Re: [Monotone-devel] [stupid user question] Renaming a tag?
On Tue, Apr 01, 2008 at 01:34:26PM -0400, Zack Weinberg wrote: > As Bruce points out, it may be possible to request that everyone > manually delete the cert from their own repositories, but this is a > major nuisance for everyone (e.g. personally, this would involve > *finding* all the repositories I've got, on multiple computers). I did know about kill_tag_locally (and use it sometimes), but to expect it to be useful in the scenario where the tag was on a publicly readable repo is not reasonable IMO (even if not maliciously, it will someday get synced back to you) > And yes, this is something that is meant to be addressed within the > general "policy branches" umbrella. -- Part of why so many things are > bundled under that umbrella, by the way, is that they all may involve > a cert-reissuing flag day, and we try very hard to minimize the number > of those. A tangentional question about certs: why are there so many of them? mtn ls certs shows 4 certs on most revisions - author, branch, the changelog contents, and the date. It seems strange to generate and verify (and store and transmit!) signatures on these things, especially considering that they don't actually contain that much data. Why not combine them into a single revision object of author/originator date SHA-1 of changelog contents (contents treated like a file) branch keyid (Aside from avoiding a flag day, of course) -Jack ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] [stupid user question] Renaming a tag?
On Tue, Apr 1, 2008 at 5:05 PM, Julio M. Merino Vidal <[EMAIL PROTECTED]> wrote: > Even with policy branches, will that be possible? Tag certs, at the > moment, cannot be identified by anything else other than their contents > itself; i.e. they do not have an author nor date attached to them. This isn't quite true; like all certs, they name the key that signed them, which is sufficient for policy-based ignore rules. Revocation could be done without that; we could have "revoke certs" that identify some other cert by hash. zw ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] Re: fatal: Botan::PRNG_Unseeded
On Tue, Apr 01, 2008 at 08:43:41PM +0200, Markus Schiltknecht wrote: > Hi, > > Jack Lloyd wrote: > >Monotone uses EMSA3 (aka PKCS#1 v1.5), which is a deterministic > >encoding. So for a fixed input, the same key will always produce the > >same signature. (The generally accepted solution for RSA signature > >padding is EMSA4 aka PSS, which is a randomized method). > > Sorry if this is a stupid question, but what prevents us from to > switching to EMSA4? Nothing that I can think of, except that it would be very much a flag day in terms of cert signatures. -Jack ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] Re: fatal: Botan::PRNG_Unseeded
On Tue, Apr 1, 2008 at 2:43 PM, Markus Schiltknecht <[EMAIL PROTECTED]> wrote: > > Monotone uses EMSA3 (aka PKCS#1 v1.5), which is a deterministic > > encoding. So for a fixed input, the same key will always produce the > > same signature. (The generally accepted solution for RSA signature > > padding is EMSA4 aka PSS, which is a randomized method). > > Sorry if this is a stupid question, but what prevents us from to switching > to EMSA4? I'm not aware of any reason other than backward compatibility, but that's a doozy: this is yet another of the changes that would require a cert-reissuing event. (It would be *great* if certs had some kind of tag describing how they were signed, so we could make a change that affected new certs only, but I don't see any way to do that without a network protocol break.) zw ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] Re: fatal: Botan::PRNG_Unseeded
Hi, Jack Lloyd wrote: Monotone uses EMSA3 (aka PKCS#1 v1.5), which is a deterministic encoding. So for a fixed input, the same key will always produce the same signature. (The generally accepted solution for RSA signature padding is EMSA4 aka PSS, which is a randomized method). Sorry if this is a stupid question, but what prevents us from to switching to EMSA4? Regards Markus ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] [stupid user question] Renaming a tag?
