Re: Bug triage report.

760 views
Skip to first unread message

David Benjamin

unread,
Jul 20, 2015, 1:28:19 PM7/20/15
to Siddharth Vijayakrishnan, Bryan McQuade, Charles 'Buck' Krasic, Misha Efimov, Ryan Sleevi, Randy Smith, Helen Li, Bence Béky, net-dev
This is half-way through my shift, but there was some miscommunication with Wednesday/Thursday's, so Charles is taking the second half of mine (Thanks!). Over to you, Charles.

Unlabeled network bugs:

Triaged up to 511718

https://crbug.com/510016 - WebSockets bug
https://crbug.com/510512 - asked for net-internals
https://crbug.com/510680 - seems some captive portal or something poisoned the HTTP cache
https://crbug.com/510935 - another one of those "Windows can't parse the cert but we can" problems
https://crbug.com/511073 - Apple cert verifier crash. Another one.
https://crbug.com/511150 - ERR_CONNECTION_RESET
https://crbug.com/511158 - auth + extensions
https://crbug.com/511659 - some CrOS-specific net-internals thing
https://crbug.com/511718 - <a download> and ServiceWorker

Investigating Cr-Internals-Network bugs:

https://crbug.com/511209 - internet disconnects
https://crbug.com/509345 - router rebooting, maybe an allergic reaction to QUIC? It's not clear
https://crbug.com/501795 - something timing out; tentatively looks like HTTP/2, but might be SSL

Look for new crashes:


Investigating crashes:
Dealing with old bugs:
(Didn't get to these with the backlog.)

On Tue, Jul 14, 2015 at 12:55 PM 'Siddharth Vijayakrishnan' via net-dev <net...@chromium.org> wrote:
On 13 July 2015 at 08:55, Bryan McQuade <bmcq...@chromium.org> wrote:
Triaged up to 509066

Identified 2 new crashers that could use additional investigations;
crbug.com/509066: Crash in net::TCPSocketWin::Write
crbug.com/509046 Crash in net::CertVerifyProcWin::VerifyInternal
both of these signatures were typically associated with malware but it appears there is also a real net stack bug for each, unrelated to malware.

Notable other issues:
crbug.com/508231 27.0% regression in Net.HttpTimeToFirstByte-android-dev on 2015-06-23 - probably nothing, but worth watching for a few more days

crbug.com/289715 Chrome_Mac: Crash Report - Magic Signature: net::URLRequestHttpJob::SaveCookiesAndNotif... - recently unassigned by battre@. Should try to find a new owner.

​What's the priority on this ? It says P1, which to me means that we should try and fix this by the next release milestone which is M46 currently. 
 
crbug.com/504416: Chrome: Crash Report - net::URLRequestJob::RecordBytesRead - I think this had been misassigned to tbansal, so just unassigned.
​Looks like Paul has started looking at this. ​
 


On Thu, Jul 9, 2015 at 10:56 AM Misha Efimov <m...@chromium.org> wrote:
Triaged up to 508528, nothing particularly exciting.

Learned that gasper alerts correlated with significant change of volume should be ignored.

Investigated a couple of issues, for example http://crbug.com/502226 which had symptoms of network related, but turned out to be caused by javascript engine changes.



On Tue, Jul 7, 2015 at 7:22 AM, Ryan Sleevi <rsl...@chromium.org> wrote:
Triaged up through 507190.

On the networking side, nothing terribly interesting.

https://code.google.com/p/chromium/issues/detail?id=507184 - Requests timing out after 30 seconds (initial looks to be a server issues, going to keep poking it)
https://code.google.com/p/chromium/issues/detail?id=505268 - The experimental 'HTTPS: 1' header for Upgrade Insecure Requests ( https://w3c.github.io/webappsec/specs/upgrade/#preference ) looks to be breaking some Wordpress blogs ( https://github.com/woothemes/woocommerce/issues/8479 for sample snippet of 'broken' code ). New headers are hard :(

On Fri, Jul 3, 2015 at 6:16 AM, 'Siddharth Vijayakrishnan' via net-dev <net...@chromium.org> wrote:
Triaged up to 506848, didn't see anything interesting. Didn't see any activity on gasper alerts other than a simple cache related one on android-dev which I don't think is important. Since I am on vacation, didn't spent time on investigating old reports. 

On 1 July 2015 at 04:08, Randy Smith <rds...@chromium.org> wrote:
On duty 6/29&6/30.  Triaged to issue 505901.

This was a light triage duty; I spent most of it evaluating
net-internals dumps on Cr-Internals-Network bugs that hadn't gotten
any attention in a while and trying to figure out why they hadn't
gotten any attention in a while.  All of crash scanning, bug scanning,
and gasper alerts were pretty simple; crash scanning because very few
releases have come out recently, and the net crashes in them had
already been registered as issues, but scanning because not many bugs
had been filed and Raman triaged well well past noon PT on his last
day, and gasper alerts because very few alerts fired.

Nothing very interesting in the bugs I looked at; the most interesting
was http://crbug.com/501795#c14, which David helped me trace down to a
very delayed response from the OS to certificate validation.  Other
potential bits of interest:

* http://crbug.com/505707: Chrome ignores "Accept Ranges".  Ricardo
thinks this is due to a problem with the media player.

* http://crbug.com/115916: Old bug about Chrome not being able to
access the internet after suspend/resume; the two net-internals dumps
I looked at were about hanging in system interfaces.

-- Randy


On Wed, Jun 24, 2015 at 4:39 PM, Helen Li <xunj...@chromium.org> wrote:
> On duty 06/23 and 06/24. Triaged to 504088. No new gasper alerts. Filed bug
> for two crashes.
>
> New network related bugs:
>
> Issue 502980:Network requests don't Vary on Client-Hints Width
>
> Issue 503488 in chromium: ServiceWorker doesn't handle trusted self-signed
> certificates
>
> Issue 504024:Chrome Repeatedly Prompts for Windows Authentication
>
>
> On Fri, Jun 19, 2015 at 3:55 PM, Bence Béky <b...@chromium.org> wrote:
>>
>> Triaged up to https://crbug.com/501991.  Filed bugs for quite a few
>> crashes.  Did bisecting to find culprits for two bugs.  No gasper
>> alerts.  Nothing extraordinary to report.
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "net-dev" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to net-dev+u...@chromium.org.
>> To post to this group, send email to net...@chromium.org.
>> To view this discussion on the web visit
>> https://groups.google.com/a/chromium.org/d/msgid/net-dev/CACMu3trze2Qrq-rQVn9dJ9bLdc69au8h8hQV%2BrTXZnBGPpqqPQ%40mail.gmail.com.
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "net-dev" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to net-dev+u...@chromium.org.
> To post to this group, send email to net...@chromium.org.
> To view this discussion on the web visit
> https://groups.google.com/a/chromium.org/d/msgid/net-dev/CAEkFr0402xB5T3Sq%3DsK_TrsbV4dHqXot7xuEb%3DYS8_QeyoTKMg%40mail.gmail.com.

--
You received this message because you are subscribed to the Google Groups "net-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to net-dev+u...@chromium.org.
To post to this group, send email to net...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/net-dev/CAFbEG_oSGs-HSnUxXY-RTQe21pGD%3Dk%2Bvpj4v23tKZduPJovZMA%40mail.gmail.com.

--
You received this message because you are subscribed to the Google Groups "net-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to net-dev+u...@chromium.org.
To post to this group, send email to net...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/net-dev/CADBDApArMzktjvQNouRMQs-hTsVqNXkuiu-w8ZDy-LMcUA3OFw%40mail.gmail.com.

--
You received this message because you are subscribed to the Google Groups "net-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to net-dev+u...@chromium.org.
To post to this group, send email to net...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/net-dev/CACvaWvajO%2BVH8wVVc2wfG9RhBHQpLZx%2B-z11_%3D_YiwERnp21CA%40mail.gmail.com.

--
You received this message because you are subscribed to the Google Groups "net-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to net-dev+u...@chromium.org.
To post to this group, send email to net...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/net-dev/CANapZYKhXVb2PtNU2SN1jU_66nHD8Dnu4rpcaZsyqLXnJPh7Kg%40mail.gmail.com.

--
You received this message because you are subscribed to the Google Groups "net-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to net-dev+u...@chromium.org.
To post to this group, send email to net...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/net-dev/CADBDApDSR9Ybmxnbt8L8h1cjYvb3%2Byt4kSBN%2BoQX4EAMbFc7-g%40mail.gmail.com.

Charles 'Buck' Krasic

unread,
Jul 20, 2015, 8:11:30 PM7/20/15
to David Benjamin, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, Helen Li, Bence Béky, net-dev
[ My first triage shift, spent most time learning the process ]

Unlabeled network bugs: 

None since 511718, triaged up to 512126

Investigating Cr-Internals-Network bugs:

crbug.com/511484 Update user agent used for "Request Desktop Site"
  - AFAICT, not a network bug. 
 
Look for new crashes:


Investigating crashes:
Dealing with old bugs:


--
Charles 'Buck' Krasic | Software Engineer | ckr...@google.com | +1 (408) 412-1141

Asanka Herath

unread,
Jul 23, 2015, 12:07:11 PM7/23/15
to Charles 'Buck' Krasic, David Benjamin, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, Helen Li, Bence Béky, net-dev
Triaged up to 513079. No trends.

Some users are complaining that WikiMedia pages are being downloaded instead of being rendered. Asked for logs. (BUG).

One crash was flagged as a UAF in extensions (BUG).


Matt Menke

unread,
Jul 24, 2015, 5:54:53 PM7/24/15
to Asanka Herath, Charles 'Buck' Krasic, David Benjamin, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, Helen Li, Bence Béky, net-dev
Triaged up to 513867.  A lot of fun and exciting new bugs.

Spent some time closing / pinging old needs-feedback issues, though we still have a lot.  Always takes a lot longer than I expect for each issue, only close those where there seems to be no patch forward without feedback.

As has been mentioned before, the crash server often doesn't show associated bugs until you update the list.  I found that adding a comment and then saving it is enough to get it to do this, though need to remember to remove the comment sand save again.


