j-g-
-
Posts
80 -
Joined
-
Last visited
-
Days Won
18
Content Type
Profiles
Forums
Blogs
Posts posted by j-g-
-
-
The flamewar was already started, and the tread was deviated(I even read a bit of a wayland rant) from the intent of the OP many posts before mine, contrary to your experience I've seen many times help given on the gentoo mailing list to people that want to try it, and lots of patience, It's a shame you didn't asked the right people, but if you go anywhere with the argument 'Theres no documentation' when there is and its even local, it's likely the one who took the effort to read and grasp the manuals(and don't say most manuals are easy, look at bash), might get unfriendly, because you don't want to put any effort, remember asking for help in the open source it's taking someone time for free, so an extra effort is required, to show appreciation of the time given to you.That's one way to start a flame war.
My objection to systemd is based only on the negative experiences I have encountered. And asking for help with systemd and getting a reply similar to quote above doesn't help. This is why I came to funtoo. I was hopeful these kinds of discussions wouldn't occur here.
-
I've mentioned this here a few times, I use systemd with Funtoo(Or should I say NotFuntoo since drobbins says Funtoo is only Funtoo with OpenRC), I would have bothered to do what l33tlinuxh4x0r did, and post a guide and even an overlay, but I anticipated a flamewar like this, for no reason, anyway getting systemd to run isn't that hard, and any competent user would figure out the changes needed.
Many of you seem to know little about systemd, or can't read manuals(Undocumented?), and yet have big opinions on it, anyway this this has shown to be the tendency in among most of the systemd detractors. I understand devs don't want the burden of having to deal with bugs related to a software they dislike, but certainly It isn't attractive either to try to contribute for the few of us who like it, when any intent would turn into a flame war.
-
Did you tried using
g++ -std=c++11 ...
to compile the program with gcc?
-
Odd - I didn't post that at 8:59. Its not even that time yet. I posted that about 3 AM. Is the site on GMT?
I'll end up re-engineering the whole thing from the ground up. I think it needs to run Funtoo completely just based on the general principle. Not sure if I'll do a liveCD with installer, or just have it boot to an installer. Comments on that welcome.
As a live CD, if I make all the packages as generic as possible (all the same USE flags for all installs) then I can dynamically create the packages from what's on the liveCD or have an "install everything" option that just copies the whole live system without bothering with packages. At that point, it may be better to have multiple install CDs for desktop vs server, etc. But ... thats probably not what people want or expect. Its more convenient to have a larger CD that can install either, which means the liveCD itself will likely just be based on whatever becomes a minimal "desktop" install. It definately won't be like the Gentoo one in scope - way too many packages and no one is going to run all that stuff on a DVD. It will be more like Sabayon's in scope (but without systemd). In fact, what I really want is Sabayon with no systemd, a "server" option, and install GCC by default.
Why don't you take a look at how sabayon is put together, and maybe even use entropy to manage your binary packages, It has the advantage of being an already known and tested suite, You might be able to use a lot of their stuff for building your project.
PD: I have never used sabayon, I know some people who have, and were pleased with it, so maybe their tools are good.
-
cgdisk /dev/sda first partition- 500M 8300 second partition- 250M EF02 third partition- 2G 8200 forth partition- 20G 8300 fith partition 40G 8300 mkfs.vfat /dev/sda1 mkdosfs -n efi-boot /dev/sda2This, the efi-boot partition MUST be the first. you could try backup sda1 and sda2, and just make those again in the correct order and copy the files back again.
-
See eclean-dist --help , it has many useful options.
-
I would do it in two steps, base system, and then @worldThanks for the info, but I don't want a tar of an existing system as a huge tarball. Although I suppose that I could simply copy the entire system, then re-emerge it all with the new flags to force the rebuild. That may be the better way to do it.
Would you happen to know the full list of emerge flags for "rebuild my whole system, every last package, because I change USE flags and CFLAGS?" command line would be. I'm thinking once I get it to my liking I'll build a set of binaries and then find some duct-tape and bailing wire to graft an installer and a liveCD/DVD into the mix.
emerge -avNe --with-bdeps=y @system
emerge -avNe --with-bdeps=y @world
-N for new USE flags, -e for empty tree:
"-emptytree (-e)
Reinstalls target atoms and their entire deep dependency tree,
as though no packages are currently installed. You should run
this with --pretend first to make sure the result is what you
expect.
