them period, because they are such an inconvenience to the average user. No matter how many spam registrations they receive.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: toffoo on April 21, 2011, 08:01:22 AM
Linux and Windows binary releases are at:
  https://sourceforge.net/projects/bitcoin/files/Bitcoin/bitcoin-0.3.21/

What about the Apple binaries? Is Mac OS no longer supported?

Well, I saw Gavin's photo in Forbes today, sitting in front of his iMac ... which made me wonder why it seems like Apple-users are second-class Bitcoin citizens.  If the lead developer is coding on a Mac, can't he build us a binary?  I don't think OSX even got an official build of 0.3.20.2 !


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: dishwara on April 21, 2011, 08:39:51 AM
Linux and Windows binary releases are at:
  https://sourceforge.net/projects/bitcoin/files/Bitcoin/bitcoin-0.3.21/

What about the Apple binaries? Is Mac OS no longer supported?

Well, I saw Gavin's photo in Forbes today, sitting in front of his iMac ... which made me wonder why it seems like Apple-users are second-class Bitcoin citizens.  If the lead developer is coding on a Mac, can't he build us a binary?  I don't think OSX even got an official build of 0.3.20.2 !
Don't get wrong. Linux is primary here for one thing open source, so do bitcoin - open source.
Due to huge number of users & request windows client made.
Soon their will be for mac.
The problem is Gavin is the only one so far working on whole official client.
Now some programmers joined with him. If you know about Mac programming you can help.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: Matt Corallo on April 21, 2011, 09:24:25 AM
After this I installed the new client on the other machine. When I turned "Map port using UPnP" the first machine's port forwarding was dropped by the router. How would I go about setting up multiple machines running bitcoin client behind 1 router?
You can't, simple as that.  You can only ever forward one port to one client thus port forwarding will only work on one (as you can't change the Bitcoin port).  Which one gets the port forward with specific who has UPnP on, etc settings depends on your router. 


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: theymos on April 21, 2011, 12:24:17 PM
Couldn't somebody interested in a DOS attack on bitcoin in this manner simply re-build their own customized client of bitcoin?

They won't be processed by the network (most of it, anyway), even if you modify your client to send them. They'll just sit at "0/unconfirmed" forever.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: dbitcoin on April 22, 2011, 06:47:59 PM
bitcoind nogui not builds without miniupnpc (default settings).
Looks like at least unix makefile and related header files do not proper handle this.

And when I completely exclude miniupnpc from build by hand:

Code:
++ -O2 -Wno-invalid-offsetof -Wformat -g -D__WXDEBUG__ -DNOPCH -DFOURWAYSSE2 -DUSE_SSL -o bitcoind obj/nogui/util.o obj/nogui/script.o obj/nogui/db.o obj/nogui/net.o obj/nogui/irc.o obj/nogui/main.o obj/nogui/rpc.o obj/nogui/init.o cryptopp/obj/sha.o cryptopp/obj/cpu.o obj/sha256.o -Wl,-Bstatic -l boost_system -l boost_filesystem -l boost_program_options -l boost_thread -l db_cxx -l ssl -l crypto -Wl,-Bdynamic -l gthread-2.0 -l z -l dl
obj/nogui/db.o: In function `CWalletDB::LoadWallet()':
/home/dlancer/bitcoin/src/bitcoin-patched/db.cpp:811: undefined reference to `fUseUPnP'
obj/nogui/db.o: In function `Unserialize<CDataStream>':
/home/dlancer/bitcoin/src/bitcoin-patched/serialize.h:136: undefined reference to `fUseUPnP'
collect2: ld returned 1 exit status


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: Matt Corallo on April 22, 2011, 07:13:03 PM
bitcoind nogui not builds without miniupnpc (default settings).
Looks like at least unix makefile and related header files do not proper handle this.

And when I completely exclude miniupnpc from build by hand:

Code:
++ -O2 -Wno-invalid-offsetof -Wformat -g -D__WXDEBUG__ -DNOPCH -DFOURWAYSSE2 -DUSE_SSL -o bitcoind obj/nogui/util.o obj/nogui/script.o obj/nogui/db.o obj/nogui/net.o obj/nogui/irc.o obj/nogui/main.o obj/nogui/rpc.o obj/nogui/init.o cryptopp/obj/sha.o cryptopp/obj/cpu.o obj/sha256.o -Wl,-Bstatic -l boost_system -l boost_filesystem -l boost_program_options -l boost_thread -l db_cxx -l ssl -l crypto -Wl,-Bdynamic -l gthread-2.0 -l z -l dl
obj/nogui/db.o: In function `CWalletDB::LoadWallet()':
/home/dlancer/bitcoin/src/bitcoin-patched/db.cpp:811: undefined reference to `fUseUPnP'
obj/nogui/db.o: In function `Unserialize<CDataStream>':
/home/dlancer/bitcoin/src/bitcoin-patched/serialize.h:136: undefined reference to `fUseUPnP'
collect2: ld returned 1 exit status
Have you tried building with USE_UPNP=


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: dbitcoin on April 22, 2011, 07:45:11 PM
Have you tried building with USE_UPNP=
Yep.
I found where is problem. Current makefiles want installed libminiupnpc.
But this is wrong. If I not need upnp support at all, why I need install it?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: dust on April 22, 2011, 07:53:38 PM
Commenting out the USE_UPNP line allowed me to build without miniupnpc.  Assigning zero still defines the symbol.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: Matt Corallo on April 22, 2011, 08:35:09 PM
Commenting out the USE_UPNP line allowed me to build without miniupnpc.  Assigning zero still defines the symbol.
If you build with USE_UPNP= then the effect is the same as commenting out that line (undefining USE_UPNP).  If you build with USE_UPNP=1, you are building with UPnP with it on by default, USE_UPNP=0 also builds with UPnP with it off by default. 