On Apr 1, 2008, at 19:34 , Zack Weinberg wrote: On Tue, Apr 1, 2008 at 1:24 PM, Jack Lloyd <[EMAIL PROTECTED]> wrote: Is this not possible? I had a feeling it wasn't, and the manual suggests the same. Can add an alternative tag easily, of course, but as soon as a tag escapes 'into the wild' it is out there forever, yes? Yes. More generally, we have no automated mechanism at present to revoke or supersede any cert once it has been propagated. As Bruce points out, it may be possible to request that everyone manually delete the cert from their own repositories, but this is a major nuisance for everyone (e.g. personally, this would involve *finding* all the repositories I've got, on multiple computers). And yes, this is something that is meant to be addressed within the general "policy branches" umbrella. -- Part of why so many things are bundled under that umbrella, by the way, is that they all may involve a cert-reissuing flag day, and we try very hard to minimize the number of those. Even with policy branches, will that be possible? Tag certs, at the moment, cannot be identified by anything else other than their contents itself; i.e. they do not have an author nor date attached to them. So, how can you ban a tag cert? Ban the name? Or in policy branches, certs get an identifier (be author/date, or whatever) that can later be used for these purposes? ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] Re: fatal: Botan::PRNG_Unseeded
On Tue, Apr 01, 2008 at 11:18:22AM +0200, Lapo Luchini wrote: > Nathaniel Smith wrote: > >>What else are random numbers used for? As I understand it, signing > >>does not use random numbers, but maybe I'm confused. > > > >IIRC signatures do include random nonces -- I think this is one of the > >things I picked up from Paul's rants about our cryptography -- but I > >could be wrong too. > > AFAIR he was complaining about the very fact that we currently use > signatures that don't, and proposed to change 'em =) > Monotone uses EMSA3 (aka PKCS#1 v1.5), which is a deterministic encoding. So for a fixed input, the same key will always produce the same signature. (The generally accepted solution for RSA signature padding is EMSA4 aka PSS, which is a randomized method). -Jack ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] [stupid user question] Renaming a tag?
On Tue, Apr 1, 2008 at 1:24 PM, Jack Lloyd <[EMAIL PROTECTED]> wrote: > Is this not possible? I had a feeling it wasn't, and the manual > suggests the same. Can add an alternative tag easily, of course, but > as soon as a tag escapes 'into the wild' it is out there forever, yes? Yes. More generally, we have no automated mechanism at present to revoke or supersede any cert once it has been propagated. As Bruce points out, it may be possible to request that everyone manually delete the cert from their own repositories, but this is a major nuisance for everyone (e.g. personally, this would involve *finding* all the repositories I've got, on multiple computers). And yes, this is something that is meant to be addressed within the general "policy branches" umbrella. -- Part of why so many things are bundled under that umbrella, by the way, is that they all may involve a cert-reissuing flag day, and we try very hard to minimize the number of those. zw ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
[Monotone-devel] Re: [stupid user question] Renaming a tag?
Jack Lloyd <[EMAIL PROTECTED]> writes: > Is this not possible? I had a feeling it wasn't, and the manual > suggests the same. Can add an alternative tag easily, of course, but > as soon as a tag escapes 'into the wild' it is out there forever, > yes? Yes. Unless you can persuade everyone to delete it in their repositories, anyway. (I think this is one of the zillion issues that might be addressed by policy branches.) ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
[Monotone-devel] [stupid user question] Renaming a tag?
