{"thread":{"id":"58681","subject":"[PATCH] docs: clarify that credential discards unrecognised attributes","startedAt":"2022-10-24T07:57:58Z","lastAt":"2022-11-14T22:40:25Z","messageCount":13,"participants":["M Hickford via GitGitGadget","Jeff King","M Hickford","Bagas Sanjaya","Junio C Hamano","Taylor Blau"],"isPatch":true,"patchVersion":1,"patchTotal":null},"messages":[{"id":"465608","messageId":"pull.1393.git.1666598268697.gitgitgadget@gmail.com","threadId":"58681","inReplyTo":null,"subject":"[PATCH] docs: clarify that credential discards unrecognised attributes","fromName":"M Hickford via GitGitGadget","fromEmail":"gitgitgadget@gmail.com","sentAt":"2022-10-24T07:57:48Z","receivedAt":"2022-10-24T07:57:58Z","isPatch":true,"sender":{"key":"mirth.hickford@gmail.com","avatar":"https://avatars.githubusercontent.com/u/105314?v=4"},"body":"From: M Hickford \n\nIt was previously unclear how unrecognised attributes are handled.\n\nSigned-off-by: M Hickford \n---\n docs: clarify that credential discards unrecognised attributes\n \n It was previously unclear how unrecognised attributes are handled.\n \n Signed-off-by: M Hickford mirth.hickford@gmail.com\n\nPublished-As: https://github.com/gitgitgadget/git/releases/tag/pr-1393%2Fhickford%2Funrecognised-attributes-v1\nFetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-1393/hickford/unrecognised-attributes-v1\nPull-Request: https://github.com/gitgitgadget/git/pull/1393\n\n Documentation/git-credential.txt | 2 ++\n Documentation/gitcredentials.txt | 1 +\n 2 files changed, 3 insertions(+)\n\ndiff --git a/Documentation/git-credential.txt b/Documentation/git-credential.txt\nindex f18673017f5..ac2818b9f66 100644\n--- a/Documentation/git-credential.txt\n+++ b/Documentation/git-credential.txt\n@@ -160,6 +160,8 @@ empty string.\n Components which are missing from the URL (e.g., there is no\n username in the example above) will be left unset.\n \n+Unrecognised attributes are silently discarded.\n+\n GIT\n ---\n Part of the linkgit:git[1] suite\ndiff --git a/Documentation/gitcredentials.txt b/Documentation/gitcredentials.txt\nindex 80517b4eb2c..e856e3c8330 100644\n--- a/Documentation/gitcredentials.txt\n+++ b/Documentation/gitcredentials.txt\n@@ -269,6 +269,7 @@ stdout in the same format (see linkgit:git-credential[1] for common\n attributes). A helper is free to produce a subset, or even no values at\n all if it has nothing useful to provide. Any provided attributes will\n overwrite those already known about by Git's credential subsystem.\n+Unrecognised attributes are silently discarded.\n \n While it is possible to override all attributes, well behaving helpers\n should refrain from doing so for any attribute other than username and\n\nbase-commit: 1fc3c0ad407008c2f71dd9ae1241d8b75f8ef886\n-- \ngitgitgadget\n"},{"id":"465675","messageId":"Y1cm2a3qy9aowwNh@coredump.intra.peff.net","threadId":"58681","inReplyTo":"pull.1393.git.1666598268697.gitgitgadget@gmail.com","subject":"Re: [PATCH] docs: clarify that credential discards unrecognised attributes","fromName":"Jeff King","fromEmail":"peff@peff.net","sentAt":"2022-10-24T23:59:21Z","receivedAt":"2022-10-25T01:03:08Z","isPatch":true,"sender":{"key":"peff@peff.net","avatar":"https://avatars.githubusercontent.com/u/45925?v=4"},"body":"On Mon, Oct 24, 2022 at 07:57:48AM +0000, M Hickford via GitGitGadget wrote:\n\n> It was previously unclear how unrecognised attributes are handled.\n\nYeah, this was always part of the intended behavior, but I agree we did\nnot say it very explicitly (aside from an in-code comment!). Both the\nintent and content of your patch look good to me.\n\nWe did discuss patches a long time ago that would let Git carry\narbitrary keys between helpers, even if Git itself didn't understand it.\nOne of the intended uses was to let helpers talk to each other about\nTTLs. So if you had say:\n\n [credential]\n helper = generate-some-token\n helper = cache\n\nwhere the first helper generates a token, and the second caches it, the\nfirst one could shove a \"ttl\" or \"expiration\" key into the protocol,\nwhich the cache could then learn to respect.