"
WARNING: You might bump into some bugs when doing a full rebuild, or compilation might strangely stop, the only time I have done this, it took 4 'emerge --resume' to get it done.(But I have limited ram and not much cores, I blame it on that)
PD: you can migrate your current system to another partition like you would take a VM and run it bare-metal, just make sure the boot process is correct.(fstab and bootloader configuration mainly)
-
I can't test. /var/portage doesn't exist on my funtoo installs so I can't even find /var/portage/tree/funtoo/profiles/package.mask/funtoo-gnome.
Those instructions are from an installation with PORTDIR="/var/portage/tree/funtoo/", the default is PORTDIR="/usr/portage". see the make.conf(5) manual.
-
RedHat has paid Poettering to develop a lot of code and then given it away for free. So what? That doesn't exempt him from criticism for his bad ideas and behavior. Poettering hasn't developed as much code as Linus but Poettering doesn't seem to feel that puts Torvalds beyond criticism.
Most people haven't chosen systemd. They chose a distro. Poettering, his funders, and supporters were successful at coercing most major distros into adopting systemd through a combination of political manoeuvring and threats that anyone who didn't fall in line would be rendered "irrelevant" as systemd consumed every project it touched and became a dependency for everything.
The two distros that put effort on systemd initially OpenSUSE and Fedora(Red Hat), aren't bigger in manpower, than the combination of Arch. Debian(and derivates), Gentoo, Canonical, and Slackware, developer comunities, I don't see how being majority 4 years ago distros that didn't used systemd, were politically forced to use it, the man power was there, if they would have wanted the organization of a better and moder init system would have happened. The political argument doesn't make practical sense.
-
No one is forcing anything, you chose Funtoo and you don't have to use it, the thing is, code does more than rant threads like this, you say Lennart is publicly a prick? Yeah that prick who has written more free software than anyone on this thread and put it available for free. and guess what? people chose to use it.I have to say that although he's arrogant and publicly a prick (don't know him personally), the work he's doing would be considered a contribution if it weren't being forced down our throats.
He's a paid employee isn't he? If he is then he should start acting like it.
He's indeed a paid employee who works to make what I think he likes the most: write free software. I don't know, but I don't think any if you given the chance(and being ~25) would turn down a Red Hat job offer, and if you did, clearly you would be very stupid, anyway, working for someone doesn't take away your freedom of speech(Sorry if you don't have that in your country), you can post whatever you want in your personal social media, and Lennart most of the time post technical and interesting stuff in his blog, unlike the insightful analogies between filming porn and developing free software I've read in this forum[1].
It's true saying today 'if you don't like systemd make your own', is stupid, given how much development systemd has had, but, the 'Rethinking PID 1' post[2] that was published in the beginnings of systemd, and describes the vision of the development, was four years ago, and none of the people that has been ranting since then, did anything more than that, just rant. And many others, those who knew what is useful to change, wrote code, but they did it for systemd(also upstart and OpenRC, but clearly they where less).
Anyway, This post is getting too long, So...
:wq
[1] http://0pointer.de/blog/projects/systemd.html
[2] http://forums.funtoo.org/index.php?/topic/111-gnome-312-is-here/?p=571
PD: Code says and does more than prose when it comes to software.
-
So far my experience has been, you can use /usr/local/portage, or an overlay of your own, I would recommend the overlay, since you can make changes to it as normal user. there's an skeleton overlay uploaded for funtoo in github, you can start from that.
Make the directory structure for the package you wan to test in your overlay dir:
$ mkdir -p ${OVERLAY_DIR}/cat/pkg/ # Edit your ebuild: $ $EDITOR ${OVERLAY_DIR}/cat/pkg/pkg-0.0.ebuild # generate a Manifest $ ebuild ${OVERLAY_DIR}/cat/pkg/pkg-0.0.ebuild manifestAdd your overlay via making a overlay.xml file in /etc/layman/overlays/ (this is documented in the layman manual) or adding to your make.confPORTDIR_OVERLAY="${PORTDIR_OVERLAY} $YOUR_OVERLAY_DIR"If you don't have the package installed already or are adding new dependencies you can do:$ sudo emerge --onlydeps '=cat/pkg-0.0'
and then you can test your ebuild step by step using the 'ebuild' command, I try before merging anything I edit, generating a package, this means all the build process is made except actually getting the package in your filesystem, you can inspect the tarball generated to see everything looks right and then merge.$ sudo ebuild ${OVERLAY_DIR}/cat/pkg/pkg-0.0.ebuild packageThat said you should read first and mainly refer to the gentoo dev documentation.[1] or 'emerge -av devmanual' -
I re-vectorized the original funtoo logo to make some wallpapers and a 3d model and posted the svg and other editable files(3d) with that logo some time ago in this tread:@Deklan do you have that as svg?
look at the bottom in the post, there's a URL to a tar.xz in dropbox.
