<input> leaks memory if <input type="password"> exists
Reported by
michael....@uni-due.de,
Oct 22
|
||||||||
Issue descriptionUserAgent: Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/70.0.3538.67 Safari/537.36 Steps to reproduce the problem: 1. website with <input type="password">, static or create with JS 2. Use webserver to load that website 3. create and remove new <input> fields What is the expected behavior? Memory usage and number of DOM nodes should stay constant (after GC). Like in "file.PNG" (website loaded with "file:///C:/..."). What went wrong? Memory usage and number of DOM nodes raises (webserver.PNG). In my real application the memory usage raises very fast because not just the <input> nodes stay in memory but also all DOM nodes in the "neighborhood". The Memory usage raises permanently until the memory is exhausted. Did this work before? N/A Does this work in other browsers? N/A Chrome version: 70.0.3538.67 Channel: stable OS Version: 6.1 (Windows 7, Windows Server 2008 R2) Flash Version: Same results with Ubuntu 18.04.
,
Oct 22
,
Oct 25
,
Oct 26
Unable to reproduce the issue on win-7 and win-10 using chrome reported version #70.0.3538.67 and latest canary #72.0.3592.0. Attached a screen cast for reference. Following are the steps followed to reproduce the issue. ------------ 1. Opened index.html using a webserver. 2. Created and removed new <input> fields. 3. Observed that memory usage and number of DOM nodes stayed constant as expected. reporter@ - Could you please check the attached screen cast and please let us know if anything missed from our end. Also please check the issue on latest canary #72.0.3592.0 by creating a new profile without any apps and extensions and please let us know if the issue still persist or not. Thanks...!!
,
Oct 26
Thanks for your reply. All my tests were done with a profile without any apps or extensions. A college of mine can reproduce the issue on win-10 with chromium #72.0.3593.0. But after watching the screen cast, I suspect we misunderstood. For this issue it is mandatory to create and remove <input> FIELDS (not just the user input). I'm doing that with setIntervall for 10 seconds (see script in index.html).
,
Oct 26
Thank you for providing more feedback. Adding the requester to the cc list. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Nov 6
I recorded the screen with Chromium #72.0.3595.2 on Ubuntu 18.04. I created a more complete example which additionally shows leaking is stopped after the password field was removed. The Steps of the test: 0) <0sec: start recording memory 1) 0sec: page refresh 2) GC triggered manually 3) 3sec: start to create and remove <input> fields (interval=300ms) 4) 13sec: stop interval (it runs 10000/300> 33 times) 5) GC triggered manually 6) 16sec: remove password field 7) GC triggered manually 8) 19sec: start to create and remove <input> fields (interval=300ms) (like step 3)) 9) 29sec: stop interval (it runs 10000/300> 33 times) 10) GC triggered manually Results: - Number of DOM nodes increases from 25 to 73 (step 2 vs step 5). - Old nodes are not removed after removing tge password field (step 5 vs step 7). - After removing the password field the number of nodes stay constant 73 (step 7 vs step 10).
,
Nov 13
As per comment #5 and comment #7, retried the issue on latest chromium 72.0.3610.0 using windows 10 and ubuntu 14.04. Steps: ----- 1. Launched chromium 2. Opened given index.html file using a web server 3. Navigated to Dev tools > Performance > started recording 4. Tried to click GC trigger manually As we are observed that the recording stops at 3 seconds @Reporter: Could you please check the attached screen cast and let us know if anything missed from our end. If possible provide any inputs for further triaging it.
,
Nov 13
Thanks for your reply. Unfortunately our tests are very different: I described 10 steps (about 30 seconds). Your test ended after step 2. You should use the other record button (and stop the recording manually after step 10).
,
Nov 13
Thank you for providing more feedback. Adding the requester to the cc list. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Nov 26
Able to reproduce the issue on reported chromium version #70.0.3538.0 and latest chromium #72.0.3622.0 using Mac 10.13.6, Windows 10 and Ubuntu 17.10. Note: As we could not download the branch builds of chromium, hence checked on the above versions. The behavior is seen from old M-60 builds(#60.0.3112.0). This is a non-regression issue, hence marking it as untriaged and requesting someone from the dev team to look into the issue. Thanks.! |
||||||||
►
Sign in to add a comment |
||||||||
Comment 1 by phanindra.mandapaka@chromium.org
, Oct 22