Timeline for How much of the Bitcoin Source code is untouchable? [CVE-2018–17144]
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 7, 2019 at 23:10 | vote | accept | Paul Kania | ||
| S May 7, 2019 at 22:51 | history | suggested | Paul Kania | CC BY-SA 4.0 |
added clarification for when a previous version of bitcoin was released
|
| May 7, 2019 at 22:47 | review | Suggested edits | |||
| S May 7, 2019 at 22:51 | |||||
| May 7, 2019 at 21:44 | comment | added | G. Maxwell | The mistake here depended on the fact that people "knew" the history and original purpose of the code and didn't know that far away consensus rules had been weakened.I think it's informative to observe that "bitcoin-abc" rewrote double spend validation for their "ctor" change and managed to preserve the same bug and not detect it in the new test cases they added. The fact that double spending within a transaction isn't the same as general double spending is exactly the sort of cryptographic subtly that comes up and is fiendishly hard to get right. | |
| May 7, 2019 at 21:42 | comment | added | G. Maxwell | I think you also miss a bit of history here: In the original code this kind of double spend was only caught in block construction/connection and so there was a bug where the mempool could could contain these transactions (which was mostly harmless except they could show up unconfirmed in the wallet). This test was specifically added to protect the mempool and nothing else. But its importance was undocumentedly (and unintentionally?) increased by later changes. | |
| May 7, 2019 at 19:10 | history | answered | Pieter Wuille | CC BY-SA 4.0 |