-
I'm no one to say when but I think a few(or many) weeks. It will take time to the gentoo developers to bump all the gnome ebuilds, and then the patches made by dantrell for funtoo would have to be tested to work and/or upgraded. It was just released today, and the work is not a simple task.
Edit:
If you want to adventure to try to make it work with funtoo, this overlay seems to be ahead of all the others on the gnome bump:
-
set timeout=3 serial --unit=0 --speed=9600 terminal --timeout=5 serial console insmod part_gpt insmod ext2 set root=(hostdisk//dev/sdb,gpt1) search --no-floppy --fs-uuid --set 79d306ed-2578-4ac9-a199-6f61a6a7e64c if loadfont /grub/unifont.pf2; then set gfxmode=text insmod gfxterm insmod vbe terminal_output gfxterm fi set menu_color_normal=cyan/blue set menu_color_highlight=blue/cyan menuentry "Funtoo Linux Dracut - vmlinuz-3.16.2-hardened-r2" { insmod part_gpt insmod ext2 set root=(hostdisk//dev/sdb,gpt1) search --no-floppy --fs-uuid --set 79d306ed-2578-4ac9-a199-6f61a6a7e64c linux /vmlinuz-3.16.2-hardened-r2 rootfstype=ext4 luks enc_root=UUID=77f735b0-3d46-4fd7-9751-34df92d460d5 lvm rd.luks.key=/boot/luks-key.gpg root=/dev/mapper/vg1-root console=tty0 console=ttyS0,9600 rd.shell set gfxpayload=text } set default=0If you crafted that manually you missed adding the initrd command to load your dracut generated initramfs, and since you are using dracut you might want to explore more of its options, you can set your kernel command line in /etc/dracut.conf, and other ineresting things, this will make the upgrading kernel process more cleaner since you don't need to add the command line every time, also make sure you have ' hostonly="yes" ' in dracut.conf.
-
I think you missed the '-t' or '--target' flag, and that is the part that interprets 'armv7a-hardfloat-linux-gnueabi', and maybe crossdev just was trying to build another compiler for your local target, that's why so many conflicts regarding 'x86_64-pc-linux-gnu', and the strange message 'Package 'cross-armv7a-hardfloat-linux-gnueabi/gcc-4.8.2-r2' NOT merged' Is it even supposed to be a package?
Also I don't know if it would give some kind trouble to use the '-S' flag if you are running 'current', seems a contradiction, is that your case?
crossdev is tricky and not really well documented, If that isn't the problem, I can't help you further, I haven't played around with arm stuff that much, and currently I don't have a decent device for that. So I would look for help in irc, from people who uses it every day(Or even its developers)
-
You know well that there are with Fun-tasstic-too many possibilities to configure your system to the top :)
# cp /usr/share/zoneinfo/Europe/Rome/ /etc/localtime [/code]
There are a few problems with your aproach.
That file should be a symbolic link(upgrades), from funtoo wiki[1]:
/etc/localtime ... Your timezone, which will default to UTC if not set. This should be a symbolic link to something located under /usr/share/zoneinfo (e.g. /usr/share/zoneinfo/America/Montreal)
This should be done before compiling the kernel.
As for the upgrade timetable, I wrote a little row in /etc/local.d/
Then you may:
# Emerge -u ntp ntpclient
# Touch /etc/local.d/local.start
# chmod u+x /etc/local.d/local.start
Then "instruct this" ........ and. g.
# echo "sleep 90 && /etc/init.d/local.start &>/dev/null 2>&1 &" >> /etc/local.d/local.start
First, there's no '/etc/init.d/local.start' file so that would print and error but you are sending it to /dev/null, and It doesn't make sense to call something in /etc/init.d in a local script, read the README inside /etc/local.d/ for information on the correct use of the 'local' service.