\n\nBut we never merged such a thing, and in practice I think people would\njust implement both parts as a single helper for simplicity. And anyway,\neven if we did want to do that, such a patch would want to modify the\ndocumentation to explain the new behavior. :)\n\n-Peff\n"},{"id":"465676","messageId":"Y1cnE5cyN1NBdmCv@coredump.intra.peff.net","threadId":"58681","inReplyTo":"Y1cm2a3qy9aowwNh@coredump.intra.peff.net","subject":"Re: [PATCH] docs: clarify that credential discards unrecognised attributes","fromName":"Jeff King","fromEmail":"peff@peff.net","sentAt":"2022-10-25T00:00:19Z","receivedAt":"2022-10-25T01:03:23Z","isPatch":true,"sender":{"key":"peff@peff.net","avatar":"https://avatars.githubusercontent.com/u/45925?v=4"},"body":"On Mon, Oct 24, 2022 at 07:59:22PM -0400, Jeff King wrote:\n\n> We did discuss patches a long time ago that would let Git carry\n> arbitrary keys between helpers, even if Git itself didn't understand it.\n> One of the intended uses was to let helpers talk to each other about\n> TTLs. So if you had say:\n> \n> [credential]\n> helper = generate-some-token\n> helper = cache\n> \n> where the first helper generates a token, and the second caches it, the\n> first one could shove a \"ttl\" or \"expiration\" key into the protocol,\n> which the cache could then learn to respect.\n\nIn case anyone is morbidly curious, it was in this thread:\n\n https://lore.kernel.org/git/20120407033417.GA13914@sigill.intra.peff.net/\n\n-Peff\n"},{"id":"465677","messageId":"20221025015116.4730-1-mirth.hickford@gmail.com","threadId":"58681","inReplyTo":"Y1cm2a3qy9aowwNh@coredump.intra.peff.net","subject":"Thanks","fromName":"M Hickford","fromEmail":"mirth.hickford@gmail.com","sentAt":"2022-10-25T01:51:16Z","receivedAt":"2022-10-25T01:52:02Z","isPatch":false,"sender":{"key":"mirth.hickford@gmail.com","avatar":"https://avatars.githubusercontent.com/u/105314?v=4"},"body":"Thanks Jeff for your reply. This is helpful to understand the background.\n\n(first message using git send-email, hopefully I followed the instructions correctly)\n"},{"id":"465688","messageId":"84143b0c-139b-63c5-519f-8d4a44254976@gmail.com","threadId":"58681","inReplyTo":"20221025015116.4730-1-mirth.hickford@gmail.com","subject":"Re: Thanks","fromName":"Bagas Sanjaya","fromEmail":"bagasdotme@gmail.com","sentAt":"2022-10-25T09:05:56Z","receivedAt":"2022-10-25T09:12:31Z","isPatch":false,"sender":{"key":"bagasdotme@gmail.com","avatar":"https://avatars.githubusercontent.com/u/40219486?v=4"},"body":"On 10/25/22 08:51, M Hickford wrote:\n> Thanks Jeff for your reply. This is helpful to understand the background.\n> \n> (first message using git send-email, hopefully I followed the instructions correctly)\n\nYou messed up the thread (you broke it).\n\nOn every message showed on lore.kernel.org/git, there is reply\ninstructions at the bottom of the message's page. Follow it closely.\n\nIf you insisted on either using git-send-email(1) or mailto: link\nto reply, you need to manually quote any portion of original message\nyou wish to give the context. Make sure every line of quoted text\nstarts with `> `.\n\nBetter yet, reply using \"reply-all\" button or key in your mail client.\nHowever, make sure your email is text/plain (not HTML) and your mail\nclient doesn't mess up with whitespaces (convert tabs to spaces,\nword wrapping).\n\nThanks.\n\n-- \nAn old man doll... just what I always wanted! - Clara\n\n"},{"id":"465750","messageId":"20221026043905.1654-1-mirth.hickford@gmail.com","threadId":"58681","inReplyTo":"84143b0c-139b-63c5-519f-8d4a44254976@gmail.com","subject":"Re: Thanks","fromName":"M Hickford","fromEmail":"mirth.hickford@gmail.com","sentAt":"2022-10-26T04:39:05Z","receivedAt":"2022-10-26T04:39:16Z","isPatch":false,"sender":{"key":"mirth.hickford@gmail.com","avatar":"https://avatars.githubusercontent.com/u/105314?v=4"},"body":"> > (first message using git send-email, hopefully I followed the instructions correctly)\n>\n> You messed up the thread (you broke it).