XS11E wrote:
"David B." "David wrote:

On 05/12/2018 18:49, XS11E wrote:
"David B." "David wrote:

Microsoft Replacing Edge With New Chromium-based Browser
By Mayank Parmar
December 4, 2018

https://www.bleepingcomputer.com/new...soft-replacing
-edge-with-new-chromium-based-browser/
Have to disagree with the article, I've been using Edge for sometime
now and have had issues so far.
Did you mean "... and have had *NO* issues so far"?

Correct, Edge works OK, my fingers don't seem to.
Thanks.

Did you try my "giant PDF file" test case ?

MSEdge causes a garbage collector for the desktop
to rail on one CPU core for around 20 seconds after
you examine a large PDF with the MSEdge PDF plugin.
During this time, any double clicks you do on
the desktop, are ignored. (You have to wait 20 seconds,
before you can start, say, Notepad.)

Microsoft patched the recurrence of those railed
CPU behaviors, but haven't cured the initial 20 second
outage. The problem is still there. The test PDF file
used for this, is a PDF file from the Microsoft site.
It contains 26,300 pages.

That's not a bug by the way, it's an architectural
shortcoming they won't be fixing.

This is what Metro applications bring to a desktop
which is HTML/js based.

Just a reminder of what Edge looks like.

https://s26.postimg.cc/6g1kartix/tab_thumb_edge.gif

And I don't set out with an "evil plan" to break stuff
in 2018. I just try stuff for the first time. And it
breaks. For example, with LibreOffice, just about
every test case I've ever run through it, "comes out wobbly".
(The first time, I was getting OpenGL-inspired crashes.)
When I ran ImageMagick (with its OpenMP library enabled
by default), it locked up the computer when multiple
threads tried to Malloc() memory buffers at the
same instant in time. I wasn't expecting that one
either. Microsoft fixed that.

There really isn't a lot of well tested code sitting
out there in 2018. When you say you "haven't found issues",
you should fully expect to find an issue in the next
24 hours. Nothing is really tested all that well.
Not programs with 600,000 source files in their
tarball. Those aren't tested all that well. There are
test benches inside, to test for "correctness" at the
module level, but not overall usability.

That's why Firefox has an entire telemetry system
to report memory usage and crashes. It allows
the developers to determine "20.4% of users saw
Firefox 52ESR to crash on this line of code". They
know this, because they have a subsystem reporting it.
Microsoft on the other hand, has its own CEIP. The
difference, is the Firefox capability, reports directly
to Mozilla, while CEIP, a developer has to log into
Microsoft, to get the info collected via CEIP. Mozilla
went to the trouble of writing that subsystem that
way, just so Microsoft would not get a copy of the data.

Microsoft will ruin that Chromium browser, but because
it will no longer look like Chromium, nobody will be
affected by the outcome. The Chromium browser they
start with will be Win32, but it won't stay that
way for long. I'd prefer for them to *deliver* the
browser first, and then happen to casually mention
what it's based on, rather than leaking some announcement
building up expectations, only to receive a letdown
when it comes out. One of the first things it'll
need, will be a telemetry transplant ("vortex").

Paul