Please upgrade to Bitcoin 0.8 and help Android/MultiBit users!

[jim618]

Things to do this weekend:

+ meet up with your friends and have a beer/ coffee together
+ check your credit card statements and avoid those sneaky late payment fees the banks slap on you
+ phone up your mum and see if everything is all right with her
+ upgrade your bitcoind to v0.8.0

:-)

[1714913952]

1714913952

[1714913952]

1714913952

[1714913952]

1714913952

[1714913952]

1714913952

[teukon]

Finished, phew!

Adding to DannyHamilton's datapoint:

Processing equiptment: One 32-bit, single core, 900 MHz CPU (bitcoin-qt limited to 50% to avoid overheating, my CPU fan is broken).
Re-indexing time: 55 hours

This should be no real problem for even somewhat dated desktops and laptops but I'd advise those attempting the upgrade on ancient/exotic hardware to prepare themselves.  I've been writing down transactions I wanted to make while waiting for my client to sync.

Anyway, I now have one almost-24-7 node in Asia with a reasonable connection (at least 10 Mb/s upload).  Hope it helps!

[Cryptoman]

So I see that upgrading to 0.8 and reindexing has created a new set of blockchain files under the "blocks" subdirectory.  However, everything seems to occupy about a gigabyte total now:

Code:

11M     ./database
166M    ./chainstate
31M     ./blocks/index
790M    ./blocks

Is this correct, and is it safe to delete the block0001.dat-block0003.dat and blkindex.dat files in the parent directory?

[Mike Hearn]

This is a common point of confusion. The old files can be deleted. They actually aren't using up the space that they may appear to be as they are hard links, but a lot of Windows apps don't know how to represent the disk space usage of hard linked files correctly. I bet if you look at the disk free measurement on the drives properties page it doesn't actually change after removing the old files.

[Jason]

Processing equiptment: One 32-bit, single core, 900 MHz CPU (bitcoin-qt limited to 50% to avoid overheating, my CPU fan is broken).
Re-indexing time: 55 hours

Doesn't take long on modern hardware:

i5-3570 overclocked at 4.5 GHz.
Re-indexing time:  47 minutes.

Nice to see it was CPU-bound on all 4 cores for a good portion of that time.

[cypherdoc]

so that's good enough?  never checked this way before.

thanks.

you only need to check if the signature is OK (by clicking "details"). the rest is not needed.

Yes. I wanted to illustrate the depth of information available. Anyone handling significant amounts of coins or sensitive data should in fact check the certification path, and look for suspicious changes in signing authorities. Case in point:
http://arstechnica.com/security/2012/06/flame-malware-was-signed-by-rogue-microsoft-certificate/

i'm a little skeptical about signing certifs after that incident last year where Diginotar got compromised.

i mean no one around here has told us how to verify that the CA you pointed out is genuine.  i see a few words that say Bitcoin Foundation in the detail and COMODO and someone's email address but how am i supposed to know that this detail is genuine?

My unqualified guess is that if you are concerned you should contact the site owner and request (via phone, PGP-authenticated communication channel, or in person) to verify the serial numbers and/or thumbprints for all the certificates in the certification chain. A pain in the ass that I never inflicted upon myself.

For the record, windows installer 0.8.0 for me shows certification path

USERTrust>Comodo Code Signing CA 2>The Bitcoin Foundation, Inc. (serial no. 00 95 58 31 df b0 68 e1 11 ee 55 2a b6 2c f7 33 62, sha1 digest 8c 94 64 e3 b5 b0 41 89 5b 89 b0 57 cc 74 b9 44 e5 b2 92 66)

fyi, i show exactly the same info. 

[pulsecat]

Does anybody know will bitcoin-qt package in Ubuntu get updated?

[giszmo]

Does anybody know will bitcoin-qt package in Ubuntu get updated?

… I'm confused about that, too. Shouldn't the ppa be updated long ago?

(I don't agree with Mike Hearn's push to upgrade to 0.8 with the reasoning that new incoming crowd might be disappointed. There is no reason to rush specifically this release. It is not a bug fix release that would deserve this pushing and as we all know, new software might have glitches, so the exposure to such glitches should be kept low. Better have people switch moderately fast, most adventurous to most regular user than to migrate faster this time than else. It will make a difference of days when Spinner and Co may use these bloom nodes anyway.)

[Mike Hearn]

Even with promotion in the forums, it's going to take a long time before even half the network runs 0.8, so I wouldn't worry too much about that. Also, it's been in testing for quite a long time by now already. There isn't much reason to hold back.

It looks like we might hit 10% in a few days. So probably Andreas will release the Android app upgrade soon.

[cypherdoc]

Does anybody know will bitcoin-qt package in Ubuntu get updated?

specifically the PPA

[Cryptoman]

This is a common point of confusion. The old files can be deleted. They actually aren't using up the space that they may appear to be as they are hard links, but a lot of Windows apps don't know how to represent the disk space usage of hard linked files correctly. I bet if you look at the disk free measurement on the drives properties page it doesn't actually change after removing the old files.

Thanks for clearing that up!

Code:

~/.bitcoin> du -h
168M    ./chainstate
11M     ./database
288K    ./encrypted_wallet_backup
31M     ./blocks/index
883M    ./blocks
8.2G    .
~/.bitcoin> ls -l blk*
-rw-r--r--  2 nobody users 2097220817 Jul 29  2012 blk0001.dat
-rw-r--r--  2 nobody users 2097158452 Dec  6 01:24 blk0002.dat
-rw-r--r--  2 nobody users 1499978472 Feb 16 00:43 blk0003.dat
-rw-r--r--  1 nobody users 1818804224 Feb 16 00:43 blkindex.dat
~/.bitcoin> rm blk*
~/.bitcoin> du -h
168M    ./chainstate
11M     ./database
288K    ./encrypted_wallet_backup
31M     ./blocks/index
6.2G    ./blocks
6.5G    .
~/.bitcoin> 

[jim618]

We are now over 10% of the network updated to v0.8.0 !