\n\nCurious. The thread overview \"5+ messages\" at https://lore.kernel.org/git/20221025015116.4730-1-mirth.hickford@gmail.com/#related looks okay to me. I followed the git send-email instructions setting the In-Reply-To header. I changed the subject -- maybe that confused some clients? Thanks for the tips.\n"},{"id":"465751","messageId":"Y1jDMW4ST5UEjKbC@coredump.intra.peff.net","threadId":"58681","inReplyTo":"20221026043905.1654-1-mirth.hickford@gmail.com","subject":"Re: Thanks","fromName":"Jeff King","fromEmail":"peff@peff.net","sentAt":"2022-10-26T05:18:41Z","receivedAt":"2022-10-26T05:18:57Z","isPatch":false,"sender":{"key":"peff@peff.net","avatar":"https://avatars.githubusercontent.com/u/45925?v=4"},"body":"On Wed, Oct 26, 2022 at 05:39:05AM +0100, M Hickford wrote:\n\n> > > (first message using git send-email, hopefully I followed the instructions correctly)\n> >\n> > You messed up the thread (you broke it).\n> \n> Curious. The thread overview \"5+ messages\" at\n> https://lore.kernel.org/git/20221025015116.4730-1-mirth.hickford@gmail.com/#related\n> looks okay to me. I followed the git send-email instructions setting\n> the In-Reply-To header. I changed the subject -- maybe that confused\n> some clients?\n\nYour reply looked fine to me. The in-reply-to you used is correct. It's\nOK to change the subject (though usually only done when the conversation\ndrifts to a different topic).\n\n-Peff\n"},{"id":"465760","messageId":"xmqqeduvkjte.fsf@gitster.g","threadId":"58681","inReplyTo":"20221026043905.1654-1-mirth.hickford@gmail.com","subject":"Re: Thanks","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2022-10-26T09:36:45Z","receivedAt":"2022-10-26T09:36:55Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"M Hickford writes:\n\n>> > (first message using git send-email, hopefully I followed the instructions correctly)\n>>\n>> You messed up the thread (you broke it).\n>\n> Curious. The thread overview \"5+ messages\" at\n> https://lore.kernel.org/git/20221025015116.4730-1-mirth.hickford@gmail.com/#related\n> looks okay to me. I followed the git send-email instructions\n> setting the In-Reply-To header. I changed the subject -- maybe\n> that confused some clients? Thanks for the tips.\n\nYour In-Reply-To: header looked OK to me. Some webmail platforms\nbreak the thread when Subject: is edited, but that is their bug. As\nlong as you keep the In-Reply-To: chain correct, retitling the\nmessage as necessary is indeed encouraged practice.\n\n"},{"id":"467186","messageId":"CAGJzqsmH0DpEZDVZYU=6XpBO_vPfbirrK7KaWm2RmKG3aAk0oA@mail.gmail.com","threadId":"58681","inReplyTo":"Y1cm2a3qy9aowwNh@coredump.intra.peff.net","subject":"Re: [PATCH] docs: clarify that credential discards unrecognised attributes","fromName":"M Hickford","fromEmail":"mirth.hickford@gmail.com","sentAt":"2022-11-12T02:21:24Z","receivedAt":"2022-11-12T02:22:05Z","isPatch":true,"sender":{"key":"mirth.hickford@gmail.com","avatar":"https://avatars.githubusercontent.com/u/105314?v=4"},"body":"On Tue, 25 Oct 2022 at 00:59, Jeff King wrote:\n>\n> On Mon, Oct 24, 2022 at 07:57:48AM +0000, M Hickford via GitGitGadget wrote:\n>\n> > It was previously unclear how unrecognised attributes are handled.\n>\n> Yeah, this was always part of the intended behavior, but I agree we did\n> not say it very explicitly (aside from an in-code comment!). Both the\n> intent and content of your patch look good to me.\n\nThanks. What happens next? I should look for this change in the seen\nbranch? https://git-scm.com/docs/MyFirstContribution#after-approval\n\n> We did discuss patches a long time ago that would let Git carry\n> arbitrary keys between helpers, even if Git itself didn't understand it.