Bugs:

Someone filed a duplicate of the WikiMedia download bug.  Asanka and David figured out it's an issue with an HTTP2/SPDY intolerant server.  We're also interpreting the SPDY frames we receive as an HTTP/0.9 response to our HTTP/1.x request.  Should we even be supporting HTTP/0.9 over TLS?.

More "HTTPS: 1" fallout.  ColdFusion uses CGI.HTTPS to indicate whether a page was requested over HTTPS or not.  In an innovative fashion, if there was an "HTTPS" request header, it takes on that value instead, saving a couple bytes of memory, and allowing a browser to indicate to the server that it likes exposing credentials over HTTP.

Another bug using HTTPS when we shouldn't be.  Almost certainly also caused by The Grand HTTPS:1 Adventure, though the site seems to be working now.

Redirect loop that looks to be a bug on Microsoft's site, when requesting it in German.  Didn't WontFix it, but suspect it's not worth digging into.

A pair of issues with Windows's cert dialog.  Davidben already has a fix for one out for review.

A pair of issues related to launching external protocol handlers - not exactly network issues, but the code dangles pretty directly off the RDH.

Four "new" crashers, two look like malware, one URLFetcher being used to fetch invalid URLs resulting in a CHECK failure, one where we try to write to a closed HTTP2 stream after reading from an UploadDataStream, resulting in a null pointer dereference.

Not a new bug, but spent some time investigating issue 509144.  When an extension tries to override a URL, the extension infrastructure doubles up headers, even if the extension just returns the original set of headers.  An "sdch, gzip" content-encoding is replaced with "sdch, gzip, sdch, gzip".  Bizarrely, we don't detect a problem when decompressing it - instead, in the example given, we extract a file that's normally ~38k (<5k with SDCH) into a 6+ gigabyte file, which I've filed as a new bug (Note that this wasn't a constructed attack file, just a normal resource).

request we cache XHRs POST responses.


David Benjamin

unread,
Jul 24, 2015, 6:07:46 PM7/24/15
to Matt Menke, Asanka Herath, Charles 'Buck' Krasic, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, Helen Li, Bence Béky, net-dev
On Fri, Jul 24, 2015 at 5:54 PM Matt Menke <mme...@chromium.org> wrote:
Someone filed a duplicate of the WikiMedia download bug.  Asanka and David figured out it's an issue with an HTTP2/SPDY intolerant server.  We're also interpreting the SPDY frames we receive as an HTTP/0.9 response to our HTTP/1.x request.  Should we even be supporting HTTP/0.9 over TLS?.

If it turns out HTTP/0.9's numbers are low enough that it's unused over TLS, I'd say go for it. Same as over port 80. But just on general principle, not this bug. This bug is not so much an intolerant server, but a completely broken SSL MITM proxy. :-)

The only way I can see for this to happen is if the SSL MITM echos ClientHello extensions it doesn't understand and then tries to modify the rest of the handshake. (Notably it echos ClientHello ALPN advertisements but strips the ServerHello ALPN selection.) This is totally unsound since there can be (and are!) extensions that change the structure of the handshake or details of the crypto.
 

Matt Menke

unread,
Jul 24, 2015, 6:12:57 PM7/24/15
to David Benjamin, Asanka Herath, Charles 'Buck' Krasic, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, Helen Li, Bence Béky, net-dev
On Fri, Jul 24, 2015 at 6:07 PM, David Benjamin <davi...@chromium.org> wrote:
On Fri, Jul 24, 2015 at 5:54 PM Matt Menke <mme...@chromium.org> wrote:
Someone filed a duplicate of the WikiMedia download bug.  Asanka and David figured out it's an issue with an HTTP2/SPDY intolerant server.  We're also interpreting the SPDY frames we receive as an HTTP/0.9 response to our HTTP/1.x request.  Should we even be supporting HTTP/0.9 over TLS?.

If it turns out HTTP/0.9's numbers are low enough that it's unused over TLS, I'd say go for it. Same as over port 80. But just on general principle, not this bug. This bug is not so much an intolerant server, but a completely broken SSL MITM proxy. :-)

Oops, yea, I left out "proxy", which rather changes meaning there.

I'll plan to add histograms for HTTP/0.9 over SSL, to see if we can remove it (HTTP/0.9 is used for 1 out of every 8,300 HTTP/[01] responses in general, according to our current metrics from M44 (It's still rolling out to stable) which is probably a bit too high to remove it entirely.  :(

Elly Jones

unread,
Jul 31, 2015, 10:48:21 AM7/31/15
to Matt Menke, David Benjamin, Asanka Herath, Charles 'Buck' Krasic, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, Helen Li, Bence Béky, net-dev
mattm reminded me that I didn't actually mail a triage report!

Did up through 514900.

Interesting/notable bugs:
https://crbug.com/514106 - SPDY protocol errors on many sites. I think this is likely to be a bad middlebox interfering with SPDY but we need a netlog to be sure.
https://crbug.com/514518 - Cache-Control directives not being honored, needs a netlog to investigate

A couple of memory leaks:
https://crbug.com/514439 - system request context
I don't really understand the practice of immediately suppressing these warnings when they appear.

Not a lot else that is interesting.

-- elly

Matt Mueller

unread,
Jul 31, 2015, 6:37:06 PM7/31/15
to Elly Jones, Matt Menke, David Benjamin, Asanka Herath, Charles 'Buck' Krasic, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, Helen Li, Bence Béky, net-dev
Triaged up to 516039.

Not too much of interest:
https://crbug.com/515772 - another report of the websockets null origin bug
https://crbug.com/515943 - crash in content::PepperUDPSocketMessageFilter::SendBindReply (possibly chromoting related)



Matt Menke

unread,
Aug 4, 2015, 1:53:01 PM8/4/15
to Matt Mueller, Elly Jones, David Benjamin, Asanka Herath, Charles 'Buck' Krasic, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, Helen Li, Bence Béky, net-dev
Just thought I'd send out a warning to triagers:  I recently landed a CL to tighten up our parsing of HTTP responses.  We no longer treat responses that are just partial headers, missing the double CRLF/LF on the end as valid responses.  We removed this behavior for HTTPS a while back, for security reasons, I just removed it for HTTP as well.

It should be pretty low impact, but if you see bugs that mention a Chrome "regression" causing them to get "ERR_RESPONSE_HEADERS_TRUNCATED", they're probably due to my change.  Feel free to send those bugs my way, or just add the Internals-Network-HTTP label.

Also plan to remove support for extra meaningless bits of data before headers fairly soon - in the case of buggy servers, this may result in the entire page, including headers, being displayed.

Eric Roman

unread,
Aug 4, 2015, 9:48:35 PM8/4/15
to Matt Menke, Matt Mueller, Elly Jones, David Benjamin, Asanka Herath, Charles 'Buck' Krasic, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, Helen Li, Bence Béky, net-dev
Triaged up to 4 August, 17:56 PDT
(bug number 516887 if you must know)

Nothing interesting. And yet I will go into nauseating detail...

Bug activity

Asked for more information. Seems like a WontFix from what I could understand of it.

[crbug.com/160401] "CHECK failed in content::ResourceDispatcher::OnSetDataBuffer"
Old bug whose root cause was never identified (invalid handle when mapping memory in renderer), continues to impact users. Looked at recent crashes, no new data.

[crbug.com/516846] "Chrome back button doesn't respect the cache-control:no-cache header"
WontFix - The back button by design does not respect caching. (Confusion over Back/Forward, and HTTP cache headers in general is a frequent source of bug reports)

[crbug.com/516763] "Basic Authentication on native mobile web app hangs Chrome"
Getting more information. Although potentially troubling that they are using basic auth unless this is being done over an encrypted connection...

[crbug.com/516808] "Un-signed certificate not being accepted. Returns err_insecure_response."
(Translation: Un-Signed = Self-Signed). Reaching out to SSL experts on this one. The SSL connection fails flakily with ERR_CERT_AUTHORITY_INVALID given the _same_ certificate.

[crbug.com/516379] "Chrome_Android: Crash Report - chrome_browser_net::NetErrorTabHelper::DidFailProvisionalLoad"
Android specific problem, looks to be deep in RenderFrameHost / Navigation. Someone else is investigating...

[crbug.com/516787] "Typing "servername.domain.local" into the addressbar triggers a search and not a dns lookup"
Magic moments with omnibox search vs navigation prediction. Punted over to omnibox team for triage.

[crbug.com/516220] "ARC runtime: sock.setKeepAlive throws"
Pretty clueless about this, but it was tagged with Cr-Internals-Network. I expect will not have anything to do with our network stack code, punting to ARC team.

[crbug.com/516237] "Chrom{e,ium} refuses to talk HTTP/2 to a server once it's seen HTTP_1_1_REQUIRED for any resource."
Punted over to an HTTP/2 guru for further triage.

[crbug.com/516305] "Chrome not respecting ipv6 prefix precedence on linux."
Probably an address sorting issue with use of async resolver and its enforcement (or lack thereof) of precedence rules in /etc/gai.conf

Crash activity (dev and canary channels)

There were several new Chrome releases that overlapped my triage days, including a new stable release.

Problems spiked on the canary in 46.0.2470.0 and 46.0.2471.0 due to a single big crasher in src/net/*, but then dropped back down in the 46.0.2472.* releases.

The filter query I used was:

product.name CONTAINS 'Chrome' AND product.version CONTAINS '46.0.247' AND custom_data.ChromeCrashProto.ptype='browser' AND custom_data.ChromeCrashProto.magic_signature_1.component CONTAINS 'src/net'

The crash which dominated and accounted for 99% of the crashes was crbug.com/515846, and has now been fixed (confirmed fixed on 2472 builds).

The second top crash is an already known one (crbug.com/289715).

Crash activity (stable channel)

44.0.2403.130 was released a few hours ago. There weren't any real new issues, although I did file a bug for one signature which wasn't in the bug system (although admittedly it was seen only by a single user).

The filter I used was:

product.name CONTAINS 'Chrome' AND product.version='44.0.2403.130' AND custom_data.ChromeCrashProto.ptype='browser' AND custom_data.ChromeCrashProto.magic_signature_1.component CONTAINS 'src/net'

The top crashes were third party problems on Windows (i.e. crbug.com/501986), and also our good friend "[Assert] net::URLRequestContext::AssertNoURLRequests"

Gasper alerts

Haven't reviewed them yet, will respond later tonight...
Nothing jumped out at me though from my first pass.