I've put the version of MultiBit with bloom filter support on the multibit.org website for general release:
https://multibit.org/releases.html

Thanks everyone for upgrading the network so swiftly.

[Mike Hearn]

Andreas rolled out the Bitcoin Wallet upgrade this morning. Thanks everyone!

[giszmo]

Andreas rolled out the Bitcoin Wallet upgrade this morning. Thanks everyone!

That's good news. Just installed it for the first time in one year and I must say, seeing "Synchronization stalled, 18 months behind" is frustrating not only for noobs. Hope that gets better. The peer monitor shows me one 0.8 out of 5 connected nodes and I could not find any setting to activate bloom filtering. I assume it filters the bloom by default Shouldn't it ask for more peers to swap out those pre 0.8 peers?

[runeks]

Version 0.8 just became available in the Bitcoin PPA.

To whomever controls this PPA: have you considered trying to get 0.8 included in the new 13.04 Ubuntu release? I'm not sure what the process entails, but it seems you've already went through all the trouble and created a package, and all they need to do is pull it into Raring.

[Mike Hearn]

That's good news. Just installed it for the first time in one year and I must say, seeing "Synchronization stalled, 18 months behind" is frustrating not only for noobs. Hope that gets better. The peer monitor shows me one 0.8 out of 5 connected nodes and I could not find any setting to activate bloom filtering. I assume it filters the bloom by default Shouldn't it ask for more peers to swap out those pre 0.8 peers?

Bloom filtering speeds up download of the part of the chain where your wallet exists. When you install it fresh for a new user, it still has to download all the headers to catch up and yes, this is too slow and yes, fixing it is my next highest priority. It needs to start up instantly for new users, no argument there.

The place you'll really notice the 0.8 difference is when it's finally caught up with the chain head and you open it a few days later, you'll see that it goes from being 1-2 days behind (since your last charge) to fully caught up almost instantly. Before it could take several minutes and use a ton of bandwidth. There are 144 blocks per day so you can imagine that if we started having 1mb sized blocks, it'd have to download and parse 144mb of data to catch up with a days worth of block chain which is just infeasible. With Bloom filtering, that's more like 30-40kb.

[cypherdoc]

Version 0.8 just became available in the Bitcoin PPA.

To whomever controls this PPA: have you considered trying to get 0.8 included in the new 13.04 Ubuntu release? I'm not sure what the process entails, but it seems you've already went through all the trouble and created a package, and all they need to do is pull it into Raring.

Matt Corallo

[giszmo]

That's good news. Just installed it for the first time in one year and I must say, seeing "Synchronization stalled, 18 months behind" is frustrating not only for noobs. Hope that gets better. The peer monitor shows me one 0.8 out of 5 connected nodes and I could not find any setting to activate bloom filtering. I assume it filters the bloom by default Shouldn't it ask for more peers to swap out those pre 0.8 peers?

Bloom filtering speeds up download of the part of the chain where your wallet exists. When you install it fresh for a new user, it still has to download all the headers to catch up and yes, this is too slow and yes, fixing it is my next highest priority. It needs to start up instantly for new users, no argument there.

The place you'll really notice the 0.8 difference is when it's finally caught up with the chain head and you open it a few days later, you'll see that it goes from being 1-2 days behind (since your last charge) to fully caught up almost instantly. Before it could take several minutes and use a ton of bandwidth. There are 144 blocks per day so you can imagine that if we started having 1mb sized blocks, it'd have to download and parse 144mb of data to catch up with a days worth of block chain which is just infeasible. With Bloom filtering, that's more like 30-40kb.

In order to use it in Chile I need to switch off pre 0.8 full downloads and actually would like to switch off mobile network downloads at all. I would want to have it sync only on wifi or on demand on mobile internet.

And yes, if it is throwing away data that does not affect the own address and it just generated the first own address, it should be instantly ready.

Edit: I installed 0.8 and it is reindexing since an hour or so at 300% CPU.

[Mike Hearn]

In order to use it in Chile I need to switch off pre 0.8 full downloads and actually would like to switch off mobile network downloads at all. I would want to have it sync only on wifi or on demand on mobile internet.

There's no option to switch off pre-0.8 downloads, but you can use the trusted peer preference and set it to a node you know runs 0.8. If you want to, you can use riker.plan99.net which is run by me, but don't forget to eventually unset this so you balance the load appropriately.

The app will sync automatically (if you don't change the setting) when plugged in and charging. So, assuming you have wifi at home and you charge at home, the daily sync will be over WiFi. Otherwise it'll only sync when you open the app. So you can avoid it using up your mobile quota by just not opening the app when you're on 3G.

On recent versions of Android, if you go into the data usage screen, you can tick a box that means the app won't be allowed to use background data when it's not got WiFi.

And yes, if it is throwing away data that does not affect the own address and it just generated the first own address, it should be instantly ready.

It is for me, so if you see something different let me know and we can investigate.

[giszmo]

And yes, if it is throwing away data that does not affect the own address and it just generated the first own address, it should be instantly ready.

It is for me, so if you see something different let me know and we can investigate.

Well, I just installed it and it ran into several ANR and out of battery and did not catch up the block chain in hours … ok, so it might get the balance straight from the start but still would have to check if the chain is a chain leading back to the genesis block, right?

Oh, funny. With the ANRs I had I put it to a charger and it ran out of battery despite the fact it was connected and switched off. Now when I click on the bitcoin wallet button I get a toast saying App isn't installed.