\n> One of the intended uses was to let helpers talk to each other about\n> TTLs. So if you had say:\n>\n> [credential]\n> helper = generate-some-token\n> helper = cache\n>\n> where the first helper generates a token, and the second caches it, the\n> first one could shove a \"ttl\" or \"expiration\" key into the protocol,\n> which the cache could then learn to respect.\n\nComposing helpers like this is how I encourage users to configure\ngit-credential-oauth [1][2]. Note that the storage helper should come\n*before* the generator, so that `credential fill` finds a stored\ncredential before it generates a fresh credential.\n\n> the first one could shove a \"ttl\" or \"expiration\" key into the protocol,\n> which the cache could then learn to respect.\n>\n> But we never merged such a thing, and in practice I think people would\n> just implement both parts as a single helper for simplicity.\n\nComposing helpers has the advantage that the user can choose their\npreferred storage. Generated credentials aren't necessarily short\nlived. GitHub OAuth tokens, for example, are good for at least one\nyear [3].\n\n[1] https://lore.kernel.org/git/CAGJzqs=+fCQzkDX53H8Mz-DjXicVVgRmmzPjkatSiOpYO7wGGA@mail.gmail.com/\n[2] https://github.com/hickford/git-credential-oauth\n[3] https://docs.github.com/en/authentication/keeping-your-account-and-data-secure/token-expiration-and-revocation#token-expired-due-to-lack-of-use\n"},{"id":"467202","messageId":"Y2/OIrSz+xrqkd+C@coredump.intra.peff.net","threadId":"58681","inReplyTo":"CAGJzqsmH0DpEZDVZYU=6XpBO_vPfbirrK7KaWm2RmKG3aAk0oA@mail.gmail.com","subject":"Re: [PATCH] docs: clarify that credential discards unrecognised attributes","fromName":"Jeff King","fromEmail":"peff@peff.net","sentAt":"2022-11-12T16:47:30Z","receivedAt":"2022-11-12T16:47:34Z","isPatch":true,"sender":{"key":"peff@peff.net","avatar":"https://avatars.githubusercontent.com/u/45925?v=4"},"body":"On Sat, Nov 12, 2022 at 02:21:24AM +0000, M Hickford wrote:\n\n> On Tue, 25 Oct 2022 at 00:59, Jeff King wrote:\n> >\n> > On Mon, Oct 24, 2022 at 07:57:48AM +0000, M Hickford via GitGitGadget wrote:\n> >\n> > > It was previously unclear how unrecognised attributes are handled.\n> >\n> > Yeah, this was always part of the intended behavior, but I agree we did\n> > not say it very explicitly (aside from an in-code comment!). Both the\n> > intent and content of your patch look good to me.\n> \n> Thanks. What happens next? I should look for this change in the seen\n> branch? https://git-scm.com/docs/MyFirstContribution#after-approval\n\nUsually the maintainer would pick it up, it would end up in seen, then\neventually 'next', and then eventually 'master'. You can check the\nperiodic \"What's Cooking\" messages from the maintainer to see more\ndiscussion of various topic branches.\n\nIn this case, though, I don't see any indication that the maintainer\npicked saw it. It sometimes happens that a topic is simply overlooked,\neven if it received positive reviews.\n\nThe usual thing to do is repost it, cc-ing the maintainer. I've also\ncc'd the interim maintainer here, so that may get things moving. :)\n\n> > We did discuss patches a long time ago that would let Git carry\n> > arbitrary keys between helpers, even if Git itself didn't understand it.\n> > One of the intended uses was to let helpers talk to each other about\n> > TTLs. So if you had say:\n> >\n> > [credential]\n> > helper = generate-some-token\n> > helper = cache\n> >\n> > where the first helper generates a token, and the second caches it, the\n> > first one could shove a \"ttl\" or \"expiration\" key into the protocol,\n> > which the cache could then learn to respect.\n> \n> Composing helpers like this is how I encourage users to configure\n> git-credential-oauth [1][2]. Note that the storage helper should come\n> *before* the generator, so that `credential fill` finds a stored\n> credential before it generates a fresh credential.