Nick Harper

unread,
Aug 5, 2015, 6:37:41 PM8/5/15
to Eric Roman, Matt Menke, Matt Mueller, Elly Jones, David Benjamin, Asanka Herath, Charles 'Buck' Krasic, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, Helen Li, Bence Béky, net-dev
Triaged through 517221.
No new gasper alerts; no new crashers.

Josh Karlin

unread,
Aug 7, 2015, 3:08:38 PM8/7/15
to Nick Harper, Eric Roman, Matt Menke, Matt Mueller, Elly Jones, David Benjamin, Asanka Herath, Charles 'Buck' Krasic, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, Helen Li, Bence Béky, net-dev
Triaged through 517952.

Nothing interesting to report. 

Bence Béky

unread,
Aug 12, 2015, 7:55:48 AM8/12/15
to Josh Karlin, Nick Harper, Eric Roman, Matt Menke, Matt Mueller, Elly Jones, David Benjamin, Asanka Herath, Charles 'Buck' Krasic, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, Helen Li, net-dev
Triaged up to https://crbug.com/519674.  Nothing extraordinary.  Filed https://crbug.com/518812 and https://crbug.com/518813 for Gasper alerts.

Ricardo Vargas

unread,
Aug 21, 2015, 9:56:24 PM8/21/15
to Bence Béky, Josh Karlin, Nick Harper, Eric Roman, Matt Menke, Matt Mueller, Elly Jones, David Benjamin, Asanka Herath, Charles 'Buck' Krasic, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, Helen Li, net-dev
Triaged up to 523527. No particular trend or spike.

Helen Li

unread,
Aug 25, 2015, 5:31:38 PM8/25/15
to Ricardo Vargas, Bence Béky, Josh Karlin, Nick Harper, Eric Roman, Matt Menke, Matt Mueller, Elly Jones, David Benjamin, Asanka Herath, Charles 'Buck' Krasic, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, net-dev
Triage up to 524667. No new crashers. Nothing very interesting to report.

Ryan Sleevi

unread,
Aug 31, 2015, 9:19:42 PM8/31/15
to Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, Nick Harper, Eric Roman, Matt Menke, Matt Mueller, Elly Jones, David Benjamin, Asanka Herath, Charles 'Buck' Krasic, Siddharth Vijayakrishnan, Bryan McQuade, Misha Efimov, Ryan Sleevi, Randy Smith, net-dev
On Tue, Aug 25, 2015 at 2:23 PM, Helen Li <xunj...@chromium.org> wrote:
Triage up to 524667. No new crashers. Nothing very interesting to report.

Took over for ttutle@, got through 526817.

On the bug side:
  - https://crbug.com/510356 - Trying to get more information, but the user is using a proxy to speak to localhost, and running into issues.
  - https://crbug.com/525945 - Here, the inverse is desired. Code for making Chrome on Android/WebView skip talking to a proxy for localhost (and all intranet domains) is causing issue. This appears like it may be a dupe of https://crbug.com/471400, and just requires "Someone" being a "Decider" and deciding how the product should work for two somewhat incompatible use cases.

Nothing interesting in crashes.

Siddharth Vijayakrishnan

unread,
Sep 3, 2015, 2:13:23 AM9/3/15
to Ryan Sleevi, Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, Nick Harper, Eric Roman, Matt Menke, Matt Mueller, Elly Jones, David Benjamin, Asanka Herath, Charles 'Buck' Krasic, Bryan McQuade, Misha Efimov, Randy Smith, net-dev
Triaged till 527560. 
Cc'ed you on a couple of bugs to investigate further. Also, one crasher : 527534

Randy Smith

unread,
Sep 9, 2015, 12:31:54 PM9/9/15
to Raman Tenneti, Ryan Sleevi, Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, Nick Harper, Eric Roman, Matt Menke, Matt Mueller, Elly Jones, David Benjamin, Asanka Herath, Charles 'Buck' Krasic, Bryan McQuade, Misha Efimov, net-dev
Triaged 528022 (best guess as 3pm Friday) - 529341. Evaluated crash
reports, gasper alerts, and last week of touched Cr-Internals-Network
bugs. Short stint as my rotation overlapped with a holiday.

Didn't manage to investigate either http://crbug.com/528665 (trying to
download a everything from an extension results in a "corrupted
extension"), or http://crbug.com/499663 (low frequency crasher in
HttpStreamParser::DoReadBody). I don't think either is a very high
priority.

Concerns:
* The gasper alert system is broken--it would be useful to put some
netstack resources into fixing it (figuring out and moving to new
system, nailing down the nature of the comparisons). I evaluated
O(15) alerts, none of which were "real" to my eye.

* http://crbug.com/289715 (somewhat common crasher in
URLRequestHttpJob::SaveCookiesAndNotifyHeadersComplete) continues to
need love and not get it. Someone randomly assigned it to Matt, but
don't be fooled by that--no one has yet accepted the challenge.

* I'm a bit concerned by http://crbug.com/259940 (Can not open IPv6
sites using Toredo), mostly because of the sob story on the latest
post on the bug:

"PS I want you to understand that we're in cold-cold soviet russia
live in 1984 and have an <agency> watching us and blocking internet
sites. In that case teredo tunneling is the simplest way to non-pro
users to avoid such restrictions."

Raman Tenneti

unread,
Sep 11, 2015, 3:56:39 PM9/11/15
to Randy Smith, Ryan Sleevi, Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, Nick Harper, Eric Roman, Matt Menke, Matt Mueller, Elly Jones, David Benjamin, Asanka Herath, Charles 'Buck' Krasic, Bryan McQuade, Misha Efimov, net-dev
Triaged up to 530658.

+ Issue 530290 25.2% regression in SimpleCache.Http.EntryCreationTime-cros-dev on 2015-08-22
+ Issue 530243 7.3% regression in Net.HttpJob.TotalTime-win-stable on 2015-09-02
+ Issue 530482 --enable-ipv6 switch is broken on Chromium 45 - we deleted the switch. Being investigated.
+ Issue 528999  Network stack documentation is not up to date

thanks
raman

Charles 'Buck' Krasic

unread,
Sep 14, 2015, 8:23:49 PM9/14/15
to Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, Nick Harper, Eric Roman, Matt Menke, Matt Mueller, Elly Jones, David Benjamin, Asanka Herath, Bryan McQuade, Misha Efimov, net-dev
Triaged up to 531681.

531639: 11.0% regression in Net.HttpJob.TotalTimeCached-android-dev on 2015-08-30

 530674: Chrome: Crash Report - base::internal::BindState<base::internal::RunnableAdapter<void (*)(net::DnsConfig const &)>,void ,base::internal::TypeList<> >::Destroy
 
530662Chrome: Crash Report - net::QuicChromiumClientSession::OnConnectionClosed

531219QUIC persistent/repeated failure accessing groups.google.com while on (presumably broken) conference wifi



cshar...@google.com

unread,
Sep 17, 2015, 11:35:30 AM9/17/15
to net-dev, rten...@google.com, rds...@chromium.org, rsl...@chromium.org, xunj...@chromium.org, rva...@chromium.org, b...@chromium.org, jka...@chromium.org, nha...@chromium.org, ero...@chromium.org, mme...@chromium.org, ma...@chromium.org, elly...@google.com, davi...@chromium.org, asa...@chromium.org, bmcq...@chromium.org, m...@chromium.org, ckr...@google.com
Triaged up to 532675.

First triage duty. Could not successfully triage crbug.com/530200.

crbug.com/289715 and crbug.com/259940 from randy's triage still need help. There was a followup to 259940, 532271 that needs feedback.

Misha Efimov

unread,
Sep 18, 2015, 3:22:37 PM9/18/15
to Charles Harrison, net-dev, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, nha...@chromium.org, Eric Roman, Matt Menke, Matt Mueller, Elly Fong-Jones, David Benjamin, Asanka Herath, bmcq...@chromium.org, ckr...@google.com
Triaged up to 533586.

Notable issues:

533545 - Crash in HttpCache from ServiceWorkerDiskCacheMigrator
533012 - Chrome OS Login failure with School PAC script
532952 - Cannot download .application files

Finishing early to pack my stuff.

David Benjamin

unread,
Sep 22, 2015, 6:57:03 PM9/22/15
to Misha Efimov, Charles Harrison, net-dev, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, nha...@chromium.org, Eric Roman, Matt Menke, Matt Mueller, Elly Fong-Jones, Asanka Herath, bmcq...@chromium.org, ckr...@google.com
Nothing terribly interesting of note.

Unlabeled network bugs:

Triaged up to 535008

https://code.google.com/p/chromium/issues/detail?id=534099 - looks like WebComponents, but related to resource loading
https://code.google.com/p/chromium/issues/detail?id=534749 - asked for net-internals, seems to be a 3rd-party cookie thing?
https://code.google.com/p/chromium/issues/detail?id=534766 - Estonia continues to be incapable of generating a certificate
https://code.google.com/p/chromium/issues/detail?id=534970 - CPU usage in a dialog, though I vaguely suspect it's the throbber

Investigating Cr-Internals-Network bugs:

https://code.google.com/p/chromium/issues/detail?id=524185 - Assigned PowerSaveResourceThrottle bug to author
https://code.google.com/p/chromium/issues/detail?id=534250 - Turned out to be a duplicate of the Comodo crash

Look for new crashes:

Asanka Herath

unread,
Sep 25, 2015, 11:25:30 AM9/25/15
to David Benjamin, Misha Efimov, Charles Harrison, net-dev, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, nha...@chromium.org, Eric Roman, Matt Menke, Matt Mueller, Elly Fong-Jones, bmcq...@chromium.org, ckr...@google.com
Triaged up to 535755

Things of note:
  • From M45, Safe Browsing handles a lot more downloaded file types. Expect an increase in reports from site owners. Add Cr-UI-Browser-SafeBrowsing label if you see those.
  • Saw some dupes of the infamous %%30%30 bug. The fix for the stable channel is being rolled out in 45.0.2454.101. Canonical bug is 533361. I added a reference to the canonical bug and WontFixed the dupes I found due to the email volume for the canonical bug.
Bug activity (partial):
Nothing too exciting.
  • [ 535629 ] : Cannot receive >16k of headers using HTTP/2.
  • [ 535192 ] : Vague bug about downloads not working. Asked for more info.
  • [ 535089 ] : Site owner suspects that new their new code signing certificate is causing their binaries to be blocked by Chrome.
  • [ 534701 ] : User angry that Chrome doesn't ship a 64 bit Windows build. Gave them the good news.
  • [ 533149 ] : [webview] getting error ERR_NAME_NOT_RESOLVED when clicking links in Pinterest app. Looks like the links are broken.
  • [ 533625 ] : Infinite login loop with CAS authentication. User responded with massive net-internals log. Need to dig into this one.
Crashes:
  • [ 535601 ] : URLFetcher OOM. A URLFetcher client is downloading something large. URLFetcherStringWriter hits OOM when trying to expand the buffer to ~700MB. The dump unfortunately didn't point the finger at the culprit.
  • Filed reports for a couple of crashes I noticed that didn't have issues against them.
Other notes:

If you are looking at crash dumps on Windows, download WinDBG from the Windows 10 SDK. This build supports Natvis visualizers. So you can use Chromium's visualizers with it. Your eyeballs will thank you.

Also I didn't look at Gasper alerts yet. Will do so this afternoon.

Bryan McQuade

unread,
Sep 28, 2015, 7:30:41 PM9/28/15
to Asanka Herath, David Benjamin, Misha Efimov, Charles Harrison, net-dev, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, nha...@chromium.org, Eric Roman, Matt Menke, Matt Mueller, Elly Fong-Jones, ckr...@google.com
Triaged up to  536989.

New crasher:
https://code.google.com/p/chromium/issues/detail?id=536771 - mmenke indicates this may have already been fixed in most recent canary.

Crashers in progress (may need followup from next triager):
https://code.google.com/p/chromium/issues/detail?id=472280 (might be malware, but good to confirm)

Bugs in progress:
https://code.google.com/p/chromium/issues/detail?id=533625: need to do additional analysis of net internals dump

Steven Valdez

unread,
Sep 30, 2015, 5:50:59 PM9/30/15
to Bryan McQuade, Asanka Herath, David Benjamin, Misha Efimov, Charles Harrison, net-dev, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, nha...@chromium.org, Eric Roman, Matt Menke, Matt Mueller, Elly Fong-Jones, ckr...@google.com
Triaged up to 537798. Nothing too exciting. Filed a bunch (~40) of reports against old net:: crashes that didn't have a bug against them and moved ~20 Cr-Internals-Network bugs to appropriate subteams.

Matt Menke

unread,
Oct 2, 2015, 5:28:54 PM10/2/15
to Steven Valdez, Bryan McQuade, Asanka Herath, David Benjamin, Misha Efimov, Charles Harrison, net-dev, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, nha...@chromium.org, Eric Roman, Matt Mueller, Elly Fong-Jones, Charles 'Buck' Krasic
Triaged up to 538764.  Remarkably few new and exciting bugs, though a big backlog of old ones.  Only interesting new one marking proxies as bad, and picked up an old bug about a bunch of image requests hanging.

Went through a lot of the old network Needs-Feedback bugs, closing some, pinging others.  A bunch of those bugs had actually ended up in our court, and we failed to follow up on them. Also dealt with some of the aftermath of Anthony's pinging stale bugs.

Matt Mueller

unread,
Oct 13, 2015, 3:48:15 PM10/13/15
to Matt Menke, Steven Valdez, Bryan McQuade, Asanka Herath, David Benjamin, Misha Efimov, Charles Harrison, net-dev, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, nha...@chromium.org, Eric Roman, Elly Fong-Jones, Charles 'Buck' Krasic
Triaged up to 542714.  Not much exciting.  Someone spamming bugs with links to their blog.

Followed up on a few older needsfeedback bugs.

Not much as far as new crashes. crbug.com/542424 is an existing crash (was in low numbers) that increased around 45.0.2454.101 for no obvious reason.



--
You received this message because you are subscribed to the Google Groups "net-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to net-dev+u...@chromium.org.
To post to this group, send email to net...@chromium.org.

Eric Roman

unread,
Oct 15, 2015, 2:50:48 PM10/15/15
to Matt Mueller, Matt Menke, Steven Valdez, Bryan McQuade, Asanka Herath, David Benjamin, Misha Efimov, Charles Harrison, net-dev, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, nha...@chromium.org, Elly Fong-Jones, Charles 'Buck' Krasic
Triaged up to 543748 (Oct 15, 11:37 PDT)
  • M46 was released to stable 2 days ago. Many of the most interesting (unclassified) bugs were regressions from M45 --> M46, such as 543320. Not much in the way of network stuff.

  • Reviewed the src/net crashes so far on stable channel. Didn't find any new types of crash, but did review some of the older ones.

  • Here is a class of crash that needs an owner to clean up the the problems in the implementation issue 532798 identified (I didn't have time to take a stab at this, also unclear who the owner is). Also some more mundane crashes like issue 543751 that are easy to fix.

  • There was an entirely uninteresting gasper alert on "SimpleCache.Http.Eviction.TimeToSelectEntries-android-beta"

    This alert in particular didn't trouble me, however looking further back at its history there is a very big latency jump (almost 3x) between 45 --> 46, that I don't think anyone has identified. Find the discussion at https://goto.google.com/dykit (internal), I believe this needs some more follow-up.

Chris Bentzel

unread,
Oct 15, 2015, 9:45:09 PM10/15/15
to Eric Roman, Matt Mueller, Matt Menke, Steven Valdez, Bryan McQuade, Asanka Herath, David Benjamin, Misha Efimov, Charles Harrison, net-dev, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, nha...@chromium.org, Elly Fong-Jones, Charles 'Buck' Krasic
Was a bug filed about the SimpleCache... latency jump?

Eric Roman

unread,
Oct 15, 2015, 10:19:04 PM10/15/15
to Chris Bentzel, Matt Mueller, Matt Menke, Steven Valdez, Bryan McQuade, Asanka Herath, David Benjamin, Misha Efimov, Charles Harrison, net-dev, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Ricardo Vargas, Bence Béky, Josh Karlin, nha...@chromium.org, Elly Fong-Jones, Charles 'Buck' Krasic

Ryan Hamilton

unread,
Oct 18, 2015, 6:02:26 PM10/18/15
to Eric Roman, net-dev
Triaged up to 544658 Oct 18, 09:00 PDT
  • Very low frequency crasher in net::TCPSocketWin::LogConnectEnd. issue 544241
  • Look like there is some sort of bug with QUIC closing UDP sockets that is not yet sorted out.
  • Some noisy gasper alerts which seem to be the result of a new Android dev release which wasn't detected correctly.

Bence Béky

unread,
Oct 21, 2015, 8:14:21 AM10/21/15
to Ryan Hamilton, Eric Roman, net-dev
Ryan, thank you for covering the weekend. My understanding is that is
would have been my responsibility.

Triaged up to 545606. Nothing exciting to report. Some crashers with
low numbers filed. Note that Chirp is working and Gaspar should
eventually be deprecated, I'll update the triage documents.

Nick Harper

unread,
Oct 22, 2015, 3:11:17 PM10/22/15
to net-dev
Triaged up to 546686. I haven't looked at crash reports from stable release 46.0.2490.80, because it hasn't been out long enough yet.


--
You received this message because you are subscribed to the Google Groups "net-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to net-dev+u...@chromium.org.
To post to this group, send email to net...@chromium.org.

Matt Menke

unread,
Oct 26, 2015, 5:21:27 PM10/26/15
to Nick Harper, net-dev
Triaged up to 547939.

A large number of issues with Save-As dialogs hanging on windows.  May be some new malware package that's all the latest rage, or could just be some brokenware.

Looks like we have an Android cache performance regression on M46.

A couple new crashers with high enough incidence to file bugs for, but nothing too frequent.  Pinged or closed some old idle bug reports from the last couple months.

Josh Karlin

unread,
Oct 28, 2015, 2:42:22 PM10/28/15
to Matt Menke, Nick Harper, net-dev
Triaged up to 548747. No new crashes or bugs of note. 

Chris Bentzel

unread,
Oct 28, 2015, 2:43:47 PM10/28/15
to Matt Menke, Nick Harper, net-dev
Missed this one - is anyone tackling the android cache performance regression?

Matt Menke

unread,
Oct 28, 2015, 2:46:54 PM10/28/15
to Josh Karlin, Nick Harper, net-dev
Josh:  Did you look at any of the finch chirps from the last two days?

Matt Menke

unread,
Oct 28, 2015, 2:57:37 PM10/28/15
to Chris Bentzel, Nick Harper, net-dev
Hasn't been any activity on the bug - http://crbug.com/547811

Josh Karlin

unread,
Oct 28, 2015, 3:11:38 PM10/28/15
to Matt Menke, Nick Harper, net-dev
I missed your mention of a performance regression. Is there a bug filed that I can follow-up on? I only looked at the last couple of days chirp reports and there weren't any. Happy to dig in further tomorrow.

Helen Li

unread,
Nov 3, 2015, 5:24:34 PM11/3/15
to Josh Karlin, Matt Menke, Nick Harper, net-dev
Triaged up to 551079. Filed a couple of low-frequency crashers. Looked through chirp reports. Most of them correspond to version releases and are within historical values. Android cache performance regression on M46 is still present (crbug.com/543913).

New bugs:

Range header isn't sent correctly.

This is reported as a regression. Reporter has permission denied errors when trying to use bisect-builds. Next triager probably should follow up on this, since this looks like a regression in M46.

Range header caching behavior isn't right. Probably not related to 550914, because this one is in M45.

Feature request:

Old feature request has been around for 5 years
Issue 70762: Can't specify ScopeID when using link-local IPv6 address.



Randy Smith

unread,
Nov 5, 2015, 8:13:16 PM11/5/15
to Helen Li, Josh Karlin, Matt Menke, Nick Harper, net-dev
Triaged up to 552050. No high level comments; crashers uninteresting
except for process (see other thread).

-- Orphaned bugs that IMO someone should look at:

* http://crbug.com/121085: There's an assertion that was put in by
Wan-Teh many years ago that's still firing, so I unarchived this bug.
It would probably be worthwhile for someone to come up to speed on the
code and figure out what the assertion is tripping.

-- Bugs that are still in triage that the next triager should look at:

* http://crbug.com/550914: Bad range requests from audio. Still in
active evaluation; the next triager should look at it when the poster
replies with a net-internals dump. The link the poster gave to me to
investigate it with was http://www.peterkeuter.nl/test, but I couldn't
reproduce with that link.

* http://crbug.com/533625: This is a redirect loop in CAS
authentication that got dropped on the floor. I didn't get a chance
to look at it, but the next triager probably should.

-- Bugs that might be of interest for other reasons:

* http://crbug.com/552092: Crash in CertVerifyProcWin::VerifyInternal.
Other bugs in cert verification have been assigned to malware, but I
couldn't convince myself that this was malware from glancing at the
third party modules, so requesting someone in SSL to take a look at
it.

* http://crbug.com/106283: A problem decoding SVGZ when delivered over
HTTP or data schemes. Someone from Opera pointed out that not
decoding was spec compliant, and there's some discussion as to whether
Chrome should support this or not; about all that's left on our end is
figuring out why it works over file: URLs, and I don't think that's
high priority :-}.

-- Randy
> https://groups.google.com/a/chromium.org/d/msgid/net-dev/CAEkFr05t-y5OpsytdNNFm_r%2B%2BmB%3DxL9oWrRM2kmoHgbeBy9AtA%40mail.gmail.com.

Matt Menke

unread,
Nov 5, 2015, 8:26:25 PM11/5/15
to Randy Smith, Helen Li, Josh Karlin, Nick Harper, net-dev
On Thu, Nov 5, 2015 at 8:13 PM, Randy Smith <rds...@chromium.org> wrote:
Triaged up to 552050.  No high level comments; crashers uninteresting
except for process (see other thread).

-- Orphaned bugs that IMO someone should look at:

* http://crbug.com/121085: There's an assertion that was put in by
Wan-Teh many years ago that's still firing, so I unarchived this bug.
It would probably be worthwhile for someone to come up to speed on the
code and figure out what the assertion is tripping.

Isn't this one due to malware / brokenware mucking with our sockets?  This one has been around for ages, and I always ignore it, because of the discussion on that bug.

Ryan Sleevi

unread,
Nov 5, 2015, 8:37:17 PM11/5/15
to Randy Smith, Helen Li, Josh Karlin, Matt Menke, Nick Harper, net-dev
On Thu, Nov 5, 2015 at 5:13 PM, Randy Smith <rds...@chromium.org> wrote:
Triaged up to 552050.  No high level comments; crashers uninteresting
except for process (see other thread).

-- Orphaned bugs that IMO someone should look at:

* http://crbug.com/121085: There's an assertion that was put in by
Wan-Teh many years ago that's still firing, so I unarchived this bug.
It would probably be worthwhile for someone to come up to speed on the
code and figure out what the assertion is tripping.

Malware.

Search for closed bugs related to this; rvargas or cpu also did some of the investigation. I think rtenetti also did some light investigation, but 'malware'

Fairly sure archived is the 'right' answer for this, but rvargas was the last to own. Otherwise, I'd suggest reaching out to the Chrome/Windows folks (brucedawson, scottmg, cpu, jschuh, etc) or the crash folks, but with IIRC one exception (a true Chrome regression that was unfortunately masked by all the malware false positives for a while), it's been malware.

Randy Smith

unread,
Nov 5, 2015, 9:22:12 PM11/5/15
to Ryan Sleevi, Helen Li, Josh Karlin, Matt Menke, Nick Harper, net-dev
Whoops, sorry, I didn't check the set of bugs closely enough. Please
ignore that flag; I'll make the breadcrumbs right.

-- Randy

Raman Tenneti

unread,
Nov 10, 2015, 1:34:50 PM11/10/15
to Randy Smith, Ryan Sleevi, Helen Li, Josh Karlin, Matt Menke, Nick Harper, net-dev
Triaged up to 553711. 

crash bugs
+ crbug.com/533545 net::HttpCache::PendingOp::~PendingOp - - top crash Canary/Dev
+ crbug.com/553531 - Crash Report - net::TransportConnectJob::GetAdditionalErrorState
+ crbug.com/553522 Crash Report - net::ParsedCookie::SetupAttributes
crashes  (very low occurances)

crbug.com/552397 QUIC - accurately report the application protocol and QUIC version

+ Need feedback (I wasn't able to reproduce the following bugs)
++ crbug.com/552845 - Chrome Returns 414 (Request-URI Too Long
++ crbug.com/552790 - User lost their cookies after restart - can not reproduce ()
++ crbug.com/553410 - Connecting to any https site ends in timeout (during certificate verification)
++ crbug.com/552594 Unable to connect to Google websites - UDP traffic being blocked

thanks
raman


--
You received this message because you are subscribed to the Google Groups "net-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to net-dev+u...@chromium.org.
To post to this group, send email to net...@chromium.org.

Charles 'Buck' Krasic

unread,
Nov 17, 2015, 8:03:03 PM11/17/15
to Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Josh Karlin, Matt Menke, Nick Harper, net-dev
Triaged up to 557226.

New, very low occurances, crashers:

Feature request related to priorities: 556607cronet - add setPriority on UrlRequest
Need feedback:
 557087:chrome.sockets.tcp.secure doesn't prompt for client certificate when requested by server
 556374:Crashes when downloading the zip
555745:ERR_SSL_VERSION_OR_CIPHER_MISMATCH, rsleevi@ subsequently responded.




I realized I really wish I went to the debugging unconference session, trying to investigate 

Cr-Internals-Network bugs was slow going.

Bryan McQuade

unread,
Nov 19, 2015, 9:34:45 PM11/19/15
to Charles 'Buck' Krasic, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Josh Karlin, Matt Menke, Nick Harper, net-dev
Triaged up to 558252. No new crashers or issues of note.


Matt Menke

unread,
Nov 23, 2015, 6:37:05 PM11/23/15
to Bryan McQuade, Charles 'Buck' Krasic, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Josh Karlin, Nick Harper, net-dev
Triaged up to 560523.  Nothing too exciting:

* Another issue with the renderer keeping 6 media requests open and paused at once.
* The SimpleCache remains a top 20 crasher on Windows Canary - https://crbug.com/533545
* Some Microsoft IP virtualization driver and QUIC is also a top crasher - https://crbug.com/471986
* Also a QUIC + upload + offline apparently results in a check failure: https://crbug.com/560495

David Benjamin

unread,
Nov 25, 2015, 5:48:34 PM11/25/15
to Matt Menke, Bryan McQuade, Charles 'Buck' Krasic, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Josh Karlin, Nick Harper, net-dev
Triaged up to 561701.

New bugs:
https://crbug.com/560781 - Linux proxy config woes.
https://crbug.com/561069 - Someone found an unnecessary casts.
https://crbug.com/561343 - Cookies received during auth flow.

Investigating bugs:
https://crbug.com/464760 - Hilarious Windows quirk leaked to other platofrms. https://codereview.chromium.org/1469153006/ to fix.
https://crbug.com/458620 - Even more problems with the cache lock. We've got to fix this thing.

Crashes:
https://crbug.com/561038 - Some Chromecast change hit a CHECK in Canary.

Misha Efimov

unread,
Dec 10, 2015, 10:45:51 AM12/10/15
to David Benjamin, Matt Menke, Bryan McQuade, Charles 'Buck' Krasic, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Josh Karlin, Nick Harper, net-dev
Triaged from 567777 up to 568257.

New bugs:

https://crbug.com/568112 - HTTP/2 Breaks Large Uploads
https://crbug.com/568093 - Feature request for Dinosaur game highscore save and sync
https://crbug.com/568089 - File upload not working in new Chrome version. (v47) yellow blinking circle without completion. V46 is working fine
https://crbug.com/568110 - filesystem:file urls return incorrect inner path - filed by chrome dev

Investigated bugs:

https://crbug.com/530200 - Performance Point Excel Services Page display casues looping
I've got all the information from the reporter to reproduce issue locally. Observed tons of HIGH_PRIORITY requests with  NavigationResourceThrottle delegate being created and immediately aborted until I switch to another tab. Once switched those requests stop and page displays successfully. 

Per clamy@ "This should be normal for a navigation. You get a NavigationResourceThrottle on each navigation, it ensures that NavigationThrottles on the UI thread may run. On Android, there is one NavigationThrottle that can cancel the navigation, the InterceptNavigationThrottle that will redirect the request to an app."

I was going to try debug local build but didn't do it due to power failure.

Finch Chirps:

Nothing particularly notable except for some Disk cache spikes, which may or may not be an indication of real issue, should keep an eye on them.

Crashes:

No new crashes.


Siddharth Vijayakrishnan

unread,
Dec 10, 2015, 1:19:23 PM12/10/15
to Misha Efimov, David Benjamin, Matt Menke, Bryan McQuade, Charles 'Buck' Krasic, Raman Tenneti, Randy Smith, Ryan Sleevi, Helen Li, Josh Karlin, Nick Harper, net-dev
On 10 December 2015 at 07:45, Misha Efimov <m...@chromium.org> wrote:
Triaged from 567777 up to 568257.

New bugs:

https://crbug.com/568112 - HTTP/2 Breaks Large Uploads
https://crbug.com/568093 - Feature request for Dinosaur game highscore save and sync
https://crbug.com/568089 - File upload not working in new Chrome version. (v47) yellow blinking circle without completion. V46 is working fine
https://crbug.com/568110 - filesystem:file urls return incorrect inner path - filed by chrome dev

Investigated bugs:

https://crbug.com/530200 - Performance Point Excel Services Page display casues looping
I've got all the information from the reporter to reproduce issue locally. Observed tons of HIGH_PRIORITY requests with  NavigationResourceThrottle delegate being created and immediately aborted until I switch to another tab. Once switched those requests stop and page displays successfully. 

Per clamy@ "This should be normal for a navigation. You get a NavigationResourceThrottle on each navigation, it ensures that NavigationThrottles on the UI thread may run. On Android, there is one NavigationThrottle that can cancel the navigation, the InterceptNavigationThrottle that will redirect the request to an app."

