4 ms·
http://www.google.com/support/chrome/bin/answer.py?answer=114836&hl=en http://www.google.com/support/chrome/bin/answer.py?answer=11... http://inari.aerdan.org/
by agl 16y ago
http://www.google.com/support/chrome/bin/answer.py?answer=114836&hl=en http://www.google.com/support/chrome/bin/answer.py?answer=11...
http://inari.aerdan.org/gallery/chromeopts.png http://inari.aerdan.org/gallery/chromeopts.png
The initial request is the GoogleURLTracker class, which manages what Google TLD you're on. On the trunk, this won't happen if your default search engine is not Google, and even if it does happen, we won't save/update any cookies as a result of the update request.
Also, the reporter is using a binary downloaded from the continuous builder, that's why RLZ is included.
- Revisor 16y agoNone of those settings seem to apply to the case in question: It seems that on first request it query google for unique client ID (getting NID Cookie) Requests to safebrowsing are not anonymous. All of them signed with cookie, getted on the first 2-3 requests EDIT: You added those two sentences. The guy claims changing the default search engine didn't change this behavior.
- elptacek 16y agoDid he clear the browsing data, too? I don't use Chromium, but I've seen similar requests with other browsers. The only way I have found to get the browser to stop sending the requests is to zorch all the browsing history, saved cookies, cached garbage, burn the village to the ground and sacrifice small birds. Of course, as soon as you search for soemthing against google.com, the requests will happily start up again.
- qeorge 16y agoHe specifically addresses the 'default search engine' possibility, and that's not it: gulbrandr: Can you please redo your test with a different default search engine than Google? Alexey: Already did it. Result is the same, nothing changed.
- cookiecaper 16y agoThe grandparent indicates that this is corrected in trunk.
- studer 16y agoThe blog post he links to: http://blog.chromium.org/2010/06/in-open-for-rlz.html http://blog.chromium.org/2010/06/in-open-for-rlz.html says "Chromium will also continue to exist as it always has, without any RLZ library included. And, you can still download a Google Chrome with no RLZ behavior at www.google.com/chrome. But now that RLZ is open, Google Chrome distributed through promotional means will include this open-source implementation of RLZ.". So what makes the continuous Chromium build a "Google Chrome distributed through promotional means"?
- agl 16y ago> What makes the continuous Chromium build a "Google Chrome distributed through promotional means"? The continuous builders exist to compile and test all the code. Therefore they build with RLZ options enabled. The continuous build binaries are literally whatever the state of the tree happened to be at the time. They're useful for bisection but people aren't expected to run them for non-development reasons.
- cake 16y agoWell that's very misleading, even Wikipedia thinks that : " Google confirmed that the RLZ tracking token [...] [is] not in versions of Chrome downloaded from the Google website directly or in any versions of Chromium." http://en.wikipedia.org/wiki/Chromium_(web_browser)#Differences_between_Chromium_and_Google_Chrome http://en.wikipedia.org/wiki/Chromium_(web_browser)#Differen...
- studer 16y agoFrom the thread: "We think that the reported RLZ ping behavior (that request to "/tools/pso/ping...") is a regression -- RLZ pings should not be happening in Chromium or non-promotional Google Chrome builds, so this is not the intended behavior."
- sprout 16y agoIs a separate ID generated for a porn mode session? What does Google do with that data? This might be a good reason not to recommend Chrome of any stripe for banking.