mozilla :: #introduction

19 Apr 2017
11:38alexofenIs this channel the right one to ask something about e10s?
11:40Aryxmaybe
11:41Aryxdepends on the question
11:45alexofenI wonder why in multi-process firefox (providing sandboxing/security as a main motivator), there is not 1 tab = 1 process, but instead I read about 2 ... 4 content processes? any idea why so?
11:48alexofenIt would seem with only 2 content processes, tabs will share a process, hence security is limited? I would imagine that on the positive side of it some memory footprint reduction might be expected, yet then why change for multiprocess in the firstplace?
11:52alexofen@Arys (in case my question is in the wrong place here #introduction, where should I reask it?)
12:13Aryxalexofen: each content process comes with overhead
12:13Aryxso one process per tab means more memory consumption
12:26alexofenAryx: So lowest memory consumption is with only 1 process, sure! But is multiprocess e10s not intentionally accepting memory overhead... for the sake of security? At least the need for security/sandboxing was always highlighted whenever talk is about e10s. Yet FF only groups xx tabs into 2 to 4 processes, not much security is gained, as still memory is shared between those content processes tabs.
12:26alexofenConsidering that e10s requires contributers such as add-on/extension coders to jump through loops, I would have imagined that 1 tab = 1 process is the goal, since else security always remains partial/broken?
12:27Aryxi wasn't aware that security got so much highlighted, had more the impression that it's about responsiveness and stability
12:31alexofenEven for stability only it would make seems to me that (1 tab = 1 process) is the more logical choice, or else a bad hang in one tab might de-stabilize / crash all tabs in the same content-process? Well, I will continue to seek some statement what e10s is indeed trying to achieve. Maybe my desire for some proper browser security (hence at least sandboxing tabs to OS processes), made me read
12:31alexofensecurity into e10s design goals more than it actually is, thank you!
23:18faserranohello, I am new and working on bug 1357832. I have my code changes ready but im a little confused on whether I should use MozReview, or attach a patch to Bugzilla
23:18firebothttps://bugzil.la/1357832 NEW, nobody@mozilla.org Use standard CSS border instead of -moz-border-*-colors in eula.css
23:34fserranohello, I am new and working on bug 1357832. I have my code changes ready but im a little confused on whether I should use MozReview, or attach a patch to Bugzilla
23:34firebothttps://bugzil.la/1357832 NEW, nobody@mozilla.org Use standard CSS border instead of -moz-border-*-colors in eula.css
23:36mbrubeckfserrano: I think pushing to MozReview is generally preferred these days
23:37mbrubeckbut either way works, especially for small/simple patches
23:37fserranook, thank you for clarifying
20 Apr 2017
No messages
   
Last message: 156 days and 15 hours ago