| Summary: | RIPEMD160 hash signed messages send incorrect header with OpenPGP/MIME | ||
|---|---|---|---|
| Product: | [Unmaintained] kmail | Reporter: | Michael Trausch <fd0man> |
| Component: | encryption | Assignee: | kdepim bugs <kdepim-bugs> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | bugreports, cybdev, joke.de.buhr, julian, michael, mutz |
| Priority: | NOR | ||
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Platform: | Ubuntu | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed In: | ||
| Sentry Crash Report: | |||
| Attachments: |
RIPEMD160 signed message, OpenPGP/MIME, from Thunderbird/Enigmail
KMail message, OpenPGP/MIME, RIPEMD160 Error Output from Thunderbird on my KMail message Patch against kdepim 4.3.2 |
||
|
Description
Michael Trausch
2006-06-07 20:23:53 UTC
Created attachment 16513 [details]
RIPEMD160 signed message, OpenPGP/MIME, from Thunderbird/Enigmail
This message is signed with OpenPGP/MIME, using Thunderbird and Enigmail as the
front-end to GnuPG. The Content-Type header contains the right information
about the message digest algorithm.
Created attachment 16514 [details]
KMail message, OpenPGP/MIME, RIPEMD160
This message has an incorrect header. (KMail, however, says that it is okay.
Appears that KMail ignores the header?)
Created attachment 16515 [details] Error Output from Thunderbird on my KMail message This error is a result of viewing the message that is in attachment 16514 [details], which appears to have the header set to the wrong digest hash algorithm (SHA1, when it should be RIPEMD160). Should any more data be required, please do not hesitate to let me know! Confirming this problem with KMail (I'm currently using version 4.1.4) I'm having the same issues but with SHA-256 instead of RIPEMD160 (dupe from http://forum.kde.org/kmail-openpgp-mime-hash-in-micalg-is-wrong-t-27682.html ) Sending a mail with thunderbird gives a header such as this: Content-Type: multipart/signed; micalg=pgp-sha256; protocol="application/pgp-signature"; boundary="------------enig6C27345EB16199DD8B1DCF31" Sending a mail with KMail gives this header instead, but same key is used: Content-Type: multipart/signed; boundary="nextPart3110028.RrybVz1ln4"; protocol="application/pgp-signature"; micalg=pgp-sha1 Please note that this only applies to OpenPGP/MIME and not OpenPGP/inline as that works just fine. *** This bug has been confirmed by popular vote. *** I can confirm that KMail inappropriately generates "micalg=pgp-sha1" even when actually using a different digest algorithm to sign the message (here: SHA-512 or SHA-256). Given that by now, SHA-1 has been seriously compromised, using SHA-1 simply isn't an acceptable alternative anymore. What will it take to get this fixed in KMail? adding Marc to the CC on this. Maybe he knows more about it. Again, what will it take to get this fixed in KMail? Who is most well versed in the mechanics involved? Can we pay that someone to have this fixed? Any updates on this? From my understanding this can lead to the recipient's client complaining about an invalid signature, even if the sender using KMail did not change anything in his GnuPG configuration, because: 1) There are clients which complain when the declared hash algorithm does not match the hash algorithm that has been used. If I remember correctly I encountered this problem with Sylpheed (and this is how I got aware of this bug in the first place). 2) If a mail is both, encrypted and signed, GnuPG by default should use the hash algorithm that is preferred according to the recipient's key preferences. So if the recipient prefers SHA256 according to his key SHA256 should be used. However, KMail will still declare it as SHA1. With people moving away from SHA1, I think the priority of this bug should be increased. Created attachment 39112 [details]
Patch against kdepim 4.3.2
Retrieves the name of the signature hash algorithm from GpgME and places it in the micalg parameter of the Content-Type header for both PGP and S/MIME.
SVN commit 1065228 by tmcguire: Specify the correct algorithm which was used for creating the hash of the signature. Thanks to Michael Gorven for the patch! CCBUG: 128784 M +8 -2 messagecomposer.cpp M +4 -0 messagecomposer.h WebSVN link: http://websvn.kde.org/?view=rev&revision=1065228 SVN commit 1065232 by tmcguire: Backport r1065228 by tmcguire from trunk to the 4.3 branch: Specify the correct algorithm which was used for creating the hash of the signature. Thanks to Michael Gorven for the patch! BUG: 128784 M +8 -2 messagecomposer.cpp M +4 -0 messagecomposer.h WebSVN link: http://websvn.kde.org/?view=rev&revision=1065232 Thanks Michael for the patch, it will be in KDE 4.3.5 and KDE 4.4.0. If you have further patches, please use Reviewboard, see http://techbase.kde.org/Contribute/Send_Patches. It's easy to miss patches here on bugzilla, good thing I saw this one :) *** Bug 218721 has been marked as a duplicate of this bug. *** |