Transaction ID formation

[No_2]

I understand each transaction in the blockchain has a unique ID, is this generated when it is created or at the time it is written to the blockchain? Or is it generated by some other means?

[domob]

I understand each transaction in the blockchain has a unique ID, is this generated when it is created or at the time it is written to the blockchain? Or is it generated by some other means?

The tx ID is basically a hash of the transaction data.  So, in some sense, it is created as soon as you create the tx.  At least when it is broadcasted to the network, the ID becomes known publicly.  So not at mining time.

There's a small caveat, though:  Due to a process called "transaction malleability", it is possible that someone (who need not be related to the tx in any way) can alter the tx in such a way that it is still valid, "means" the same thing but has a different ID.  In that sense, you only know the (final) ID for sure once the tx is confirmed.

[No_2]

...
There's a small caveat, though:  Due to a process called "transaction malleability", it is possible that someone (who need not be related to the tx in any way) can alter the tx in such a way that it is still valid, "means" the same thing but has a different ID.  In that sense, you only know the (final) ID for sure once the tx is confirmed.

Thanks.

There was a lot of news about how transaction malleability had been fixed in Bitcoin Core 0.8/0.9 did this not resolve transaction malleability? Or are you saying one cannot assume that all nodes on the network are conforming to not causing transaction malleability when relaying transactions?

[Carlton Banks]

There was a lot of news about how transaction malleability had been fixed in Bitcoin Core 0.8/0.9 did this not resolve transaction malleability? Or are you saying one cannot assume that all nodes on the network are conforming to not causing transaction malleability when relaying transactions?

It wasn't fixed, but the range of circumstances under which it can be performed were mitigated somehow. And Core client reports when a doppelganger txid shows up in the mempool, whereas previously it reported both tx's as being valid (as well as reporting the wallet balance to be similarly valid  ) . I believe the BIP66 implementation in 0.10.0 is supposed to add a little more mitigation/step towards being fixed, but that's not even meaningful until the version 3 block creation threshold is reached.