Query: Patches break with Microsoft exchange server.
in [Kernel]
|
Prev: [RFC PATCH 1/2] debug_core: move all watch dog syncs to a single function
Next: x86: remove __phys_reloc_hide
From: Mihai Donțu on 9 Aug 2010 14:20 On Monday 09 August 2010 20:55:08 Justin P. Mattock wrote: > On 08/09/2010 07:35 AM, Mihai Donțu wrote: > > On Monday 09 August 2010 12:43:16 Justin P. Mattock wrote: > >> On 08/09/2010 02:35 AM, viresh kumar wrote: > >>> On 8/9/2010 2:31 PM, Matti Aarnio wrote: > >>>> On Mon, Aug 09, 2010 at 12:26:24PM +0530, viresh kumar wrote: > >>>>>> I missed this information in my last mail. We are using git > >>>>>> send-email for sending patches. As patches will go through > >>>>>> Microsoft exchange server only, so they are broken. > >>>> > >>>> Let your boss complain to your IT keepers. > >>>> "These are Machine-to-Machine messages, they must not be modified!" > >>>> > >>>> > >>>> It would probably be "against corporate policy" to use gmail for these > >>>> emails... > >>> > >>> We got one solution: Upgrade Exchange server to SP2. > >>> Lets see if our IT department does this upgradation. > >> > >> that or just blast them with some cryptology..i.e. pretty sure if your > >> message was encapsulated(AH/ESP) they couldn't tweak it.. but then > >> sending such encryption to a public list would require a _key_ on the > >> other side.. wishful thinking... > >> (just a thought)... > > > > Shouldn't just signing the message be enough? The server (normally) would > > not alter it, otherwise it will break the signature (which is a too > > obvious bug even for Microsoft). Or am I missing something here? > > > > PS: A local SMTP with DKIM signing capabilities could be another > > possibility, assuming Exchange does not break such signatures. > > yeah that would probably be just enough to get through without Microsoft > mucking around with the font etc.., but the biggest problem(I see) with > the encryption is having the key on the other end of the line. Wait. I don't think we're on the same page here. I'm talking about message signing (which does not require the receiving end to have any key - it's the same plain text e-mail with a blob after it) while you refer to actually encrypting the message. Mm? Or am I being extremely slow today? :-) -- Mihai Donțu -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Justin P. Mattock on 9 Aug 2010 15:00 On 08/09/2010 11:15 AM, Mihai Donțu wrote: > On Monday 09 August 2010 20:55:08 Justin P. Mattock wrote: >> On 08/09/2010 07:35 AM, Mihai Donțu wrote: >>> On Monday 09 August 2010 12:43:16 Justin P. Mattock wrote: >>>> On 08/09/2010 02:35 AM, viresh kumar wrote: >>>>> On 8/9/2010 2:31 PM, Matti Aarnio wrote: >>>>>> On Mon, Aug 09, 2010 at 12:26:24PM +0530, viresh kumar wrote: >>>>>>>> I missed this information in my last mail. We are using git >>>>>>>> send-email for sending patches. As patches will go through >>>>>>>> Microsoft exchange server only, so they are broken. >>>>>> >>>>>> Let your boss complain to your IT keepers. >>>>>> "These are Machine-to-Machine messages, they must not be modified!" >>>>>> >>>>>> >>>>>> It would probably be "against corporate policy" to use gmail for these >>>>>> emails... >>>>> >>>>> We got one solution: Upgrade Exchange server to SP2. >>>>> Lets see if our IT department does this upgradation. >>>> >>>> that or just blast them with some cryptology..i.e. pretty sure if your >>>> message was encapsulated(AH/ESP) they couldn't tweak it.. but then >>>> sending such encryption to a public list would require a _key_ on the >>>> other side.. wishful thinking... >>>> (just a thought)... >>> >>> Shouldn't just signing the message be enough? The server (normally) would >>> not alter it, otherwise it will break the signature (which is a too >>> obvious bug even for Microsoft). Or am I missing something here? >>> >>> PS: A local SMTP with DKIM signing capabilities could be another >>> possibility, assuming Exchange does not break such signatures. >> >> yeah that would probably be just enough to get through without Microsoft >> mucking around with the font etc.., but the biggest problem(I see) with >> the encryption is having the key on the other end of the line. > > Wait. I don't think we're on the same page here. I'm talking about message > signing (which does not require the receiving end to have any key - it's the > same plain text e-mail with a blob after it) while you refer to actually > encrypting the message. Mm? Or am I being extremely slow today? :-) > no were on the same page.. keep in mind though I'm not sure how the message signing thing really works, if it's just a signature verifying that it's from you without the other end(recipient) accepting anything, then the question is will microsoft still scan the email and garble it up? Now if it's a signature where the other end needs to accept the sender then im guessing there's a little bit of encryption there to keep microsoft database scanner from doing anything(but keep in mind I never really setup the signature thing on e-mails so I could totally be wrong) Justin P. Mattock -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: David Woodhouse on 9 Aug 2010 17:30 On Mon, 2010-08-09 at 21:15 +0300, Mihai Donțu wrote: > > Wait. I don't think we're on the same page here. I'm talking about message > signing (which does not require the receiving end to have any key - it's the > same plain text e-mail with a blob after it) while you refer to actually > encrypting the message. Mm? Or am I being extremely slow today? :-) Only when you assume that Exchange would pass signed messages without corrupting them. It really is that broken. -- dwmw2 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Justin P. Mattock on 9 Aug 2010 18:00 On 08/09/2010 02:28 PM, David Woodhouse wrote: > On Mon, 2010-08-09 at 21:15 +0300, Mihai Donțu wrote: >> >> Wait. I don't think we're on the same page here. I'm talking about message >> signing (which does not require the receiving end to have any key - it's the >> same plain text e-mail with a blob after it) while you refer to actually >> encrypting the message. Mm? Or am I being extremely slow today? :-) > > Only when you assume that Exchange would pass signed messages without > corrupting them. It really is that broken. > figured the encryption would be kind of a last resort situation..but if it's that broken to where it wont pass it along without corrupting, then the best solution is to figure out what Microsoft needs in terms of encoding, i.e. is there a way to have the scanner scan but not throw everything around after it scans.(if this is what it's doing) Justin P. Mattock -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Valeo de Vries on 9 Aug 2010 18:20
On 9 August 2010 22:56, Justin P. Mattock <justinmattock(a)gmail.com> wrote: > On 08/09/2010 02:28 PM, David Woodhouse wrote: >> >> On Mon, 2010-08-09 at 21:15 +0300, Mihai Donțu wrote: >>> >>> Wait. I don't think we're on the same page here. I'm talking about >>> message >>> signing (which does not require the receiving end to have any key - it's >>> the >>> same plain text e-mail with a blob after it) while you refer to actually >>> encrypting the message. Mm? Or am I being extremely slow today? :-) >> >> Only when you assume that Exchange would pass signed messages without >> corrupting them. It really is that broken. >> > > figured the encryption would be kind of a last resort situation..but if it's > that broken to where it wont pass it along without corrupting, then the best > solution is to figure out what Microsoft needs in terms of encoding, i.e. is > there a way to have the scanner scan but not throw everything around after > it scans.(if this is what it's doing) The link I posted earlier seems to give the impression that quoted-printable might do that. I may have misread that, though... Valeo -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ |