Still LGTM on my end as well! On Thu, Aug 17, 2023 at 6:05 PM Julio Cardozo <[email protected]> wrote:
> Obrigado pela atençao de voces > > Em qui., 17 de ago. de 2023 11:28, Andrew Williams <[email protected]> > escreveu: > >> Thanks Daniel and Chris! I've enabled this change in the Chrome >> tip-of-tree and updated the chromestatus.com entry accordingly >> >> -Andrew >> >> On Wed, Aug 16, 2023 at 11:32 AM Chris Harrelson <[email protected]> >> wrote: >> >>> Same here. No need to wait for further approvals, go ahead with the >>> subset (but please don't forget to update the chromestatus.com entry >>> accordingly). >>> >>> On Wed, Aug 16, 2023 at 8:24 AM Daniel Bratell <[email protected]> >>> wrote: >>> >>>> My LGTM still stands for the remaining subset. I don't think you need a >>>> new thread for it. Good detecting to figure all of that out. >>>> >>>> /Daniel >>>> On 2023-08-16 17:11, Andrew Williams wrote: >>>> >>>> Hey everyone, >>>> >>>> A few updates on this - TL;DR: we'd like to proceed with a subset of >>>> what was originally proposed. >>>> >>>> - During implementation we found some differences in how cookies >>>> containing \x00, \x0D, and \x0A characters are handled when set via >>>> JavaScript and set via Set-Cookie headers. For cookies set via headers, >>>> Chrome rejects the entire HTTP response if a \x00 character is present. For >>>> the \x0D and \x0A characters, the HTTP header parsing code treats each >>>> character as if it was a \x0D\x0A (CRLF) character sequence, truncating >>>> Set-Cookie header strings that contain these characters before the cookie >>>> parsing code gets them. This behavior is not specific to parsing the >>>> Set-Cookie header, and seems counter to the spec >>>> <https://www.rfc-editor.org/rfc/rfc9110.html#fields.values:~:text=Field%20values%20containing,downstream%20HTTP%20parser).>. >>>> It >>>> would be great to change this if possible, but we would need new metrics to >>>> gauge how common this is since our existing metrics were only collected >>>> from the point of view of the cookie parsing code. Also, we'd want a new >>>> blink-dev thread for this, since we'd likely want to change this behavior >>>> for header parsing in general. We'd like to defer on changing the header >>>> parsing behavior for now. One step we did take towards this is updating the >>>> existing cookie control character handling WPTs to test setting via headers >>>> as well so that behavior by other browsers can be seen >>>> <https://wpt.fyi/results/cookies/value/value-ctl.html?label=master&label=experimental&aligned> >>>> . >>>> >>>> - Whether to allow the tab character inside of a cookie name, value, >>>> attribute name, or attribute value (i.e. internal tabs) is currently >>>> <https://github.com/httpwg/http-extensions/issues/2262> being >>>> discussed and the spec position might be reconsidered. We'd like to defer >>>> on changing Chrome's current handling of this until that is resolved. Since >>>> Chrome blocks cookies containing internal tabs today, changing it to allow >>>> cookies with internal tabs should cause minimal breakage, so it should be >>>> safe to accompany this change with a blink-dev PSA. >>>> >>>> - What's remaining is to block cookies set via Javascript from >>>> containing \x00, \x0D, and \x0A characters. For this we've had deprecation >>>> reporting and DevTools warnings (both implemented via a call to >>>> CountDeprecation) in place for a while, and we recently added an enterprise >>>> policy so that enterprises can disable this change if needed once it's >>>> enabled. Also, we re-checked the existing metrics for this recently and >>>> usage is less than it was in our previous analysis for all characters. >>>> >>>> Do we need a completely new I2S for shipping the subset of these >>>> changes (affecting only cookies set via JavaScript), or can we receive new >>>> LGTMs on this thread and proceed with the rollout afterward? >>>> >>>> Thanks! >>>> >>>> -Andrew >>>> >>>> On Fri, Dec 3, 2021 at 2:43 PM Andrew Williams <[email protected]> >>>> wrote: >>>> >>>>> Awesome, thanks for the feedback, Yoav. >>>>> >>>>> The Chrome Enterprise team also recommended having an enterprise >>>>> policy in for this change just in case, and they recommended having it be >>>>> available for at least a couple of milestones after we've implemented the >>>>> change. We will plan on that as well. >>>>> >>>>> I'll post back here once we have an update on implementing the >>>>> DevTools Issue + deprecation reporting. >>>>> >>>>> -Andrew >>>>> >>>>> On Fri, Dec 3, 2021 at 1:19 AM Yoav Weiss <[email protected]> >>>>> wrote: >>>>> >>>>>> >>>>>> >>>>>> On Thu, Dec 2, 2021 at 8:14 PM Andrew Williams <[email protected]> >>>>>> wrote: >>>>>> >>>>>>> Thanks for the link, Daniel, and thank you to Mike as well for >>>>>>> sending me a Chrome Enterprise PoC to reach out to. >>>>>>> >>>>>>> Yoav, do you think waiting two or three months between when we >>>>>>> implement DevTools Issues / deprecation reports for this and when we do >>>>>>> the >>>>>>> removal would be sufficient? >>>>>>> >>>>>> >>>>>> That sounds reasonable to me. The known risk seems low, I'm mostly >>>>>> concerned about the unknown risk. I'm hoping that a couple of months of >>>>>> deprecation reports + enterprise policy can help us reduce risks on that >>>>>> front. >>>>>> >>>>>> >>>>>>> Is there a precedent for the minimum amount of time to provide >>>>>>> notice for changes like this that don't seem like they will have >>>>>>> widespread >>>>>>> impact but that we want to proceed with caution on nonetheless? >>>>>>> >>>>>>> -Andrew >>>>>>> >>>>>>> On Wed, Dec 1, 2021 at 4:02 PM Yoav Weiss <[email protected]> >>>>>>> wrote: >>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> On Wed, Dec 1, 2021 at 7:02 PM Andrew Williams < >>>>>>>> [email protected]> wrote: >>>>>>>> >>>>>>>>> Thanks for the LGTMs Yoav, Mike, and Daniel! >>>>>>>>> >>>>>>>>> Yoav, we haven't implemented warnings of any kind (DevTools Issues >>>>>>>>> or deprecation reports via the Reporting API) for this yet... >>>>>>>>> >>>>>>>> >>>>>>>> That part seems paramount before starting with the actual removal. >>>>>>>> What timelines do you have in mind? >>>>>>>> >>>>>>>> >>>>>>>>> This effort has been on the backburner but I aim to revisit it >>>>>>>>> soon. Also, I'm happy to reach out to Chrome Enterprise about this - >>>>>>>>> do you >>>>>>>>> have a recommendation regarding a good PoC? Also, our plan is still to >>>>>>>>> implement this behind a feature flag that could be disabled via Finch >>>>>>>>> in >>>>>>>>> the event that significant breakage is encountered (although I know >>>>>>>>> Finch >>>>>>>>> updates are not a panacea). >>>>>>>>> >>>>>>>>> -Andrew >>>>>>>>> >>>>>>>>> On Wed, Dec 1, 2021 at 7:26 AM Daniel Bratell <[email protected]> >>>>>>>>> wrote: >>>>>>>>> >>>>>>>>>> LGTM3 >>>>>>>>>> >>>>>>>>>> /Daniel >>>>>>>>>> On 2021-12-01 12:31, Mike West wrote: >>>>>>>>>> >>>>>>>>>> LGTM2, with less caution than Yoav. :) >>>>>>>>>> >>>>>>>>>> This seems safe to me, and philosophically correct for the >>>>>>>>>> platform. Good luck shipping it! >>>>>>>>>> >>>>>>>>>> -mike >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> On Wed, Dec 1, 2021 at 11:41 AM Yoav Weiss < >>>>>>>>>> [email protected]> wrote: >>>>>>>>>> >>>>>>>>>>> LGTM1 to ship given the low numbers, although we still want to >>>>>>>>>>> do this carefully. >>>>>>>>>>> >>>>>>>>>>> Are such cookies currently issuing deprecation reports? Have we >>>>>>>>>>> talked to Chrome Enterprise folks about their thoughts on adding an >>>>>>>>>>> enterprise policy? >>>>>>>>>>> >>>>>>>>>>> On Tuesday, November 30, 2021 at 2:17:27 PM UTC+1 Andrew >>>>>>>>>>> Williams wrote: >>>>>>>>>>> >>>>>>>>>>>> Regarding cookie strings that get truncated in Chrome today but >>>>>>>>>>>> that still get treated as valid: >>>>>>>>>>>> >>>>>>>>>>>> 0.0007 % of valid cookie strings were truncated by a \x0a >>>>>>>>>>>> character >>>>>>>>>>>> <0.0001 % of valid cookie strings were truncated by a \x0d >>>>>>>>>>>> character >>>>>>>>>>>> <0.0001 % of valid cookie strings were truncated by a \x00 >>>>>>>>>>>> character >>>>>>>>>>>> >>>>>>>>>>>> Thus, with the proposed change to reject cookie strings >>>>>>>>>>>> containing any of these characters, 0.0007% of cookie strings >>>>>>>>>>>> would be >>>>>>>>>>>> rejected. >>>>>>>>>>>> >>>>>>>>>>>> -Andrew >>>>>>>>>>>> >>>>>>>>>>>> On Wed, Nov 24, 2021 at 12:14 PM Andrew Williams < >>>>>>>>>>>> [email protected]> wrote: >>>>>>>>>>>> >>>>>>>>>>>>> Hi Chris, >>>>>>>>>>>>> >>>>>>>>>>>>> We've collected 7 days worth of metrics since M96 went live, >>>>>>>>>>>>> and the number of cookie strings containing a \x00, \x0d, or \x0a >>>>>>>>>>>>> character >>>>>>>>>>>>> is very small compared to the total number of cookie strings >>>>>>>>>>>>> processed. >>>>>>>>>>>>> I'll post the percentages here once I get approval to do so, >>>>>>>>>>>>> likely next >>>>>>>>>>>>> week given the holiday in the U.S.. >>>>>>>>>>>>> >>>>>>>>>>>>> -Andrew >>>>>>>>>>>>> >>>>>>>>>>>>> On Wed, Nov 24, 2021 at 11:29 AM Chris Harrelson < >>>>>>>>>>>>> [email protected]> wrote: >>>>>>>>>>>>> >>>>>>>>>>>>>> Hi Andrew, >>>>>>>>>>>>>> >>>>>>>>>>>>>> M96 has now shipped. Is the UseCounter data now available? >>>>>>>>>>>>>> >>>>>>>>>>>>>> On Thu, Nov 4, 2021 at 7:33 PM Andrew Williams < >>>>>>>>>>>>>> [email protected]> wrote: >>>>>>>>>>>>>> >>>>>>>>>>>>>>> Hi Daniel, >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> Thanks for following up on this. UMA metrics to count the >>>>>>>>>>>>>>> prevalence of \x00, \x0d, and \x0a characters in cookie strings >>>>>>>>>>>>>>> will roll >>>>>>>>>>>>>>> out with the M96 release. We'll post back here once those >>>>>>>>>>>>>>> metrics are >>>>>>>>>>>>>>> available. >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> Regarding deprecation warnings, we've mapped out how to >>>>>>>>>>>>>>> generate DevTools Issues that would warn developers when >>>>>>>>>>>>>>> cookies containing >>>>>>>>>>>>>>> these characters are detected, but we haven't implemented >>>>>>>>>>>>>>> anything yet. >>>>>>>>>>>>>>> Also, we haven't implemented the sending of deprecation reports >>>>>>>>>>>>>>> yet. Both >>>>>>>>>>>>>>> are still on our radar, though. >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> -Andrew >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> On Thu, Nov 4, 2021 at 2:37 PM Daniel Bratell < >>>>>>>>>>>>>>> [email protected]> wrote: >>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> The last thing happening in this thread was that we decided >>>>>>>>>>>>>>>> to wait for data. What is the current status of those >>>>>>>>>>>>>>>> usecounters, have >>>>>>>>>>>>>>>> they reached the user base now? >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> /Daniel >>>>>>>>>>>>>>>> On 2021-09-20 07:59, Yoav Weiss wrote: >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> On Fri, Sep 17, 2021 at 6:36 PM Steven Bingler < >>>>>>>>>>>>>>>> [email protected]> wrote: >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> Hi Ian and Yoav, >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> I believe the general guidance now for warning users of >>>>>>>>>>>>>>>>> some change is to use DevTools Issues rather than console >>>>>>>>>>>>>>>>> warnings. Would >>>>>>>>>>>>>>>>> using Issues, instead of console warnings, be acceptable to >>>>>>>>>>>>>>>>> you? (This >>>>>>>>>>>>>>>>> would be in addition to the reports.) >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> I don't believe the API OWNERS have a stand on console >>>>>>>>>>>>>>>> warnings vs. issues for deprecations. Whatever is the general >>>>>>>>>>>>>>>> guidance that >>>>>>>>>>>>>>>> will make this visible for developers seems good to me, >>>>>>>>>>>>>>>> assuming that >>>>>>>>>>>>>>>> issues are prominent in the UI and manage to grab the median >>>>>>>>>>>>>>>> developer's >>>>>>>>>>>>>>>> attention. >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> Also, for posterity, it is possible to emit a console >>>>>>>>>>>>>>>>> warning starting from EmitCookieWarningsAndMetrics() with a >>>>>>>>>>>>>>>>> little work. We >>>>>>>>>>>>>>>>> used to do just that for SameSite warnings before we >>>>>>>>>>>>>>>>> transitioned them to >>>>>>>>>>>>>>>>> DevTools Issue [1 >>>>>>>>>>>>>>>>> <https://chromium.googlesource.com/chromium/src.git/+/1cc31f46c2e6ae658a97b92fbc8c556eba382d3e/content/browser/frame_host/cookie_utils.cc#128>] >>>>>>>>>>>>>>>>> [2 >>>>>>>>>>>>>>>>> <https://chromium.googlesource.com/chromium/src.git/+/1cc31f46c2e6ae658a97b92fbc8c556eba382d3e/content/browser/frame_host/render_frame_host_impl.cc#8309>] >>>>>>>>>>>>>>>>> (many refactors ago). It looks like most of the necessary >>>>>>>>>>>>>>>>> functions still >>>>>>>>>>>>>>>>> exist, so it shouldn't be too hard to recreate that >>>>>>>>>>>>>>>>> functionality if >>>>>>>>>>>>>>>>> needed. >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> Thanks, >>>>>>>>>>>>>>>>> Steven >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>> On Friday, September 17, 2021 at 8:45:10 AM UTC-7 Andrew >>>>>>>>>>>>>>>>> Williams wrote: >>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>> Thanks for the feedback Mike, Yoav, and Ian. I will >>>>>>>>>>>>>>>>>> explore the feasibility of using CountDeprecation (or >>>>>>>>>>>>>>>>>> something similar) >>>>>>>>>>>>>>>>>> from the cookie-related code and will report back once I >>>>>>>>>>>>>>>>>> have an update on >>>>>>>>>>>>>>>>>> this. >>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>> -Andrew >>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>> On Fri, Sep 10, 2021 at 10:11 AM Ian Clelland < >>>>>>>>>>>>>>>>>> [email protected]> wrote: >>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>> That looks right -- that code path won't get you >>>>>>>>>>>>>>>>>>> anywhere near adding a console message, as far as I can >>>>>>>>>>>>>>>>>>> tell, but you would >>>>>>>>>>>>>>>>>>> be able to queue a report that way. Ideally, we'd have >>>>>>>>>>>>>>>>>>> something like >>>>>>>>>>>>>>>>>>> deprecation.cc for browser-side that would handle the UMA >>>>>>>>>>>>>>>>>>> as well as >>>>>>>>>>>>>>>>>>> formatting the report body consistently. As a first pass, >>>>>>>>>>>>>>>>>>> until we have >>>>>>>>>>>>>>>>>>> more that one browser-generated deprecation report, just >>>>>>>>>>>>>>>>>>> generating and >>>>>>>>>>>>>>>>>>> queuing it would work. >>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>> On Fri, Sep 10, 2021 at 6:42 AM Yoav Weiss < >>>>>>>>>>>>>>>>>>> [email protected]> wrote: >>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>> I may very well be wrong, but it seems like >>>>>>>>>>>>>>>>>>>> CookieUtils::EmitCookieWarningsAndMetrics >>>>>>>>>>>>>>>>>>>> <https://source.chromium.org/chromium/chromium/src/+/main:content/browser/renderer_host/cookie_utils.cc;l=97;drc=8afc9e45a7e96afda8f22ef044d1e7cdc5a6f75a;bpv=1;bpt=1> >>>>>>>>>>>>>>>>>>>> has >>>>>>>>>>>>>>>>>>>> the right plumbing to reach RenderFrameHost, and from it, >>>>>>>>>>>>>>>>>>>> get a >>>>>>>>>>>>>>>>>>>> ReportingSource >>>>>>>>>>>>>>>>>>>> <https://source.chromium.org/chromium/chromium/src/+/main:content/browser/renderer_host/render_frame_host_impl.cc;l=1730;drc=8afc9e45a7e96afda8f22ef044d1e7cdc5a6f75a?q=RenderFrameHost&ss=chromium%2Fchromium%2Fsrc> >>>>>>>>>>>>>>>>>>>> that >>>>>>>>>>>>>>>>>>>> can enable us to send >>>>>>>>>>>>>>>>>>>> <https://source.chromium.org/chromium/chromium/src/+/main:content/browser/renderer_host/render_frame_host_impl.cc;l=10676;drc=8afc9e45a7e96afda8f22ef044d1e7cdc5a6f75a;bpv=1;bpt=1?q=RenderFrameHost&ss=chromium%2Fchromium%2Fsrc> >>>>>>>>>>>>>>>>>>>> deprecation reports (even if through a different mechanism >>>>>>>>>>>>>>>>>>>> than >>>>>>>>>>>>>>>>>>>> CountDeprecation). >>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>> Ian - thoughts on the above? >>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>> On Thu, Sep 9, 2021 at 9:21 PM Mike West < >>>>>>>>>>>>>>>>>>>> [email protected]> wrote: >>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> I don't think `countDeprecation` is going to work >>>>>>>>>>>>>>>>>>>>> here, insofar