Second, why didn't you just used the init script for ntp? I see you didn't even added ntp to your default runlevel services, that would be:
rc-update -v add ntp default
Check that's the acutal script name you want to add, in /etc/init.d, I don't have ntp here so didn't check that, If unsure ntp is the actual provider for the service, use ` equery f ntp | grep /etc/init.d ` to confirm
[1] http://www.funtoo.org/Funtoo_Linux_Installation#Configuring_your_system
-
Are you using crossdev for that?(tell what you typed, not only the output ) I've only once compiled an armv7a-hardfloat-gnueabi and it didn't use those paths warned for installing, I can't recall everything I did, but the documentation helped me
-
I looked at the repo you posted, and the last commit[1] is a defconfig for the beagle bone, I would use that, run make help in the directory where you have that, and read about using defconfigs.
[1] https://github.com/beagleboard/linux/blob/3.14/arch/arm/configs/bb.org_defconfig -
...
emerge --info net-misc/networkmanager
Returns too much info. Will have to output it to a file somehow and pastebin it. EDIT: see attached file. File did not attach. Please see http://pastebin.com/hnpudReX for this info. Thanks.I see at the bottom of that, the USE flags used for building, and there is '-wifi', wich means the NetworManager you compiled doesn't have wifi capabilities, since you seem to be new to portage, I'll explain you some stuff, but refer to 'man 5 portage' to learn more about portage's configuration files.
You can enable wireless for NetworkManager either adding 'wifi' to the USE variable in /etc/portage/make.conf,
... USE="... wifi " ...
or for using wifi only for networkmanager, add this line in /etc/portage/package.use (You can create this file, if you don't have it yet):
net-misc/networkmanager wifi
After any of those, remerge networkmanager with 'emerge -av net-misc/networkmanager' , this will show you the USE flags portage will use to build, check wifi is used, and let it build again, this should fix your problem.
Also, what flavor did you use, when selecting a profile with 'eselect profile'? I think if you use a workstation or desktop flavor, wifi USE flag is enabled by default, but I'm not sure about this, I'll check later.
-
If you are using a chroot, you don't need to configure the wifi network inside it, just copy your local /etc/resolv.conf to ${FUNTOO_CHROOT_DIR}/etc/resolv.conf, or if you know what that file is for make sure you have some dns servers in there.Thanks for the suggestion. I think the driver is installed and loaded but not sure.
# lsmod | grep wifi iwlwifi 87837 1 iwldvm cfg80211 412130 3 iwlwifi,mac80211,iwldvm
I did chroot from a live usb and emerge rfkill. Now after rebooting and running the addwifi command I get a different error:Error: Connection activation failed: Device not managed by NetworkManager or unavailable
rfkill list returned two entries, neither of which was soft or hard blocked.When in the chroot, just make sure you build your kernel modules needed for your card, if you build debian-sources kernel, you are more likely OK.
After you leave the chroot, and reboot, use only one network manager, wicd , NetworkManager, etc.., if you will use NetworkManager and you don't have a GUI you can use nmcli, I use it that way, a quick list of commands for managing wifi trough nmcli:
nmcli c s # connection status nmcli d # list devices nmcli d w l # list wifi connections available (SSID,BSSID, etc) nmcli d w c [SSID] password <password> # connect to a network password can be eg. WPA2 or WEP
if you need more help after rebooting, save and paste here after a log of:ifconfig -a # or better ip a # you will have to install iproute2 in the chroot for that cat /proc/net/wireless cat /proc/net/dev emerge --info [network_manager_installed]
-
Post the complete outputs, the above post really says nothing.
Edit: I see you have already been given directions on the bugtracker, so there's nothing else to say.
-
Edit: I wrote a rather lengthy opinion on System U, and later realized it derails from the topic, and I care more about the topic so I put in a spoiler.
The idea of System U is good but just the concept, the project is lacking many things, one of them is realism and to know what they are doing, I see it failing to be just another distro with a maintenance nightmare of forks, maybe it will be for people with ADHD, since quoting its 'founder'(I don't know the proper term for people that do a kickstart)
And you know it will be quoting: 'Fully Customizable'...' From changing the cursor theme to adding an entire dock, everything between, and beyond, we make it possible for everyone to have an ideal system with seamless functionality.''I'm Andrew Bernstein, 17-year-old high-school student from Short Hills, NJ. I also have ADHD and I'm not a very flexible person. I really just don't like change.'
