<oembed><type>rich</type><version>1.0</version><author_name>npub1m230cem2yh3mtdzkg32qhj73uytgkyg5ylxsu083n3tpjnajxx4qqa2np2</author_name><author_url>https://nostr.ae/npub1m230cem2yh3mtdzkg32qhj73uytgkyg5ylxsu083n3tpjnajxx4qqa2np2</author_url><provider_name>njump</provider_name><provider_url>https://nostr.ae</provider_url><html>📅 Original date posted:2017-02-23&#xA;📝 Original message:Worth noting: the impact of the SHA1 collison attack on Git is *not* limited&#xA;only to maintainers making maliciously colliding Git commits, but also&#xA;third-party&#39;s submitting pull-reqs containing commits, trees, and especially&#xA;files for which collisions have been found. This is likely to be exploitable in&#xA;practice with binary files, as reviewers aren&#39;t going to necessarily notice&#xA;garbage at the end of a file needed for the attack; if the attack can be&#xA;extended to constricted character sets like unicode or ASCII, we&#39;re in trouble&#xA;in general.&#xA;&#xA;Concretely, I could prepare a pair of files with the same SHA1 hash, taking&#xA;into account the header that Git prepends when hashing files. I&#39;d then submit&#xA;that pull-req to a project with the &#34;clean&#34; version of that file. Once the&#xA;maintainer merges my pull-req, possibly PGP signing the git commit, I then take&#xA;that signature and distribute the same repo, but with the &#34;clean&#34; version&#xA;replaced by the malicious version of the file.&#xA;&#xA;-- &#xA;https://petertodd.org &#39;peter&#39;[:-1]@petertodd.org&#xA;-------------- next part --------------&#xA;A non-text attachment was scrubbed...&#xA;Name: signature.asc&#xA;Type: application/pgp-signature&#xA;Size: 455 bytes&#xA;Desc: Digital signature&#xA;URL: &lt;http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20170223/fe8c9937/attachment.sig&gt;</html></oembed>