| 1 |
On 170421-07:04+0100, Peter Humphrey wrote: |
| 2 |
> On Thursday 20 Apr 2017 17:59:13 Walter Dnes wrote: |
| 3 |
> > On Tue, Apr 18, 2017 at 10:06:12AM +0100, Peter Humphrey wrote |
| 4 |
> > |
| 5 |
> > > On Monday 17 Apr 2017 13:05:59 Walter Dnes wrote: |
| 6 |
> > > > 2) USE="-system-libs" is recommended. Yes, this does make the binary |
| 7 |
> > > > slightly larger. But it avoids problems where changing API/ABI in the |
| 8 |
> > > > system lib causes subtle problems for the program. It may also reduce |
| 9 |
> > > > dependancies pulled in, depending on what your other programs have |
| 10 |
> > > > pulled in. |
| 11 |
> > > |
| 12 |
> > > I don't see any of those USE flags. This is from the octopus overlay: |
| 13 |
https://github.com/Bfgeshka/octopus |
| 14 |
where it sticks out (very much, for me, because I actively decrypt |
| 15 |
rather often): |
| 16 |
palemoon with system-nss support |
| 17 |
|
| 18 |
which could be a good thing. But I'm not sure it would really be |
| 19 |
necessary. palemoon-overlay with stock palemoon: |
| 20 |
https://github.com/deuiore/palemoon-overlay |
| 21 |
, which packages its own nss, not system's nss, and it hasn't shown to |
| 22 |
be deficient, in my experience (but I'm not very advanced). |
| 23 |
|
| 24 |
I'll gladly read more reviews if anyone is willing to share, about |
| 25 |
octopus vs palemoon-overlay (which differently compile Pale Moon). |
| 26 |
|
| 27 |
> > > |
| 28 |
> > > [ebuild R ~] www-client/palemoon-27.2.1::octopus USE="alsa dbus |
| 29 |
> > > ffmpeg gtk2 official-branding optimize printing speech spell wave webm |
| 30 |
> > > -devtools -gtk3 -jemalloc -necko-wifi -pulseaudio -shared-js |
| 31 |
> > > -strip-binaries -system-cairo -system-compress -system-images |
| 32 |
> > > -system-libevent -system-pixman -system-spell -system-sqlite -system-vpx |
| 33 |
> > > -valgrind -webrtc" 0 KiB |
| 34 |
> > |
| 35 |
> > That ebuild does things differently. The other palemoon ebuild was |
| 36 |
> > all-or-nothing for a bunch of system libs. The octopus ebuild is |
| 37 |
> > granular, allowing separate choices for system-cairo system-compress |
| 38 |
> > system-images system-libevent system-pixman system-spell system-sqlite |
| 39 |
> > and system-vpx. |
| 40 |
> |
| 41 |
> As I'm now on sys-devel/gcc-5.4.0-r3 I'll have to assess what to do next. |
| 42 |
> That will be after I've finished sorting out another problem I have. |
| 43 |
|
| 44 |
That's just fine. Take your time. I'll also be interested to know. |
| 45 |
Maybe I get an answer to my questions (such as the one above) without |
| 46 |
much (more) investigating myself about it (I myself often get lost in |
| 47 |
the amount of learning to get the right things done). |
| 48 |
|
| 49 |
Just in case, you can have more then one gcc, i.e. you can have both 4.x |
| 50 |
and 5.x gcc, like: |
| 51 |
|
| 52 |
# equery l gcc |
| 53 |
* Searching for gcc ... |
| 54 |
[I-O] [ ] cross-arm-unknown-linux-gnueabi/gcc-5.4.0-r3:5.4.0 |
| 55 |
[IP-] [ ] sys-devel/gcc-4.9.4:4.9.4 |
| 56 |
[IP-] [ ] sys-devel/gcc-5.4.0-r3:5.4.0 |
| 57 |
# |
| 58 |
if you set: |
| 59 |
|
| 60 |
# grep multislot /etc/portage/package.use/package.use.file |
| 61 |
sys-devel/gcc multislot fortran |
| 62 |
# |
| 63 |
|
| 64 |
[if you set] the "multislot" use flag. I previously followed the |
| 65 |
recommandation to build with gcc 4-x. I changed later. |
| 66 |
|
| 67 |
Regards! |
| 68 |
-- |
| 69 |
Miroslav Rovis |
| 70 |
Zagreb, Croatia |
| 71 |
https://www.CroatiaFidelis.hr |
Archives