I was going to try debug local build but didn't do it due to power failure.

Finch Chirps:

Nothing particularly notable except for some Disk cache spikes, which may or may not be an indication of real issue, should keep an eye on them.

​What's the deal with these ? I've seen them coming on a regular basis - it'd be nice to establish the right baseline that will help us separate a real issue from false alarms. ​
 

Randy Smith

unread,
Dec 10, 2015, 1:32:00 PM12/10/15
to Siddharth Vijayakrishnan, Misha Efimov, David Benjamin, Matt Menke, Bryan McQuade, Charles 'Buck' Krasic, Raman Tenneti, Ryan Sleevi, Helen Li, Josh Karlin, Nick Harper, net-dev
On Thu, Dec 10, 2015 at 1:19 PM, Siddharth Vijayakrishnan <si...@google.com> wrote:


On 10 December 2015 at 07:45, Misha Efimov <m...@chromium.org> wrote:
Triaged from 567777 up to 568257.

New bugs:

https://crbug.com/568112 - HTTP/2 Breaks Large Uploads
https://crbug.com/568093 - Feature request for Dinosaur game highscore save and sync
https://crbug.com/568089 - File upload not working in new Chrome version. (v47) yellow blinking circle without completion. V46 is working fine
https://crbug.com/568110 - filesystem:file urls return incorrect inner path - filed by chrome dev

Investigated bugs:

https://crbug.com/530200 - Performance Point Excel Services Page display casues looping
I've got all the information from the reporter to reproduce issue locally. Observed tons of HIGH_PRIORITY requests with  NavigationResourceThrottle delegate being created and immediately aborted until I switch to another tab. Once switched those requests stop and page displays successfully. 

Per clamy@ "This should be normal for a navigation. You get a NavigationResourceThrottle on each navigation, it ensures that NavigationThrottles on the UI thread may run. On Android, there is one NavigationThrottle that can cancel the navigation, the InterceptNavigationThrottle that will redirect the request to an app."

I was going to try debug local build but didn't do it due to power failure.

Finch Chirps:

Nothing particularly notable except for some Disk cache spikes, which may or may not be an indication of real issue, should keep an eye on them.

​What's the deal with these ? I've seen them coming on a regular basis - it'd be nice to establish the right baseline that will help us separate a real issue from false alarms. ​
 

Seconded.  I suspect they're real and I'd value seeing a more detailed analysis.

-- Randy

Matt Menke

unread,
Dec 10, 2015, 2:41:45 PM12/10/15
to Randy Smith, Siddharth Vijayakrishnan, Misha Efimov, David Benjamin, Bryan McQuade, Charles 'Buck' Krasic, Raman Tenneti, Ryan Sleevi, Helen Li, Josh Karlin, Nick Harper, net-dev
Misha:  Thanks so much for the detailed report!  The second upload issue looks like a significant breakage.  Think it would have likely slipped through the cracks without your email.

Misha Efimov

unread,
Dec 10, 2015, 2:47:47 PM12/10/15
to Matt Menke, Randy Smith, Siddharth Vijayakrishnan, David Benjamin, Bryan McQuade, Charles 'Buck' Krasic, Raman Tenneti, Ryan Sleevi, Helen Li, Josh Karlin, Nick Harper, net-dev
My pleasure, sorry that I haven't investigated much. 
BTW, what would be appropriate label for upload bugs?

Matt Menke

unread,
Dec 10, 2015, 2:49:56 PM12/10/15
to Misha Efimov, Randy Smith, Siddharth Vijayakrishnan, David Benjamin, Bryan McQuade, Charles 'Buck' Krasic, Raman Tenneti, Ryan Sleevi, Helen Li, Josh Karlin, Nick Harper, net-dev
There isn't one, unfortunately.  :(  I'd just use internals-network.

Raman Tenneti

unread,
Dec 14, 2015, 2:38:23 PM12/14/15
to Misha Efimov, David Benjamin, Matt Menke, Bryan McQuade, Charles 'Buck' Krasic, Randy Smith, Ryan Sleevi, Helen Li, Josh Karlin, Nick Harper, net-dev
Triaged from 568257 to 569627

new bugs:

 https://crbug.com/569437: 1 second lag when opening URLs from the command line
 https://crbug.com/569541569541: Encryption Key Problem (feature request??)
https://crbug.com/543913 : Latency increase in "SimpleCache.Http.Eviction.TimeToSelectEntries"

thanks
raman

Charles Harrison

unread,
Dec 17, 2015, 6:06:58 PM12/17/15
to Raman Tenneti, Misha Efimov, David Benjamin, Matt Menke, Bryan McQuade, Charles 'Buck' Krasic, Randy Smith, Ryan Sleevi, Helen Li, Josh Karlin, Nick Harper, net-dev
Triaged to 570823

new bugs:
Looks like this might be a devtools reporting issue. There's some funky javascript that loads an image then alters the src on an on-page image to the loaded image. What ends up on devtools is 200 (from cache) instead of 304.

Possible issue with launching an external protocol.

Weird chromebooks networking. Waiting on feedback.

So many proxy auth prompts! Might need someone skilled with reading netlogs.

interesting triaged bugs:

Sync XHR returns emptyr responses when cache is hit after etag verification. Tossed this to japhet@ after a bisect.

CrOS OpenVPN issue. Tossed to CrOS folks.

Chrome blocks SVG (base64) data uris. Tossed to safebrowsing


No new crashers.
No alerts.



Nick Harper

unread,
Dec 21, 2015, 3:34:34 PM12/21/15
to Charles Harrison, Raman Tenneti, Misha Efimov, David Benjamin, Matt Menke, Bryan McQuade, Charles 'Buck' Krasic, Randy Smith, Ryan Sleevi, Helen Li, Josh Karlin, net-dev
Triaged to 571443. No new crashers.

I filed a few bugs based on alerts: crbug.com/571441, crbug.com/571446, and crbug.com/571447. They're all SimpleCache on Android with the change in metric happening with the release of M49 to dev. 

Steven Valdez

unread,
Dec 29, 2015, 3:33:53 PM12/29/15
to Nick Harper, Charles Harrison, Raman Tenneti, Misha Efimov, David Benjamin, Matt Menke, Bryan McQuade, Charles 'Buck' Krasic, Randy Smith, Ryan Sleevi, Helen Li, Josh Karlin, net-dev
Triaged up through 572939. Some new alerts, though most of them are probably due to volume changes based on the holidays.

Bence Béky

unread,
Jan 4, 2016, 3:33:24 PM1/4/16
to Steven Valdez, Nick Harper, Charles Harrison, Raman Tenneti, Misha Efimov, David Benjamin, Matt Menke, Bryan McQuade, Charles 'Buck' Krasic, Randy Smith, Ryan Sleevi, Helen Li, Josh Karlin, net-dev
Triaged up to 574182.  Found one interesting network issue: https://crbug.com/573518, I'll follow up on that.

Since it was five days' worth of bugs in a single day rotation, I didn't get to dive into Chirp alerts or look at recent Cr-Internals-Network bugs, sorry.

Chris Bentzel

unread,
Jan 4, 2016, 4:00:38 PM1/4/16
to Bence Béky, Steven Valdez, Nick Harper, Charles Harrison, Raman Tenneti, Misha Efimov, David Benjamin, Matt Menke, Bryan McQuade, Charles 'Buck' Krasic, Randy Smith, Ryan Sleevi, Helen Li, Josh Karlin, net-dev
Do you think you will be able to look into them? I don't want to leave a gap of things dropped on the floor.

Randy Smith

unread,
Jan 4, 2016, 4:04:17 PM1/4/16
to Chris Bentzel, Bence Béky, Steven Valdez, Nick Harper, Charles Harrison, Raman Tenneti, Misha Efimov, David Benjamin, Matt Menke, Bryan McQuade, Charles 'Buck' Krasic, Ryan Sleevi, Helen Li, Josh Karlin, net-dev
I've thought of the bug triage rotation as "Do what you can during your time, and pass what you can't on to the next person".  So I think the best way to deal with this is for the next person to extend the chirp alerts they look at back in time.  We should probably include some official position on this issue in the triage docs, though.

-- Randy

Matt Menke

unread,
Jan 4, 2016, 4:33:15 PM1/4/16
to Randy Smith, Chris Bentzel, Bence Béky, Steven Valdez, Nick Harper, Charles Harrison, Raman Tenneti, Misha Efimov, David Benjamin, Bryan McQuade, Charles 'Buck' Krasic, Ryan Sleevi, Helen Li, Josh Karlin, net-dev
That was also my intention of how things should work when drafting the rotation docs.

Randy Smith

unread,
Jan 12, 2016, 4:46:49 PM1/12/16
to net-dev
[-everyone except net-dev@, just for cleanup.]

My triage report M/T this week.  Nothing really exciting:
* I triaged up to 576833.
* Crashers took the plurality of my time; I looked at 12 platform/channel combinations (7 releases) the first day, and 4 platform/release combinations (4 releases) the second day).  I did notice that when you look at a platform/release combination in chromecrash you get the bug, note, and edit links inline in the magic signature lists, and if you look at just releases without a platform, you don't.  So that may be a reason to look at platform/releases in combination.

* I filed a couple of new bugs based on crash reports:
** http://crbug.com/576403: QUIC crasher.
** http://crbug.com/576407: Cros Crasher; closed after Matt pointed out a recent CL has renderered it moot.
** http://crbug.com/576442: Android assertion crash: PrecacheFetcher calling URLFetcherCore with invalid URL
** http://crbug.com/576445: Low frequency Android crasher in HPACK
** http://crbug.com/576860: QUIC crasher.

* I'm concerned about http://crbug.com/537232, which is showing up a lot on ChromeOS stable (0.28% of crashes on that channel, which ok, isn't sky high, but is a solid presence).  Currently assigned to Raman because it was originally thought to have to do with QUIC, but doesn't actually have to do with QUIC.  Looks like it'd be a PITA to debug, but still my candidate from this week for "bug that might be worth putting more energy into".