as that's a Blink-layer concept, and the >>>>>>>>>>>>>>>>>>>>> network stack isn't >>>>>>>>>>>>>>>>>>>>> going to have an understanding of page views or use >>>>>>>>>>>>>>>>>>>>> counters or etc. If >>>>>>>>>>>>>>>>>>>>> we've wired things up such that deprecation reports can >>>>>>>>>>>>>>>>>>>>> be triggered from >>>>>>>>>>>>>>>>>>>>> the network stack, lovely, but I'm not sure that's the >>>>>>>>>>>>>>>>>>>>> case. >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> Another approach that might be reasonable to approach >>>>>>>>>>>>>>>>>>>>> might be to roll this out on a percentage-basis, starting >>>>>>>>>>>>>>>>>>>>> with a >>>>>>>>>>>>>>>>>>>>> substantial portion of beta, then rolling to stable iff >>>>>>>>>>>>>>>>>>>>> we're confident in >>>>>>>>>>>>>>>>>>>>> that experience? >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> This feels like both the right directional and >>>>>>>>>>>>>>>>>>>>> philosophical thing to do with cookies. I'd like to see >>>>>>>>>>>>>>>>>>>>> it ship, and a >>>>>>>>>>>>>>>>>>>>> staged rollout might well be a reasonable way of gaining >>>>>>>>>>>>>>>>>>>>> confidence in our >>>>>>>>>>>>>>>>>>>>> ability to do so? >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> -mike >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>> On Thu, Sep 9, 2021 at 1:03 PM Yoav Weiss < >>>>>>>>>>>>>>>>>>>>> [email protected]> wrote: >>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>> Sounds good! Can you please ping this thread once >>>>>>>>>>>>>>>>>>>>>> results start coming in? Thanks! :) >>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>> On Wednesday, September 8, 2021 at 3:59:36 AM UTC+2 >>>>>>>>>>>>>>>>>>>>>> Andrew Williams wrote: >>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>> Sounds good - we will add the CountDeprecation >>>>>>>>>>>>>>>>>>>>>>> metrics. Thanks for the suggestion, Yoav, and thank you >>>>>>>>>>>>>>>>>>>>>>> Ian for the >>>>>>>>>>>>>>>>>>>>>>> additional info. >>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>> -Andrew >>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>> On Fri, Sep 3, 2021 at 10:07 AM Ian Clelland < >>>>>>>>>>>>>>>>>>>>>>> [email protected]> wrote: >>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>> On Fri, Sep 3, 2021 at 4:55 AM Yoav Weiss < >>>>>>>>>>>>>>>>>>>>>>>> [email protected]> wrote: >>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>> Hey Andrew, >>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>> Given that the metrics are not a superset of what >>>>>>>>>>>>>>>>>>>>>>>>> you're trying to deprecate, could you please add >>>>>>>>>>>>>>>>>>>>>>>>> CountDeprecation >>>>>>>>>>>>>>>>>>>>>>>>> <https://source.chromium.org/chromium/chromium/src/+/main:third_party/blink/renderer/core/frame/deprecation.cc;drc=f6f22e82bcd0d50f390b23ee9688c58de5ae0bdc;bpv=1;bpt=1;l=702?q=deprecation&ss=chromium> >>>>>>>>>>>>>>>>>>>>>>>>> metrics of the case you are intending to deprecate? >>>>>>>>>>>>>>>>>>>>>>>>> That would ensure .e.g >>>>>>>>>>>>>>>>>>>>>>>>> deprecation reports are sent to folks that happen to >>>>>>>>>>>>>>>>>>>>>>>>> have such cookies. >>>>>>>>>>>>>>>>>>>>>>>>> Even though you haven't really asked, from my >>>>>>>>>>>>>>>>>>>>>>>>> perspective, it's also fine to add a console >>>>>>>>>>>>>>>>>>>>>>>>> deprecation message at this >>>>>>>>>>>>>>>>>>>>>>>>> point, in parallel to the metrics. >>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>> FYI, CountDeprecation will take care of adding that >>>>>>>>>>>>>>>>>>>>>>>> console message for you, as well as: >>>>>>>>>>>>>>>>>>>>>>>> - Generating a report object which can be seen >>>>>>>>>>>>>>>>>>>>>>>> with a ReportingObserver, >>>>>>>>>>>>>>>>>>>>>>>> - Sending that report to any configured endpoints >>>>>>>>>>>>>>>>>>>>>>>> for the document, and >>>>>>>>>>>>>>>>>>>>>>>> - Counting the usage for UMA, so that we can track >>>>>>>>>>>>>>>>>>>>>>>> the (hopefully) declining usage of the deprecated >>>>>>>>>>>>>>>>>>>>>>>> feature. >>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>> Ian >>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>> Cheers :) >>>>>>>>>>>>>>>>>>>>>>>>> Yoav >>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>> On Wednesday, September 1, 2021 at 5:05:44 PM >>>>>>>>>>>>>>>>>>>>>>>>> UTC+2 Andrew Williams wrote: >>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>> Here is the percentage for the metric mentioned >>>>>>>>>>>>>>>>>>>>>>>>>> in my last email: over a 7 day period, 0.00004% of >>>>>>>>>>>>>>>>>>>>>>>>>> cookies seen in the >>>>>>>>>>>>>>>>>>>>>>>>>> stable version of Chrome had truncated names and/or >>>>>>>>>>>>>>>>>>>>>>>>>> values. >>>>>>>>>>>>>>>>>>>>>>>>>> Ultimately our plan is to ship this feature >>>>>>>>>>>>>>>>>>>>>>>>>> behind a kill switch that we could flip if major >>>>>>>>>>>>>>>>>>>>>>>>>> issues are reported. With >>>>>>>>>>>>>>>>>>>>>>>>>> that in mind, and given the low number of truncated >>>>>>>>>>>>>>>>>>>>>>>>>> cookie names/values >>>>>>>>>>>>>>>>>>>>>>>>>> observed via our existing metrics, would it make >>>>>>>>>>>>>>>>>>>>>>>>>> sense to implement and >>>>>>>>>>>>>>>>>>>>>>>>>> collect the new metrics in parallel with rolling out >>>>>>>>>>>>>>>>>>>>>>>>>> the changes described >>>>>>>>>>>>>>>>>>>>>>>>>> in this I2P&S? Or do you think taking the more >>>>>>>>>>>>>>>>>>>>>>>>>> cautious approach and >>>>>>>>>>>>>>>>>>>>>>>>>> implementing/collecting the new metrics before >>>>>>>>>>>>>>>>>>>>>>>>>> landing this change is a >>>>>>>>>>>>>>>>>>>>>>>>>> better way forward (despite taking more time)? >>>>>>>>>>>>>>>>>>>>>>>>>> -Andrew >>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>> On Fri, Aug 27, 2021 at 1:45 PM Andrew Williams < >>>>>>>>>>>>>>>>>>>>>>>>>> [email protected]> wrote: >>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>> Thanks for the feedback/questions Yoav and >>>>>>>>>>>>>>>>>>>>>>>>>>> Daniel. >>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>> We have some metrics >>>>>>>>>>>>>>>>>>>>>>>>>>> <https://source.chromium.org/chromium/chromium/src/+/700dc7fe1578ab5e0e50a6304f2a1960005b8f8b:tools/metrics/histograms/metadata/cookie/histograms.xml;l=56;bpv=1;bpt=0> >>>>>>>>>>>>>>>>>>>>>>>>>>> on Chrome's existing behavior to truncate cookie >>>>>>>>>>>>>>>>>>>>>>>>>>> lines containing \x00, >>>>>>>>>>>>>>>>>>>>>>>>>>> \x0d, and \x0a (specifically, in cases where the >>>>>>>>>>>>>>>>>>>>>>>>>>> truncation affects the >>>>>>>>>>>>>>>>>>>>>>>>>>> cookie name or the cookie value). The percentage >>>>>>>>>>>>>>>>>>>>>>>>>>> of cookies with truncated >>>>>>>>>>>>>>>>>>>>>>>>>>> names or values is quite low, although I'm still >>>>>>>>>>>>>>>>>>>>>>>>>>> waiting on approval to >>>>>>>>>>>>>>>>>>>>>>>>>>> release the exact percentage. We don't have any >>>>>>>>>>>>>>>>>>>>>>>>>>> metrics for cases where >>>>>>>>>>>>>>>>>>>>>>>>>>> truncation affected cookie attribute parsing (for >>>>>>>>>>>>>>>>>>>>>>>>>>> example, the malicious >>>>>>>>>>>>>>>>>>>>>>>>>>> case this intent aims to address) or where >>>>>>>>>>>>>>>>>>>>>>>>>>> truncation was harmless (for >>>>>>>>>>>>>>>>>>>>>>>>>>> example, a newline as the last character in the >>>>>>>>>>>>>>>>>>>>>>>>>>> cookie line), though. >>>>>>>>>>>>>>>>>>>>>>>>>>> Especially for the latter case, it does seem >>>>>>>>>>>>>>>>>>>>>>>>>>> plausible that certain sites >>>>>>>>>>>>>>>>>>>>>>>>>>> could be constructing cookie lines in such a way >>>>>>>>>>>>>>>>>>>>>>>>>>> that control characters >>>>>>>>>>>>>>>>>>>>>>>>>>> slip in unnoticed. We will add new metrics to >>>>>>>>>>>>>>>>>>>>>>>>>>> cover these cases so that we >>>>>>>>>>>>>>>>>>>>>>>>>>> can better predict the level of breakage that these >>>>>>>>>>>>>>>>>>>>>>>>>>> changes may have. >>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>> -Andrew >>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>> On Thu, Aug 26, 2021 at 2:22 PM Daniel Bratell < >>>>>>>>>>>>>>>>>>>>>>>>>>> [email protected]> wrote: >>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>> Even if browsers are currently slightly >>>>>>>>>>>>>>>>>>>>>>>>>>>> incompatible, it seems this change will short term >>>>>>>>>>>>>>>>>>>>>>>>>>>> make them more >>>>>>>>>>>>>>>>>>>>>>>>>>>> incompatible. As Yoav said, it would be good to >>>>>>>>>>>>>>>>>>>>>>>>>>>> have an idea about how >>>>>>>>>>>>>>>>>>>>>>>>>>>> common this is, i.e. how often will cookies that >>>>>>>>>>>>>>>>>>>>>>>>>>>> are today truncated >>>>>>>>>>>>>>>>>>>>>>>>>>>> instead be rejected? >>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>> /Daniel >>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>> On 2021-08-25 16:18, Yoav Weiss wrote: >>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>> Hey Andrew! Thanks for working on this, this >>>>>>>>>>>>>>>>>>>>>>>>>>>> seems like a significant compatibility gap (with >>>>>>>>>>>>>>>>>>>>>>>>>>>> security implications) >>>>>>>>>>>>>>>>>>>>>>>>>>>> that would be great to close. >>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>> On Tuesday, August 24, 2021 at 3:45:50 PM UTC+2 >>>>>>>>>>>>>>>>>>>>>>>>>>>> Andrew Williams wrote: >>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> Contact emails [email protected], >>>>>>>>>>>>>>>>>>>>>>>>>>>>> [email protected] Explainer >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> https://github.com/httpwg/http-extensions/issues/1531 >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> https://github.com/httpwg/http-extensions/pull/1589 >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> Specification >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> https://github.com/httpwg/http-extensions/blob/main/draft-ietf-httpbis-rfc6265bis.md >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> Summary >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> Updates how control characters in cookie data >>>>>>>>>>>>>>>>>>>>>>>>>>>>> are handled. Specifically, the tab character is >>>>>>>>>>>>>>>>>>>>>>>>>>>>> now permitted, but all >>>>>>>>>>>>>>>>>>>>>>>>>>>>> other control characters cause the entire cookie >>>>>>>>>>>>>>>>>>>>>>>>>>>>> to be rejected (previously >>>>>>>>>>>>>>>>>>>>>>>>>>>>> the \x00, \x0D, and \x0A characters in a cookie >>>>>>>>>>>>>>>>>>>>>>>>>>>>> line caused it to be >>>>>>>>>>>>>>>>>>>>>>>>>>>>> truncated instead of rejected entirely, which >>>>>>>>>>>>>>>>>>>>>>>>>>>>> could have enabled malicious >>>>>>>>>>>>>>>>>>>>>>>>>>>>> behavior in certain circumstances). This behavior >>>>>>>>>>>>>>>>>>>>>>>>>>>>> is also in line with the >>>>>>>>>>>>>>>>>>>>>>>>>>>>> latest drafts of RFC6265bis. >>>>>>>>>>>>>>>>>>>>>>>>>>>>> Blink component >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> Internals>Network>Cookies >>>>>>>>>>>>>>>>>>>>>>>>>>>>> <https://bugs.chromium.org/p/chromium/issues/list?q=component:Internals%3ENetwork%3ECookies> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> Motivation >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> In the case where attacker controlled data is >>>>>>>>>>>>>>>>>>>>>>>>>>>>> used to set a new cookie, having certain control >>>>>>>>>>>>>>>>>>>>>>>>>>>>> characters truncate the >>>>>>>>>>>>>>>>>>>>>>>>>>>>> cookie line could result in security-related >>>>>>>>>>>>>>>>>>>>>>>>>>>>> cookie attributes being >>>>>>>>>>>>>>>>>>>>>>>>>>>>> ignored. This behavior may also lead to cookie >>>>>>>>>>>>>>>>>>>>>>>>>>>>> data corruption when >>>>>>>>>>>>>>>>>>>>>>>>>>>>> control characters are introduced, which may >>>>>>>>>>>>>>>>>>>>>>>>>>>>> cause unpredictable behavior >>>>>>>>>>>>>>>>>>>>>>>>>>>>> on the application side (more so than cookies not >>>>>>>>>>>>>>>>>>>>>>>>>>>>> being set, which is a >>>>>>>>>>>>>>>>>>>>>>>>>>>>> case that applications should already handle). >>>>>>>>>>>>>>>>>>>>>>>>>>>>> Having control characters >>>>>>>>>>>>>>>>>>>>>>>>>>>>> result in the whole cookie being rejected helps >>>>>>>>>>>>>>>>>>>>>>>>>>>>> mitigate these concerns and >>>>>>>>>>>>>>>>>>>>>>>>>>>>> aligns Chrome with RFC6265bis. For the tab >>>>>>>>>>>>>>>>>>>>>>>>>>>>> character, although it falls in >>>>>>>>>>>>>>>>>>>>>>>>>>>>> the control character range (\x00 - \x1F, \x7F), >>>>>>>>>>>>>>>>>>>>>>>>>>>>> it’s a printable character >>>>>>>>>>>>>>>>>>>>>>>>>>>>> and allowed by other browsers. Treating it the >>>>>>>>>>>>>>>>>>>>>>>>>>>>> same way that the space >>>>>>>>>>>>>>>>>>>>>>>>>>>>> character is treated makes sense intuitively, >>>>>>>>>>>>>>>>>>>>>>>>>>>>> eliminates a potential >>>>>>>>>>>>>>>>>>>>>>>>>>>>> fingerprinting vector, and aligns Chrome with >>>>>>>>>>>>>>>>>>>>>>>>>>>>> RFC6265bis. >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>> In the past, moving to a stricter models that >>>>>>>>>>>>>>>>>>>>>>>>>>>> forbade certain characters resulted in at least >>>>>>>>>>>>>>>>>>>>>>>>>>>> some breakage of >>>>>>>>>>>>>>>>>>>>>>>>>>>> non-malicious content. I doubt this one would be >>>>>>>>>>>>>>>>>>>>>>>>>>>> significantly different. >>>>>>>>>>>>>>>>>>>>>>>>>>>> Do you have a sense of the resulting breakage? >>>>>>>>>>>>>>>>>>>>>>>>>>>> If not, I think it'd make sense to add metrics to >>>>>>>>>>>>>>>>>>>>>>>>>>>> our cookie parsing >>>>>>>>>>>>>>>>>>>>>>>>>>>> algorithm and see what that breakage would look >>>>>>>>>>>>>>>>>>>>>>>>>>>> like. >>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> Initial public proposal TAG review >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> N/A >>>>>>>>>>>>>>>>>>>>>>>>>>>>> <https://groups.google.com/a/chromium.org/g/blink-api-owners-discuss/c/uBxq9uCpKx0/m/A5LI0NbyAAAJ>: >>>>>>>>>>>>>>>>>>>>>>>>>>>>> this change is already specified in RFC 6265bis >>>>>>>>>>>>>>>>>>>>>>>>>>>>> and is a relatively minor >>>>>>>>>>>>>>>>>>>>>>>>>>>>> change to what's already implemented in Chrome >>>>>>>>>>>>>>>>>>>>>>>>>>>>> (to improve spec compliance). >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>> I agree that this change is in lower layers >>>>>>>>>>>>>>>>>>>>>>>>>>>> than those the TAG usually deals with. >>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> TAG review status Not applicable >>>>>>>>>>>>>>>>>>>>>>>>>>>>> Risks >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> N/A >>>>>>>>>>>>>>>>>>>>>>>>>>>>> Interoperability and Compatibility >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> WebKit / Safari: >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> - All control characters except the tab >>>>>>>>>>>>>>>>>>>>>>>>>>>>> character cause the cookie to be rejected if >>>>>>>>>>>>>>>>>>>>>>>>>>>>> present in the name and cause >>>>>>>>>>>>>>>>>>>>>>>>>>>>> the rest of the cookie line to be truncated if >>>>>>>>>>>>>>>>>>>>>>>>>>>>> present in the value >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> Gecko / Firefox: >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> - 0x00 in the cookie value causes the rest of >>>>>>>>>>>>>>>>>>>>>>>>>>>>> the value to be truncated (but subsequent >>>>>>>>>>>>>>>>>>>>>>>>>>>>> attributes are preserved) >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> - 0x00 in the cookie name causes the rest of >>>>>>>>>>>>>>>>>>>>>>>>>>>>> the name and the value to be truncated (but >>>>>>>>>>>>>>>>>>>>>>>>>>>>> subsequent attributes are >>>>>>>>>>>>>>>>>>>>>>>>>>>>> preserved) >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> - 0x0d and 0x0a cause the entire cookie line >>>>>>>>>>>>>>>>>>>>>>>>>>>>> to be truncated (attributes ignored) >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> - 0x01 through 0x09 (the tab character), 0x0b >>>>>>>>>>>>>>>>>>>>>>>>>>>>> through 0x0c, and 0x0e through 0x1f cause the >>>>>>>>>>>>>>>>>>>>>>>>>>>>> cookie to be rejected if they >>>>>>>>>>>>>>>>>>>>>>>>>>>>> are present in the name, but are allowed in the >>>>>>>>>>>>>>>>>>>>>>>>>>>>> cookie value >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> - 0x7f is allowed in the cookie name and >>>>>>>>>>>>>>>>>>>>>>>>>>>>> cookie value >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> >>>>>>>>>>>>>>>>>>>>>>>>>>>>> -- You received this message because you are subscribed to the Google Groups "blink-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAL5BFfWfpj0HTMXPo9sYrVwywme4ZaK%2BkU%2BV2Ew62CdZfBbAbg%40mail.gmail.com.