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: dbitcoin on April 22, 2011, 08:44:58 PM
Commenting out the USE_UPNP line allowed me to build without miniupnpc.  Assigning zero still defines the symbol.
If you build with USE_UPNP= then the effect is the same as commenting out that line (undefining USE_UPNP).  If you build with USE_UPNP=1, you are building with UPnP with it on by default, USE_UPNP=0 also builds with UPnP with it off by default. 

Yep, with commenting all builds ok.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: Matt Corallo on April 22, 2011, 08:46:56 PM
Commenting out the USE_UPNP line allowed me to build without miniupnpc.  Assigning zero still defines the symbol.
If you build with USE_UPNP= then the effect is the same as commenting out that line (undefining USE_UPNP).  If you build with USE_UPNP=1, you are building with UPnP with it on by default, USE_UPNP=0 also builds with UPnP with it off by default. 

Yep, with commenting all builds ok.
You shouldn't comment out that line.  There is no point changing the source when all you have to do is undefine USE_UPNP by specifying USE_UPNP= when you build.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: dbitcoin on April 22, 2011, 08:50:04 PM
You shouldn't comment out that line.  There is no point changing the source when all you have to do is undefine USE_UPNP by specifying USE_UPNP= when you build.

For custom build this is doesn't matter. I anyway apply several patches for pool.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: jorgen on April 22, 2011, 09:31:31 PM
When I try to edit or add new entry in address book, I can't add or edit bitcoinaddress field. It looks like thin line under the field with contact name


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: Matt Corallo on April 22, 2011, 09:48:50 PM
When I try to edit or add new entry in address book, I can't add or edit bitcoinaddress field. It looks like thin line under the field with contact name
I don't think you ever could.  Though that would be a nice feature to add.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: ShadowOfHarbringer on May 17, 2011, 05:16:11 PM
RE: changing the confusing transaction fee message:  good idea, and yes, "we" can.

How about:

This transaction requires a transaction fee of at least 0.0N because of its amount, complexity, or use of recently received funds


I don't want to use the word "recommend", because the GUI doesn't let you try to send them without a fee.


Gavin, i don't doubt your good intentions when requiring 0.01 fee, but this produces serious issues currently in my Bitcoin trading.

Meaningly, when i tried to transfer about 9 bitcoins again after receiving 7 confirmations, i still got the "0.01 fee required because of blah blah" message. When i downgrade to 0.3.20, it does not produce this error.

So what is the point of this feature if somebody can "fix it" just by downgrading to a slightly lower version ?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: ShadowOfHarbringer on May 17, 2011, 05:22:20 PM
This is the receiving address I am talking about:

http://blockexplorer.com/address/1LBK2EespTDpcxnR1J4utP9tBtbA6X5SF9

As you can see, i received 9.95 bitcoins on it, and then i was unable to push the bitcoins further even though i had 7 confirmations.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: Gavin Andresen on May 17, 2011, 05:47:00 PM
So what is the point of this feature if somebody can "fix it" just by downgrading to a slightly lower version ?

Downgrading does not "fix" it.  The earlier version would broadcast the transaction, and then it would (most likely) either sit in the network with 0 confirmations for half a day or sit on your machine with 0 confirmations for a few hours, until it aged enough to graduate from the 'very low priority' category.

The rule of thumb is 1 BTC you just received needs 144 confirmations (1 day) before it is considered 'normal priority.'  10 bitcoins requires 1/10'th that-- 15 confirmations would do it.

Given the recent rise in bitcoin value, I think it makes sense to modify that formula so 1 BTC takes (perhaps) 10 confirmations to be considered 'not likely to be spam'.


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: josef on May 17, 2011, 06:26:31 PM
When I try to edit or add new entry in address book, I can't add or edit bitcoinaddress field. It looks like thin line under the field with contact name
I don't think you ever could.  Though that would be a nice feature to add.

Wouldn't this be futile unless you could also could copy in the private key that goes with that address?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: gigabytecoin on May 17, 2011, 08:37:23 PM
Nobody saying anything about bitcoin gui not working on ubuntu 11.04 out of the box?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: ShadowOfHarbringer on May 18, 2011, 09:24:23 AM
So what is the point of this feature if somebody can "fix it" just by downgrading to a slightly lower version ?

Downgrading does not "fix" it.  The earlier version would broadcast the transaction, and then it would (most likely) either sit in the network with 0 confirmations for half a day or sit on your machine with 0 confirmations for a few hours, until it aged enough to graduate from the 'very low priority' category.

Perhaps the theory is correct, but the practice confirms you are wrong.

The 9.95 was sent by me about 1 hour after receiving it using 0.3.20 version and later I got 3 confirmations after about 45 minutes.

Perhaps the criteria for requiring transaction fees are wrong ? Yes, i am absolutely certain we got 3 confirmations and i think that 45 minutes is quickly enough, isn't it ? So why require a fee ?


Title: Re: Please help test: Bitcoin version 0.3.21 release candidate
Post by: elmom on May 19, 2011, 08:02:21 AM
Nobody saying anything about bitcoin gui not working on ubuntu 11.04 out of the box?

EDIT:
Just saw the issue on github, should've looked there first :)

OLD:
Same here.

Starting the client, it uses some cycles, but gives no kind of output. No gui and nothing on stdout/err. Maybe wxwidgets 2.9 doesn't like Unity?

I've tried with both 0.3.19 and *.21.

Can someone try without Unity?