-- Randy


Josh Karlin

unread,
Jan 19, 2016, 7:23:19 AM1/19/16
to net-dev
Short triage duty (just Friday as Monday was a holiday). Triaged up to 578480. A few unnecessary chirps. No particularly interesting bugs of note. No new crashers of note.

--
You received this message because you are subscribed to the Google Groups "net-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to net-dev+u...@chromium.org.
To post to this group, send email to net...@chromium.org.

Helen Li

unread,
Jan 20, 2016, 5:41:34 PM1/20/16
to Josh Karlin, net-dev
Triage duty Tuesday and Wednesday (1/20). No new crashers and notable bugs. 
ERR_CONTENT_DECODING_FAILED went up in M47 stable and triggered a chirp alert. I looked through commit history and SDCH UMA data, but didn't find any correlation. Next triager maybe could follow up on this.

Charles 'Buck' Krasic

unread,
Jan 25, 2016, 6:49:21 PM1/25/16
to Helen Li, Josh Karlin, net-dev
Sorry for the late report.

My time was entirely spent on crashers.   

New crashers:


Issue 580725:

Chrome: Crash Report - net::URLRequestHttpJob::NotifyBeforeNetworkStart


Issue 580727:

Chrome: Crash Report - [Assert] net::`anonymous namespace'::IsEOL


Issue 580729:

Chrome: Crash Report - net::URLRequestHttpJob::StartTransactionInternal


Issue 580731:

Chrome: Crash Report - net::internal::ClientSocketPoolBaseHelper::Group::RemovePendingRequest




Ryan Hamilton

unread,
Jan 27, 2016, 2:44:05 PM1/27/16
to Charles 'Buck' Krasic, Helen Li, Josh Karlin, net-dev
Triaged bugs up to 581464. Only 1 possible network bug (about a proxy not working,but I'm sure this is not actually a chrome problem.)

This crash has been around for a while and needs some attention:
Chrome: Crash Report - net::HttpCache::PendingOp::~PendingOp


Bryan McQuade

unread,
Jan 29, 2016, 9:05:04 AM1/29/16
to Ryan Hamilton, Charles 'Buck' Krasic, Helen Li, Josh Karlin, net-dev
Triaged bugs up to 582437.

Issue "Crash Report - net::QuicConnection::CloseConnection" https://code.google.com/p/chromium/issues/detail?id=546668 appears to be happening on stable and should probably get an owner assigned so we can investigate and fix. Stack:
net::QuicConnection::CloseConnection(net::QuicErrorCode,bool)
net::QuicConnection::SendConnectionCloseWithDetails(...)
net::QuicConnection::SendConnectionClose(net::QuicErrorCode)
net::QuicConnection::CheckForTimeout()
net::`anonymous namespace'::TimeoutAlarm::OnAlarm()
net::QuicAlarm::Fire()

Nothing else notable.


Matt Menke

unread,
Feb 2, 2016, 12:20:47 PM2/2/16
to Bryan McQuade, Ryan Hamilton, Charles 'Buck' Krasic, Helen Li, Josh Karlin, net-dev
Triaged up to 583142

Issue that I think still needs more general investigation:
https://crbug.com/580770 and https://crbug.com/581588 - login failing (One a google site, one not).  May or may not be some sort of cookie issue.  Second issue looks to be a 48 to 49 regression.  Second not yet clear, but seems likely to be in the same range.
https://crbug.com/581288 - issue where we're marking a filtering proxy as bad (Unclear if this is a Chrome bug, but we're apparently not matching behavior of other browsers here).

Selected other new net bugs:
https://crbug.com/582810 - HTTPS cert dialog focus issue on OSX.
https://crbug.com/580906 - some sort of download issue with lots of.crdownload files sticking around.
https://crbug.com/582623 - A recently added DCHECK in URLRequestHttpJob is triggering, on Android, for unclear reasons.  I'm taking ownership.

Asanka Herath

unread,
Feb 3, 2016, 8:49:19 PM2/3/16
to Matt Menke, Bryan McQuade, Ryan Hamilton, Charles 'Buck' Krasic, Helen Li, Josh Karlin, net-dev
Triaged up to 584107. No big trends.

There was some metrics related alerts due to the M48 release of Chrome on iOS switch to WkWebView having a large effect on its use of //net. There may be more alerts as things settle down.

David Benjamin

unread,
Feb 9, 2016, 7:12:05 PM2/9/16
to Asanka Herath, Matt Menke, Bryan McQuade, Ryan Hamilton, Charles 'Buck' Krasic, Helen Li, Josh Karlin, net-dev
Triaged up to 585622. No big trends (as usual, lots of malware). One thing of note is Matt and I determined https://crbug.com/556928 is probably yet another of our classic AssertNoURLRequests shutdown bug, but due to lifetime problems in ProfileIOData, manifests very differently.

Charles Harrison

unread,
Feb 11, 2016, 5:08:17 PM2/11/16
to David Benjamin, Asanka Herath, Matt Menke, Bryan McQuade, Ryan Hamilton, Charles 'Buck' Krasic, Helen Li, Josh Karlin, net-dev, Julia Tuttle
Triaged up 586276

No new crashers of note

Gasper alerts: 2 alerts for DNS related regressions on Linux. We definitely need to figure out how these happened. Matt noticed these are not tied to any release, but to a date. Did someone merge a change? What happened here? Julia, any insight into these metrics?

New(ish) issues:

Navigation to a page with HTTP status 304 redirects me to the previous page
This one seems to be a problem with following a 302 Found. Includes pcap. I couldn't see anything wrong with what the endpoint gave us, but we were definitely doing something horribly wrong. We also seem to be going over our redirect limit (20).

prompt_for_download appears when set to False
Edge case for auto downloads where the directory is not available. Filer said they could repro even when the directory exists. Possible guess is that unmounting causes this behavior and remounting does not fix it?

Chrome keeping mobile radio active
This phone is sending out way too many packets on the mobile network. Asked for a net-export.


Eric Roman

unread,
Feb 22, 2016, 3:12:50 PM2/22/16
to net-dev
Triaged up through 588814

* The Switchover to the new issue-tracker happened during my watch. Went smoothly, just a few differences to the workflow, being updated in the documentation:

   - Way to subscribe to bugs has changed (https://codereview.chromium.org/1720053002/)
   - Labels have changed some stuff has been re-branded "components" (https://codereview.chromium.org/1722763002/)
   - Preview (h) no longer works when browsing bugs (TODO: I don't have a solution yet)
   - Been getting a few spurious 500's and slowness, but nothing a refresh or 2 didn't fix, or otherwise affected productivity.

* Only a few "network" bugs that were new or had activity:
  
* I didn't get around to investigating this (old) issue but would be a good task for future triagers if you have time. Occasional leak of IOThread::globals_. Some ideas in comment #6 to get started.

* There were some internal alerts on a few histograms. I have followed up on the internal threads. Mainly noise, with the exception of possibly SimpleCache.App.CreationToIndex being a real one. It had fired earlier for non-Cros platforms, and then fired during my rotation for Cros. Although even that looks suspiciously like noise, see bug.

Matt Mueller

unread,
Feb 29, 2016, 10:30:23 PM2/29/16
to net-dev
Triaged up through 590949. Nothing exciting.

Only one new crasher, which was already addressed by the time I saw it:

Didn't look at chirps. Will try to do that tomorrow.

Misha Efimov

unread,
Mar 3, 2016, 10:37:52 AM3/3/16
to Matt Mueller, net-dev
Triaged up through 591445.

Spend some time looking into:

Issue 589121:⚑Unable to resume browsing on a existing page after disable and enable of data
Issue 590487:⚑Browser tries to downlaod instead of rendering plain text TCP stream
WONTFIX: 591438: Files are CHANGED (corrupted) on DOWNLOAD unexpectedly

Filed

DUP: 591093: - Chrome_Android: Crash Report - [Assert] net::QuicHttpStream::ResetStream

Fixed the link to query for old bugs that need feedback in net bug triage suggested workflow.
Pinged / closed old bugs without feedback.


Bence Béky

unread,
Mar 5, 2016, 4:45:45 PM3/5/16
to net-dev
Triaged up through https://crbug.com/592106.  There's lots of complaints about page actions disappearing from Omnibar, like https://crbug.com/572522, https://crbug.com/578051https://crbug.com/590320https://crbug.com/591301https://crbug.com/591485https://crbug.com/592092, https://crbug.com/592106.

No new crashers.  Spent some time trying to tackle existing crasher bugs, no progress made.

Rejected the single Chirp alert during my rotation.


Steven Valdez

unread,
Mar 8, 2016, 5:06:26 PM3/8/16
to Bence Béky, net-dev
Triaged up through: 593132. Nothing too exciting, some cert dialog issues with Android that might be corp-related (591797).

Filed a few spiking crasher bugs, but nothing exciting. Still more malware related crashes and a few iOS crashes.

Rejected one noise Chirp alert and filed 593105 for an Android SimpleCache spike.

Matt Menke

unread,
Mar 10, 2016, 5:05:47 PM3/10/16
to Steven Valdez, Bence Béky, net-dev
Triaged up to 593889.

Nothing exciting, just the usual bunch of reports that more often than not are due to server / middlebox weirdness.

There's a new sheriffbot in town that likes to assign Needs-Feedback issues to the person who added the Needs-Feedback label, after anyone makes a comment.  If it incorrectly assigns a bug to you, feel free to just unassign the bug, particularly if you were the triager when you added the label.  Hope we'll get that behavior removed, sooner rather than later.

Ryan Sleevi

unread,
Mar 10, 2016, 5:10:57 PM3/10/16
to Matt Menke, Steven Valdez, Bence Béky, net-dev
On Thu, Mar 10, 2016 at 2:05 PM, Matt Menke <mme...@chromium.org> wrote:
Triaged up to 593889.

Nothing exciting, just the usual bunch of reports that more often than not are due to server / middlebox weirdness.

There's a new sheriffbot in town that likes to assign Needs-Feedback issues to the person who added the Needs-Feedback label, after anyone makes a comment.  If it incorrectly assigns a bug to you, feel free to just unassign the bug, particularly if you were the triager when you added the label.  Hope we'll get that behavior removed, sooner rather than later.

Correction: If you see something, say something

https://www.chromium.org/issue-tracking/autotriage lists the logic, as well as the short-links to filing bug for things you think are borked

For example, I filed https://bugs.chromium.org/p/chromium/issues/detail?id=593824 for what Matt mentioned, based on seeing the behaviour when he pinged a bug. 

Matt Menke

unread,
Mar 10, 2016, 5:13:35 PM3/10/16
to Ryan Sleevi, Steven Valdez, Bence Béky, net-dev
I also filed a bug, actually...As it's not fixed yet, my comment seems to stand.

Jana Iyengar

unread,
Mar 14, 2016, 10:41:45 PM3/14/16
to Ryan Sleevi, Matt Menke, Steven Valdez, Bence Béky, net-dev
Triaged through 594628. Nothing particularly exciting.
Also dug into and updated crasher crbug/499663, looks like a use-after-free bug. Responded to a few interesting finch-chirp alerts.


--
You received this message because you are subscribed to the Google Groups "net-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to net-dev+u...@chromium.org.
To post to this group, send email to net...@chromium.org.

Helen Li

unread,
Mar 22, 2016, 5:30:02 PM3/22/16
to net-dev
Triaged on Monday and Tuesday. No new chirp alerts. Filed a couple low frequency crashers. Tried digging into some of the crashers, but didn't get far.

Josh Karlin

unread,
Mar 24, 2016, 3:17:40 PM3/24/16
to Helen Li, net-dev
Triaged Wed. and Thu. up to 597678. No new crashers w/ > 1 user. Chirp seems to have picked up on https://bugs.chromium.org/p/chromium/issues/detail?id=589473 which is leading to increased SimpleCache volume and delay.

Miriam Gershenson

unread,
Mar 28, 2016, 6:02:16 PM3/28/16
to Josh Karlin, Helen Li, net-dev
Triaged up to 598373. Looked at a few chirp alerts. No new crashers. Spent some time looking at a few Internals>Network bugs, didn't really get anywhere.

Randy Smith

unread,
Mar 30, 2016, 5:15:08 PM3/30/16
to Miriam Gershenson, Josh Karlin, Helen Li, net-dev
[I may do some more triage work tonight, but I think I can write the
handoff email now.]

TL;DR: Mostly harmless.

* Triaged up to 599214.

* Note on active bug: http://crbug.com/595968: The next step is
probably to pull in someone who knows about tab process creation and
its use in navigation; I was going to grab nasko@, but was waiting on
the last thing I asked the user to do, so the next triager may want to
pull in nasko@.

* I lucked out on crashes; I only needed to look at canaries, and none
of them had anything interesting in them. There were a surge up in
instances of crashes in URLFetcherCore, but 10/11 of them came from a
single users, so ignorable--I took the opportunity to clear out the
(all Fixed/Verified) list of bugs associated with that CHECK and
update the note.

I think there's a problem in the ChromeOS crash reporting
pipeline--the most recent activity on
https://crash.corp.google.com/root?platform=cros is two months ago.
If anyone knows who to ping about this, please let me know.

* Added some minor tweaks to triage documentation
(https://codereview.chromium.org/1836343002).

* As Matt mentioned, there were a lot of old unresponded to chirp
alerts, and I arbitrarily decided to ignore everything before last
Friday (the existence of a response is how future triagers know that
they don't need to look at the alert).

-- Randy
> https://groups.google.com/a/chromium.org/d/msgid/net-dev/CAKKC82ckj9BTidKhBBUW93LKoC%3DYTNKAdBA_r2syWAk3zL8D-w%40mail.gmail.com.

Eric Roman

unread,
Mar 30, 2016, 5:38:31 PM3/30/16
to Randy Smith, Miriam Gershenson, Josh Karlin, Helen Li, net-dev
On Wed, Mar 30, 2016 at 2:15 PM, Randy Smith <rds...@chromium.org> wrote:
[I may do some more triage work tonight, but I think I can write the
handoff email now.]

TL;DR: Mostly harmless.

* Triaged up to 599214.

* Note on active bug: http://crbug.com/595968: The next step is
probably to pull in someone who knows about tab process creation and
its use in navigation; I was going to grab nasko@, but was waiting on
the last thing I asked the user to do, so the next triager may want to
pull in nasko@.

* I lucked out on crashes; I only needed to look at canaries, and none
of them had anything interesting in them.  There were a surge up in
instances of crashes in URLFetcherCore, but 10/11 of them came from a
single users, so ignorable--I took the opportunity to clear out the
(all Fixed/Verified) list of bugs associated with that CHECK and
update the note.

I think there's a problem in the ChromeOS crash reporting
pipeline--the most recent activity on
https://crash.corp.google.com/root?platform=cros is two months ago.
If anyone knows who to ping about this, please let me know.

Follow-up is happening on internal bug 27926763.

The problem is in how recent builds are summarized on that page, not in actually collecting ChromeOS crashes (i.e. the data is there, you just don't see links to the recent releases).

Shivani Sharma

unread,
Apr 1, 2016, 4:57:54 PM4/1/16
to net-dev, Randy Smith, Miriam Gershenson, Josh Karlin, Helen Li
Triaged upto 599972.
No new chirp alerts.
No new crashers.

Asanka Herath

unread,
Apr 7, 2016, 9:54:33 PM4/7/16
to Shivani Sharma, net-dev, Randy Smith, Miriam Gershenson, Josh Karlin, Helen Li
Triaged up to to 601599. No trends.

Bryan McQuade

unread,
Apr 11, 2016, 4:24:45 PM4/11/16
to Asanka Herath, Shivani Sharma, net-dev, Randy Smith, Miriam Gershenson, Josh Karlin, Helen Li
Triaged up to 602421. Nothing notable.

Bence Béky

unread,
Apr 13, 2016, 5:13:48 PM4/13/16
to Bryan McQuade, Asanka Herath, Shivani Sharma, net-dev, Randy Smith, Miriam Gershenson, Josh Karlin, Helen Li

Charles 'Buck' Krasic

unread,
Apr 15, 2016, 8:22:42 PM4/15/16
to Bence Béky, Bryan McQuade, Asanka Herath, Shivani Sharma, net-dev, Randy Smith, Miriam Gershenson, Josh Karlin, Helen Li
Partial report.  Triaged up to 604037.  Nothing exciting.   I didn't get time to check crashers,  will see how much I can do by Monday.

Charles Harrison

unread,
Apr 19, 2016, 4:45:07 PM4/19/16
to Charles 'Buck' Krasic, Bence Béky, Bryan McQuade, Asanka Herath, Shivani Sharma, net-dev, Randy Smith, Miriam Gershenson, Josh Karlin, Helen Li
Triaged to 604867

No new crashers.

One confirmed bug:
- crbug.com/604779 - invalid redirects cause terrible UX.

David Benjamin

unread,
Apr 21, 2016, 8:21:07 PM4/21/16
to Charles Harrison, Charles 'Buck' Krasic, Bence Béky, Bryan McQuade, Asanka Herath, Shivani Sharma, net-dev, Randy Smith, Miriam Gershenson, Josh Karlin, Helen Li
Triaged up to 605789

Nothing terribly exciting, apart from noticing that https://crbug.com/594226 was probably another instance of auto-reload masking stuff, so we really need to be watching those metrics.

I started putting together a fix for https://crbug.com/604779 as part of dealing with existing bugs since that code is otherwise unowned.

Eric Roman

unread,
Apr 25, 2016, 9:14:49 PM4/25/16
to David Benjamin, Charles Harrison, Charles 'Buck' Krasic, Bence Béky, Bryan McQuade, Asanka Herath, Shivani Sharma, net-dev, Randy Smith, Miriam Gershenson, Josh Karlin, Helen Li
Triaged up to 606610.

Nothing too interesting. Some excerpts:

* Realm string no longer shown in HTTP auth dialog -- https://crbug.com/606392

* Rewinding system clock does not cause chrome to accept (presumably expired) cookies -- https://crbug.com/606132

* Port number 65535 not allowed in navigations -- https://crbug.com/605367

Josh Karlin

unread,
Apr 29, 2016, 3:02:45 PM4/29/16
to Eric Roman, Asanka Herath, Bryan McQuade, Helen Li, Randy Smith, Shivani Sharma, Miriam Gershenson, Bence Béky, David Benjamin, Charles 'Buck' Krasic, net-dev, Charles Harrison

Triaged up to 607944. Nothing exciting to report.

Jana Iyengar

unread,
May 3, 2016, 6:38:27 PM5/3/16
to Josh Karlin, Eric Roman, Asanka Herath, Bryan McQuade, Helen Li, Randy Smith, Shivani Sharma, Miriam Gershenson, Bence Béky, David Benjamin, Charles 'Buck' Krasic, net-dev, Charles Harrison
Triaged up through 608900. Nothing specifically interesting to report.

Matt Mueller

unread,
May 9, 2016, 10:07:57 PM5/9/16
to Jana Iyengar, Josh Karlin, Eric Roman, Asanka Herath, Bryan McQuade, Helen Li, Randy Smith, Shivani Sharma, Miriam Gershenson, Bence Béky, David Benjamin, Charles 'Buck' Krasic, net-dev, Charles Harrison
Triaged up to 610541. No interesting bugs.

No new crashes that meet the bar.

Chirps:
EMPTY_RESPONSE spike on Android, all channels: https://bugs.chromium.org/p/chromium/issues/detail?id=610436
Spike in HttpCache.EntryLockWait on Win dev: https://bugs.chromium.org/p/chromium/issues/detail?id=610439
Spike in Net.HttpJob.TotalTimeNotCached on Win, all channels: https://bugs.chromium.org/p/chromium/issues/detail?id=610458
Spike in SimpleCache.Http.EntryCreationTime on Cros Beta: https://bugs.chromium.org/p/chromium/issues/detail?id=610487



Raman Tenneti

unread,
May 11, 2016, 11:20:00 PM5/11/16
to Matt Mueller, Jana Iyengar, Josh Karlin, Eric Roman, Asanka Herath, Bryan McQuade, Helen Li, Randy Smith, Shivani Sharma, Miriam Gershenson, Bence Béky, David Benjamin, Charles 'Buck' Krasic, net-dev, Charles Harrison
Triaged up to 611265. No interesting bugs/crashes.

It is loading more messages.
0 new messages