\n\nRight, it's been a while since I've constructed an example like this. ;)\n\nWhat you're doing works fine with the code as-is; you just can't carry\nextra data (like a ttl) between the two.\n\nThe thread I linked earlier also discusses (in the very top-level patch)\na change in behavior that would break the flow you're relying on here\n(because it may unexpectedly propagate credentials between helpers). But\nI don't think anybody is interested in pursuing that, and it has been 10\nyears now.\n\n> > But we never merged such a thing, and in practice I think people would\n> > just implement both parts as a single helper for simplicity.\n> \n> Composing helpers has the advantage that the user can choose their\n> preferred storage. Generated credentials aren't necessarily short\n> lived. GitHub OAuth tokens, for example, are good for at least one\n> year [3].\n\nYeah, the composability was one of the goals of the system. I just think\nin practice that not many people use it. You can also compose outside of\nGit (I think the thread I linked earlier has an example of a wrapper\nthat does so), but again, I don't think anybody really does so in\npractice.\n\nI agree for GitHub's tokens that the times involved make auto-expiration\nnot that important. The example back in that thread was something more\ntime-limited (like minutes or hours). I don't know how often that kind\nof things is in the wild.\n\n-Peff\n"},{"id":"467206","messageId":"CAGJzqsnd3UbCuj6xfsAgVEFz5OaDBhnCTNeAvOVPK9NWBDnDRA@mail.gmail.com","threadId":"58681","inReplyTo":"Y2/OIrSz+xrqkd+C@coredump.intra.peff.net","subject":"Re: [PATCH] docs: clarify that credential discards unrecognised attributes","fromName":"M Hickford","fromEmail":"mirth.hickford@gmail.com","sentAt":"2022-11-12T19:08:42Z","receivedAt":"2022-11-12T19:09:23Z","isPatch":true,"sender":{"key":"mirth.hickford@gmail.com","avatar":"https://avatars.githubusercontent.com/u/105314?v=4"},"body":"On Sat, 12 Nov 2022 at 16:47, Jeff King wrote:\n> > > We did discuss patches a long time ago that would let Git carry\n> > > arbitrary keys between helpers, even if Git itself didn't understand it.\n> > > One of the intended uses was to let helpers talk to each other about\n> > > TTLs. So if you had say:\n> > >\n> > > [credential]\n> > > helper = generate-some-token\n> > > helper = cache\n> > >\n> > > where the first helper generates a token, and the second caches it, the\n> > > first one could shove a \"ttl\" or \"expiration\" key into the protocol,\n> > > which the cache could then learn to respect.\n> >\n>\n> What you're doing works fine with the code as-is; you just can't carry\n> extra data (like a ttl) between the two.\n\nFWIW I have a draft patch that adds password_expiry_utc and\noauth_refresh_token attributes to credential\nhttps://github.com/gitgitgadget/git/pull/1394 introducing expiry logic\nin the credential layer. I'll share a RFC sometime in future.\n\n> I agree for GitHub's tokens that the times involved make auto-expiration\n> not that important. The example back in that thread was something more\n> time-limited (like minutes or hours). I don't know how often that kind\n> of things is in the wild.\n\nGitLab OAuth tokens expire after 2 hours (the refresh tokens are valid\nlonger). This is a security improvement over long-lived tokens.\n"},{"id":"467213","messageId":"Y3B48E04XfBiSz2V@nand.local","threadId":"58681","inReplyTo":"Y2/OIrSz+xrqkd+C@coredump.intra.peff.net","subject":"Re: [PATCH] docs: clarify that credential discards unrecognised attributes","fromName":"Taylor Blau","fromEmail":"me@ttaylorr.com","sentAt":"2022-11-13T04:56:16Z","receivedAt":"2022-11-13T04:56:23Z","isPatch":true,"sender":{"key":"me@ttaylorr.com","avatar":"https://avatars.githubusercontent.com/u/301000140?v=4"},"body":"On Sat, Nov 12, 2022 at 11:47:30AM -0500, Jeff King wrote:\n> On Sat, Nov 12, 2022 at 02:21:24AM +0000, M Hickford wrote:\n>\n> > On Tue, 25 Oct 2022 at 00:59, Jeff King wrote:\n> > >\n> > > On Mon, Oct 24, 2022 at 07:57:48AM +0000, M Hickford via GitGitGadget wrote:\n> > >\n> > > > It was previously unclear how unrecognised attributes are handled.