<sarcasm> Woahh!!! do they mean I will be able to change how my desktop looks, awesome! In the 6+ distros I've ever tried I was never able to do that </sarcasm> (I even did it with windows google bblean)
We funtoo/gentoo users do know what is 'Fully Customizable', really, I think he just need a psychologist to assist him dealing with his ADHD, you know, this world as a whole will always change. I really don't know why he doesn't just use debian-stable and contributes there, sounds like his perfect distro. Please don't think I dismiss him because he's a kid, I'm not old either, I know other people is involved and there are geniuses in this world.
Lets contribute to the discussion, and share some ideas about that thing people names 'User experience' and I agree with you on the concept of surveys, or something like that, but ultimately, one can't generalize the concept of 'users', there are many different types of users, and a person uses a computer as means to do something, not because of using the computer(but I bet many of us *ntoo users do exactly that), so the 'tools' used that achieve the main goal of the user, are the essential part as I see it, so efforts to make well designed tool-chains for different use cases, I think would be a better approach(eg. office, graphics edition, CAD, Business management), and take someone who isn't a power-user, and I think these are the majority, and put a launcher in his desktop for his 'work' tools, a web-brower, and a file manager for him to copy files to USB pen-drives, I bet you will have a happy user most of the time. In this regard I see the fedora spins as a really good Idea, I also always see many people saying windows is easier to use for 'end users', but really how many people you know that actually 'know how to use it', and don't just click any OK button that gets in their way, and yes, they break things this ways sometimes.
So, it is my opinion use cases, a profile of a set of regular users of a specific tool-chain, surveys made to people that actually uses the software, would be better directions for developers. I think many times, FOSS developers just scratch their own itches, thus making not really well planned and studied changes, that might be a headache for a user who might have never needed the new change(feature), and I know many times for a developer, to make that 'well planned' change requires time, which many FOSS contributors don't have because they do these new features in their free time, I think fund-raising to hire paid developers to improve/make/integrate tool-chains *UPSTREAM*(I really want to emphasize that) wold make the FOSS world much better than more distro fragmentation.
I have written too much, for this post, so I leave sharing a video-conference on User Experience, that taught me some very useful things I will apply as a developer, http://youtu.be/ 7OSkB4BCx00 (remove the space).
PD: Excuse my English, I'm not a native speaker.
-
Do it the other way around, don't set static-libs gloabally, you will have more troubles, instead set it per package in package.use.
-
I think the first post is a pretended emerge, and ~amd64 is set globally on his make.conf, so no, it wont help.what is your emerge -apv xbmc output? Correct me if I'm wrong anybody but from my understanding it looks like the amd64 is masked since the multilib use flag is in conflict with a 32bit ebuild. Could a ~amd64 keyword in package.accept.keyword(for nvidia-cg-toolkit-3.1.0013-r3) fix this issue?
What OP could try, is unmasking the 'multilib' useflag, if his system actually supports multilib, and try emerge again, but I don't think it was masked arbitrarily, so it's likely the build fails or runtime problems might arise.
echo '-multilib' >> /etc/portage/package.use.mask
It might be also needed to enable the multilib use flag in package.use for media-gfx/nvidia-cg-toolkit. But I warn again, it's likely to be broken somewhere and that caused the developers to do the use flag masking (and if you look carefully, this is coming from gentoo, so I would search the gentoo bugtracker to see if I found out, why multilib is masked)
Systemd
in General Discussion
Posted
This is actually a strong point of systemd, remember you can have a shell-less boot, and run in your machine just to find a few examples:
$ egrep -rn ' sed | awk | grep ' /etc/init.d/
And the fast boot is rather than a intended feature, a consequence of running less programs at boot to get services going. about dbus, I haven't dealt with it directly, so I can't say much. but supposedly kdbus is an intent to take the good of it and ditch the bad.
I don't think companies are all good intentions, but I don't really see much problem, when they:
- Want to hire open source developers, this is a really good thing if you are a developer, doing what you like and get paid for it.
-The code produced is being open sourced.
The fact that most companies are unlike this, is a shame, everyone would have so much better software if more talented developers would get paid to improve open source tools, and not so much home-grown stuff, totally closed.