[blfs-support] Firefox issues

Alan Feuerbacher alanf00 at comcast.net
Fri Sep 26 22:05:21 PDT 2014


On 9/26/2014 9:45 PM, Ken Moffat wrote:
> On Fri, Sep 26, 2014 at 07:57:10PM -0500, Bruce Dubbs wrote:
>> I'm starting a new thread because that was what should have been done
>> several messages earlier.  It really isn't realted to "LFS and BLFS Version
>> 7.6 is released:.
> ...
>   You seem to be saying that running make with -j1 worked, and
> therefore you cannot know where it failed.

What I'm saying is that uncommenting the mozconfig line with -j1 
compiled ok, whereas keeping it commented failed. I'm not competent to 
figure out why.

> Is this i686 or x86_64 ?

x86_64

> You said you are using systemd, but I cannot think of anything there
> that would ... Oh, are you using pulse ?

Yes.

> That is the one thing I do
> not build - worked for me when I was testing gnome a year or two
> back, but when I tried before that, I was totally unable to make it
> work (that was in icewm, not a desktop environment, I sort of assume
> Lennart expected a lot of daemons to be running).

All of this is completely experimental for me, and I'm doing it as a 
learning experience, so I have a long way to go to get sound stuff working.

>> For me it happened with BOTH 32.0.1 and 32.0.3.
>
>   I'm not sure how many times I have installed 32.0.1, but I am
> fairly sure that I built it with -j4 on x86_64, and with -j3 on the
> i686 system which only has two cores [ and then built it again on
> x86_64 with -j1 to time it, after I knew it worked ].
>
>   The original error on i686 was during the install (inluding DESTDIR
> installs), and was a Python problem which gave a backtrace of Python
> modules.
>
>>>> I also installed the full development environment. What do you get when
>> you
>>>> install that? Where can I learn more about this?
>>>>
>>>    No idea. It does not ring any bells for me.
>>
>> Ok.
>>
>> But *someone* with BLFS must know what this means -- otherwise he would not
>> have put some verbiage about installing the full dev environment in the BLFS
>> book. Who might that be?
>>
>
>   It might be a lingering reference to when xulrunner was in the
> book.

Actually I just found (see my earlier email replying to Bruce) that the 
statement is in the 7.6 systemd book, but not in the non-systemd book. 
It's not in any of the earlier books in the BLFS archives, either.

> Back in the day, it looked as if xulrunner would be the basis
> for a quicker build of both firefox and thunderbird, but in practice
> they were hardly ever on the same version, and everything linked
> against it had to be recompiled for each new version.  At one time,
> some things such as icedtea and gnash needed xulrunner to get some
> headers.  Now, I think NPAPI-SDK has replaced that.

I don't fully understand all that, but I remember that when I compiled 
firefox and thunderbird last November, the process was more painful than 
this time around. So, good job, guys!

Alan



More information about the blfs-support mailing list