\n> > >\n> > > Yeah, this was always part of the intended behavior, but I agree we did\n> > > not say it very explicitly (aside from an in-code comment!). Both the\n> > > intent and content of your patch look good to me.\n> >\n> > Thanks. What happens next? I should look for this change in the seen\n> > branch? https://git-scm.com/docs/MyFirstContribution#after-approval\n>\n> Usually the maintainer would pick it up, it would end up in seen, then\n> eventually 'next', and then eventually 'master'. You can check the\n> periodic \"What's Cooking\" messages from the maintainer to see more\n> discussion of various topic branches.\n>\n> In this case, though, I don't see any indication that the maintainer\n> picked saw it. It sometimes happens that a topic is simply overlooked,\n> even if it received positive reviews.\n>\n> The usual thing to do is repost it, cc-ing the maintainer. I've also\n> cc'd the interim maintainer here, so that may get things moving. :)\n\nI think that Junio may have missed it when picking up topics. But it\nhappened before I start combing the list more finely, so I had already\npurged it from my inbox without picking it up, either.\n\nI'll take a look now and queue it to start merging down. Thanks for the\nnudge.\n\nThanks,\nTaylor\n"},{"id":"467277","messageId":"Y3LD1Ilq8kLPlJMO@coredump.intra.peff.net","threadId":"58681","inReplyTo":"CAGJzqsnd3UbCuj6xfsAgVEFz5OaDBhnCTNeAvOVPK9NWBDnDRA@mail.gmail.com","subject":"Re: [PATCH] docs: clarify that credential discards unrecognised attributes","fromName":"Jeff King","fromEmail":"peff@peff.net","sentAt":"2022-11-14T22:40:20Z","receivedAt":"2022-11-14T22:40:25Z","isPatch":true,"sender":{"key":"peff@peff.net","avatar":"https://avatars.githubusercontent.com/u/45925?v=4"},"body":"On Sat, Nov 12, 2022 at 07:08:42PM +0000, M Hickford wrote:\n\n> On Sat, 12 Nov 2022 at 16:47, Jeff King wrote:\n> > > > We did discuss patches a long time ago that would let Git carry\n> > > > arbitrary keys between helpers, even if Git itself didn't understand it.\n> > > > One of the intended uses was to let helpers talk to each other about\n> > > > TTLs. So if you had say:\n> > > >\n> > > > [credential]\n> > > > helper = generate-some-token\n> > > > helper = cache\n> > > >\n> > > > where the first helper generates a token, and the second caches it, the\n> > > > first one could shove a \"ttl\" or \"expiration\" key into the protocol,\n> > > > which the cache could then learn to respect.\n> > >\n> >\n> > What you're doing works fine with the code as-is; you just can't carry\n> > extra data (like a ttl) between the two.\n> \n> FWIW I have a draft patch that adds password_expiry_utc and\n> oauth_refresh_token attributes to credential\n> https://github.com/gitgitgadget/git/pull/1394 introducing expiry logic\n> in the credential layer. I'll share a RFC sometime in future.\n\nNeat.\n\nI'm not _totally_ opposed to introducing these as something Git\nunderstands, but I think it makes more sense to just teach Git to relay\nunknown entries between helpers.\n\nThe oauth thing is going to be very helper specific, and not something I\nthink Git would ever do anything with itself.\n\nIn theory Git might care about expiration, but in practice I think it\ndoesn't. It's very unlikely for a token to expire in the course of Git\nusing it. It's only much later, when we ask for it back, that a helper\nwill notice it's expired. Git could save the helper some work by\nnoticing this on read, but since the helper has to learn to store and\nreport the expiration in the first place, not much is gained.\n\nAnd in the case of something like credential-cache, we want to do more\nthan just store; we'd actually drop the credential entirely (and maybe\neven cause the daemon to exit) if it expires before the usual timeout.\n\n-Peff\n"}]}