Re: Re: Connection at browser first run or update

[chippy]

Hi,
> En 07/06/24 06:40, chippy escribió:
> > Hi all,
> > 
> > I noticed with the last builds, that the browser will open gnu.org
> > on
> > first run.
> 
> That sounds like a feature Mozilla implemented and got inherited
> downstream.
> I've seen that happening on Thunderbird/Icedove opening the donation 
> page on first run.
> 
I came to the same conclusion after my email.

> I suppose it would be possible to unset the link and leave it blank
> or 
> point it to 127.0.0.1
> 
> > 
> > It will also transparently ping gnu.org when opening about:addons.
> > I did not investigate further...
> > 
> > Why is that? I (and a lot of users) did not expect a similar
> > behavior...
> Well, the transparent pinging to gnu.org sounds like a global sed 
> replacement removing some url match and pointing back to gnu.org,
> again 
> this could be found and set to point nowhere at all.
> > 
> > This doesn't seem different from the first-run telemetry ping to
> > Mozilla by Firefox.
> > 
> > *** What is happening? ***
> > 
> So far, I've seen that pinging is easily fixed by looking where it's 
> pointing to and replacing or removing it.
> The opening pages on first-run, it would take longer to investigate, 
> maybe a value could prevent that feature to be executed or maybe it
> is 
> necessary to manually remove such code.
> 
I'm actually having trouble reproducing it with 115.12. Removing Icecat
and the profile folder in the home, then reinstalling it, does not
trigger the thing again.

> Either way I think you are looking it on the wrong light, I genuinely
> don't thing gnu.org has any intention to track pings 
I believe so too.
> (nor the 
> infrastructure in place to do so, like Mozilla which in fact does) so
> *it's different* maybe not ideal but on the end result there is no 
> tracking. It is highly possible you are looking the result of early 
> limited cleaning routines.
> 
> Now that you are using a newer / more advance approach you have
> better 
> information that previous developers did.
> 
> I would document it, send a patch, propose a new routine, implement 
> locally (so my own copy have those changes), test and report back. So
> the codebase gets improves for everyone (me included).
> 
I'll look into it again, but as I said I couldn't reproduce. The idea
is to replace all the gnu.org occurrences with some counter so to have
them all different and being able to tell one from another in one go
(read in one build). 

> > 
> > Thanks in advance,
> > Chippy
> >  Thank you.
> Regards.
> 

Thanks,
Chippy