#18266 closed bug (duplicate)

WebPositive: memory leak triggered Debugger

Reported by: cocobean Owned by: pulkomandy
Priority: normal Milestone: Unscheduled
Component: Kits/Web Kit Version: R1/Development
Keywords: Cc:
Blocked By: #15934 Blocking:
Platform: All

Description

Haiku hrev56773 x86 memory leak test:

WebPositive 1.3 alpha
HaikuWebKit 1.9.3 (WebKit 615.1.15)
  • Test URL: https://news.yahoo.com
  • Test time: 170m (almost 3 hours)
  • Memory consumed: 1.5GiB RAM (triggered debugger at this point)

NOTE: I didn't use computer for three hours after launching WebPositive. This was not a "out of memory" issue (3.43 GiB RAM total, 1.70 GiB RAM used (i.e. About Haiku info)).

Attachments (3)

WebPositive-1100-debug-17-02-2023-15-49-07.txt (72.5 KB ) - added by cocobean 15 months ago.
HaikuWebKit 1.9.3 memory leak - hrev56773 x86
WebPositive-766-debug-02-04-2023-03-58-06.report.txt (25.9 KB ) - added by un_spacyar 13 months ago.
WebPositive-776-debug-03-04-2023-09-39-28.report (25.8 KB ) - added by bipolar 13 months ago.

Download all attachments as: .zip

Change History (7)

by cocobean, 15 months ago

HaikuWebKit 1.9.3 memory leak - hrev56773 x86

comment:1 by cocobean, 15 months ago

NOTE: Yahoo news webpage consumed only 134-150 MiB RAM at launch.

comment:2 by un_spacyar, 13 months ago

Hi. I have the same issue, but when open any Github page. I attach my Webpositive debug report (WebPositive-766-debug-02-04-2023-03-58-06.report.txt)

Haiku hrev56889 (32 bit) 2 GB RAM

Last edited 13 months ago by un_spacyar (previous) (diff)

comment:3 by bipolar, 13 months ago

Same on beta4 32 bits, 2 GB RAM, VBox VM, with only an idle (forgotten in a different Workspace) page open on https://pypi.org/project/pyaes.

comment:4 by pulkomandy, 10 months ago

Blocked By: 15934 added
Keywords: memory leak removed
Resolution: duplicate
Status: newclosed

We have 3 debug reports here, none of which are an out of memory issue and they are all in completely different places in WebKit.

This does not make my life easy. Please open a separate issue for each crash you encounter. I can review the bugreports and decide to close them as duplicates if needed. But I can't split a ticket into multiple ones.

The first bug report has a very large number of "crypto queue" threads. So I'm closing this as a duplicate of #15934. Please open new tickets if you have other problems.

Note: See TracTickets for help on using tickets.