Is this not possible? I had a feeling it wasn't, and the manual suggests the same. Can add an alternative tag easily, of course, but as soon as a tag escapes 'into the wild' it is out there forever, yes? Jack ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] interface version / command matrix
On Tue, Apr 1, 2008 at 3:16 AM, <[EMAIL PROTECTED]> wrote: > > Quoting Daniel Atallah <[EMAIL PROTECTED]>: > > > On Sun, Mar 30, 2008 at 2:24 AM, Stephen Leake <[EMAIL PROTECTED]> wrote: > >> For me and Emacs DVC, I think ssh-agent is the best solution to that. > >> That's on my todo list (ssh-agent doesn't work with mtn on Win32, at > >> least not generally). > > > > What about it have you seen not working? I've been using it since I > > submitted the patch to implement it in September 2007. > > It requires a particular ssh-agent implementation, and looks for it in > a particular way. Ah, yes it does, I tied it to Putty's Pageant as that seems to be the closest to a "native" implementation on Windows. > I use Cygwin ssh-agent. monotone doesn't find that by default. Are you using a cygwin monotone or the native win32 build? I wouldn't expect the cygwin build to use the Pageant ssh-agent implementation, but perhaps it does (I don't use the cygwin binary). > One of these days we need to put in a Lua hook to configure that. That would be a nice thing to be able to configure, alternatively, it shouldn't be too tricky to to update the win32 ssh-agent implementation to support both the Pageant and the cygwin ssh-agent. -D ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] mtn:// sync
On Tue, 2008-04-01 at 14:22 +0200, Ulf Ochsenfahrt wrote: > Richard Levitte wrote: > > In message <[EMAIL PROTECTED]> > on Thu, 20 Mar 2008 15:08:56 +, Peter Stirling > <[EMAIL PROTECTED]> said: > > > > peter> Is using '&' a good idea > > > > & is the standard parameter divisor for parametrised URLs, so you > > can't get away from that without getting trouble... > > I've just found out that the semicolon ';' is also an acceptable > parameter separator, according to the HTML 4 spec. > > See here: > http://www.w3.org/TR/html4/appendix/notes.html#h-B.2.2 ';' has the same shell-related annoyances as '&' does, so it isn't really an improvement. $ echo foo;bar foo bash: bar: command not found ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] mtn:// sync
Ulf Ochsenfahrt schrieb: I've just found out that the semicolon ';' is also an acceptable parameter separator, according to the HTML 4 spec. See here: http://www.w3.org/TR/html4/appendix/notes.html#h-B.2.2 The Semikolon also works as command separator on various unix shells. A URL would then have to be enclosed in quotation marks. ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] mtn:// sync
Richard Levitte wrote: > In message <[EMAIL PROTECTED]> on Thu, 20 Mar 2008 15:08:56 +, Peter Stirling <[EMAIL PROTECTED]> said: > > peter> Is using '&' a good idea > > & is the standard parameter divisor for parametrised URLs, so you > can't get away from that without getting trouble... I've just found out that the semicolon ';' is also an acceptable parameter separator, according to the HTML 4 spec. See here: http://www.w3.org/TR/html4/appendix/notes.html#h-B.2.2 Cheers, -- Ulf smime.p7s Description: S/MIME Cryptographic Signature ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
[Monotone-devel] Re: fatal: Botan::PRNG_Unseeded
Nathaniel Smith wrote: What else are random numbers used for? As I understand it, signing does not use random numbers, but maybe I'm confused. IIRC signatures do include random nonces -- I think this is one of the things I picked up from Paul's rants about our cryptography -- but I could be wrong too. AFAIR he was complaining about the very fact that we currently use signatures that don't, and proposed to change 'em =) Lapo ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] Non-anonymous pulls
Quoting Markus Schiltknecht <[EMAIL PROTECTED]>: Hi, Jack Lloyd wrote: Wouldn't it make sense though for a pull to (if nothing else) retry using available keys if anonymous access were refused? (Ie, why make the user do it when Monotone can easily guess). Sure. That has been proposed before, and we know it would make sense. However, nobody got around implementing it, yet. I think it would be better to store the key in _MTN/options if --set-default is specified on the 'pull' command line, and then also use the key from _MTN/options if it is there. That's easy to implement; I just did it for sync. -- Stephe ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel
Re: [Monotone-devel] interface version / command matrix
Quoting Daniel Atallah <[EMAIL PROTECTED]>: On Sun, Mar 30, 2008 at 2:24 AM, Stephen Leake <[EMAIL PROTECTED]> wrote: For me and Emacs DVC, I think ssh-agent is the best solution to that. That's on my todo list (ssh-agent doesn't work with mtn on Win32, at least not generally). What about it have you seen not working? I've been using it since I submitted the patch to implement it in September 2007. It requires a particular ssh-agent implementation, and looks for it in a particular way. I use Cygwin ssh-agent. monotone doesn't find that by default. One of these days we need to put in a Lua hook to configure that. For now I just hard-code my passphrase in the passphrase Lua hook; there are other things to fix that are more of a problem. -- Stephe ___ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel