Re: [gentoo-user] How the HAL are you supposed to use these files?
On 2/8/2010 5:20 PM, Alan Mackenzie wrote: . Am I the only person that finds this semantic gibberish? Is there any explanation somewhere of what a "policy" aka "device rule" is? What is the semantic significance of a "device rule"? What does it mean, to "rule a device", or what sort of restrictions are being placed on this device? Given that one might desire a "basic working keyboard/mouse combination", what is the chain of reasoning that ends up selecting the file called "10-input-policy.fdi" from all the other ones? This file is an inpenetrable stanza of uncommented XML. Are its verbs documented somewhere? What do "" and"" mean, for example? The way HAL works, in a nutshell, is to scan your system for every known piece of hardware it can find, and stores the information in a tree-like database of key/value pairs. Software can then query this database for information about whatever hardware you have. The information includes things like the bus location of the hardware, the manufacturer information, state information, and a lists of known capabilities like "keyboard", "mouse", "disk", etc. Device Rules are simply ways for the user to change values in the database after a device has been detected. The XML files work in two steps: 1. an existing node in the database, 2. or For example, the hplip printer/scanner drivers include a set of HAL rules that match HP devices by their PCI device information, then append "scanner" to their list of capabilities. Other software can then scan the HAL database for "all scanners" and find them. The synaptics touchpad driver (if you build it +hal) includes a set of HAL rules that overrides the standard 'mouse' rules to make your touchpad more useful. That bit about you having to do anything to make a "basic working" setup function, though, is wrong. On Gentoo, at least, everything you need for a working keyboard and mouse in X should be installed properly for you by default. You'd only need to mess with the rules if something didn't work. But, see below. Can this new-style fragmented XML configuration do anything that a good old-fashioned, human-readable and compact xorg.conf can't? If so, what? What am I missing here? HAL manages a *lot* more than just your X configuration. It's intended to be a complete hardware management layer, one that was able to keep pace with new hardware more quickly than the kernel could. If you run the HAL database dump utility "lshal" you'll see more information about your hardware than you could ever possible care to know. Please, somebody, tell me all this HAL stuff is straightforwardly explained in an easily accessible Gentoo document, so that I can hang my head in shame and apologise for the noise! ;-) Oddly enough, the most complete explanation of HAL I've ever found was on the Gentoo wiki and I think the page may be lost. It was never really documented that well, though there are a number of places you can find specific ways to do specific things (like using a touchpad) with HAL. The key point here, though, is that HAL is going away. Not because it was hard to configure, though -- because the code is an "unmaintainable mess" and because other software, like udev, duplicated much of its purpose. At this point, if it's not working for you out of the box, turn it back off and revert to the old style configuration file.
Re: [gentoo-user] How the HAL are you supposed to use these files?
On 2/9/2010 3:16 AM, Dale wrote: On Mon, 8 Feb 2010 21:17:08 -0500, Walter Dnes wrote: My solution to simplify Gentoo... waltd...@d531 ~ $ cat /etc/portage/package.mask sys-libs/pam sys-apps/dbus sys-apps/hal You'll have to do a manual depclean (very carefully) and revdep-rebuild, but it's worth the effort to purify your Gentoo system. Simpler than that, just add -hal to xorg stuff in package.use and then run emerge -uvDNa world. I'm not saying your way won't work but I think mine is easier. His way is also *way* more Luddite than yours. Note the 'pam' and 'dbus', two things basically standard (and very stable) on modern Linux desktop systems. --K
Re: [gentoo-user] How the HAL are you supposed to use these files?
On 2/10/2010 2:12 AM, Alan McKinnon wrote: On Wednesday 10 February 2010 03:29:50 Dale wrote: Well, actually, if hal would have worked I wouldn't have cared if it uses xorg.conf at all. That was the point of using hal. Thing is, I followed the howto and it didn't work. The fact that the config files are in xml only became a problem after hal locked me out of my GUI and required a hard shutdown. hal is a classic "Second System Effect" case But I thought we thrashed this to death a while ago and all agreed to never speak of this abomination again, while we await the Third System Effect aka DeviceKit? Last I heard DeviceKit was "deprecated before it even happened" and they're just going to move everything into udev. --Mike
Re: [gentoo-user] Re: Has semantic-desktop really become compulsatory for kmail?
On 2/11/2010 7:00 AM, Walter Dnes wrote: And if I really wanted a glitzy, bloated, slow-as-molasses, pointy-clicky-touchy-feely-oowee-GUI, I would've stayed with Windows, thank you. I started with Blckbox and am now on ICEWM. So, to summarize: * You don't like modern desktop environment design philosophy * You have no need for all of the convenience and productivity enhancements they provide * You are perfectly happy to use your previous window manager * Your previous window manager continues to work well for you and do everything you require from a window manager, and, * You just really like to complain about things. --Mike
Re: [gentoo-user] Re: Has semantic-desktop really become compulsatory for kmail?
On 2/12/2010 10:14 AM, Zeerak Waseem wrote: But then the question isn't whether there are a number of soft dependencies, but in the case of semantic-desktop whether -it- is a soft dependency. Like previously stated, I don't use kmail, nor do I intend to (I at least think I mentioned it). This is just my take on the matter of whether it is truly necessary, or even a good idea to have symantic-desktop as a hard dependency. No, it's not a soft dependency. Yes, it's a hard dependency. Yes, it's a good idea. Yes, it's necessary. No, no amount of pointless whining about "bloat" is going to change things. If you want to use KDE-4 applications, you use semantic desktop. If you don't want to use semantic desktop, you don't use KDE-4 applications. Yay for choice. --Mike
Re: [gentoo-user] How the HAL are you supposed to use these files?
On 2/15/2010 2:20 PM, Enrico Weigelt wrote: J. Roeleveld wrote: And *IF* some application is interested in the such information, why not just using the filesystem ? Because on flash-drives (Which are used in small devices and netbooks) you don't want every single status update to be written to the filesystem. And with minimal memory, I don't want to have a ram-disk gobbling up the memory I have. Why not simply using tmpfs ? Or an specific synthetic filesystem ? 9P makes this really easy, and network agnostic. I'm kinda stunned that your arguments against D-Bus seems to boil down to "just use 9p instead" given that plumber is a basic element of 9p and does essentially the same job D-Bus does. So you're just swapping one system-wide general-purpose IPC service out for another one?
Re: [gentoo-user] How the HAL are you supposed to use these files?
On 2/16/2010 3:23 AM, J. Roeleveld wrote: Netbook: 1GB of ram, with Linux, I can easily run all the software I want , without need of any swap. Can I do the same with 9P? Eg. will I be able to run all the software I use on my netbook without having to spent time on porting it all? Is also all the hardware supported in 9P? Linux supports all the hardware in my netbook. Unless the answer to this is a 100% yes, 9P is never going to be an option. Just for reference, 9p is not Plan 9, it's only the Plan 9 network protocol/distributed file system, which you can use on Linux with the appropriate file system modules. --Mike
Re: [gentoo-user] Re: KDE? Get me out of here!
On 2/24/2010 8:41 AM, Zeerak Mustafa Waseem wrote: > On Wed, Feb 24, 2010 at 03:38:09PM +0200, Nikos Chantziaras wrote: >> On 02/24/2010 04:27 AM, Kevin O'Gorman wrote: >>> I've been using KDE for a long time, for reasons that are no longer >>> important to me. I have remained out of pure inertia. >>> I use gnome happily at work, both on Fedora and Ubuntu. All I need from >>> any of them is a panel with some favorites, and a pager for multiple >>> desktops. >>> I spend most of my time in vim, in the C program and documentation >>> toolchains or in a browser. >>> >>> The reason I bring this up is that my account just froze on me from >>> running out of disk space. A little research showed that an >>> odd-sounding thing called nepomuk was using 7.2 G (SEVEN GIGS) in some >>> dotfiles. It turns out to be a KDE client - whatever that is. I've got >>> a lot of space here and there, but my /home partition was never near >>> full before. >> >> Put "-semantic-desktop" in your make.conf. emerge -auDN world. emerge >> -a --depclean. That should do it. >> >> > > Is that even possible? Won't a number of KDE apps demand the semantic-desktop > use flag set? > For KDE 4.4, +semantic-desktop is mandatory, though you can still turn off the services after installing them. Honestly, for what the OP appears to need out of a desktop environment, he'd be more than happy with Xfce or something and save a ton of disk space. --Mike
Re: [gentoo-user] Re: KDE? Get me out of here!
On 2/24/2010 11:59 AM, Nikos Chantziaras wrote: > How do you know what he needs? He probably wants KDE but without the > whole "Semantic Desktop" thingy. Well, mostly based on him telling us what he needs, and that he doesn't really "want" KDE: On 02/24/2010 04:27 AM, Kevin O'Gorman wrote: > I've been using KDE for a long time, for reasons that are no longer > important to me. I have remained out of pure inertia. > I use gnome happily at work, both on Fedora and Ubuntu. All I need from > any of them is a panel with some favorites, and a pager for multiple > desktops. it appears that all he needs is a panel and a pager, which any decent window manager will have. Therefore, if KDE is using up more resources than he feels warranted, perhaps it's time to switch. --Mike
Re: [gentoo-user] New openssh install message?
On 3/12/2010 2:07 PM, Mark Knecht wrote: > Hi, >I don't remember seeing this message on previous openssh updates: > Installing (1 of 1) net-misc/openssh-5.3_p1-r1 > * >>> SetUID: [chmod go-r] /usr/lib64/misc/ssh-keysign ... > [ ok ] > * Remember to merge your config files in /etc/ssh/ and then > * reload sshd: '/etc/init.d/sshd reload'. > >Is this a new message or have I just missed in the past? I don't > know anywhere else for ssh configuration files to exist except > /etc/ssh and I normally just do The message isn't new; every version of openssh in portage includes it as part of the post-install. As far as I can determine, it's just a reminder to restart ssh after merging your config changes e.g. with conf-update or whatever you use, so you can take advantage of whatever security bugfixes were present in the new version. --Mike
Re: [gentoo-user] Re: CUPS and hp OfficeJet Pro
On 3/18/2010 3:55 PM, Dale wrote: > I think avahi is a KDE thing. I don't really know what zeroconf is. If > I recall correctly, some package said it had to have that so I turned it > on. No clue what it is even after looking up the definition with euse. > May as well be Greek. ;-) Zeroconf is a set of technologies that are supposed to generate a fully working IP network with no user or operator intervention. It includes three basic parts: link-local network config (e.g. IPv4LL), distribution hostname resolution (multicast DNS), and automatic service and device discovery (DNS service discovery). Used in the context of applications or services, you're usually talking specifically about the autodiscovery portion, which allows applications to find services and network devices automatically. It was primarily invented at Apple, who developed mDNS and DNS-SD, and is built into OS X as Bonjour. Avahi is just a free-software implementation of Bonjour (which was originally under the not-entirely-free Apple Public License), and from what I've read has practically overtaken Bonjour in terms of performance and features. Back onto the topic at hand: emerging cups with +zeroconf allows it to respond to service discovery requests. By default CUPS uses mDNSResponder, which is Apple's implementation; with +avahi is uses avahi instead. This means any Mac on your network will automatically see CUPS printers, as will any Linux client with avahi properly installed. Windows machines with iTunes or Safari installed probably have Bonjour as well, so they'd also benefit. On a side-note: CUPS 1.4 stopped supporting Avahi and only supports Apple's implementation, so the Gentoo devs have disabled zeroconf support completely until CUPS 1.5 (or whatever) brings back native Avahi support. --Mike
Re: [gentoo-user] Re: help
On 3/22/2010 3:40 PM, Mick wrote: > On Monday 22 March 2010 19:21:26 KH wrote: >> Am 22.03.2010 20:17, schrieb Mick: >>> TBH, I wouldn't pay money for it but as many OEM impose a MSWindows tax >>> on all of us I had no other option if I wanted to buy this particular >>> laptop. >> >> You can refuse the license agreement and give windows back. If you are >> lucky, the vendor will give you some money back. > > Getting money back from Dell?!! It'll be like squeezing blood out of a > stone. ;-) > > Seriously though, I've asked them to take it off for me and they said that > this is "not an option"! Particularly annoying is the fact that Dell claims to be "Linux friendly". Which is apparently shorthand for: "Sure, we'll happily sell you one of three crappy laptop models with Ubuntu pre-installed, at a slight discount, while bombarding you with 'Dell Recommends Windows' ads while you shop. What's that? You want a desktop machine with Linux? Are you insane?"
Re: [gentoo-user] KDE 4 - error when closing window
On 3/22/2010 5:46 PM, Dale wrote: > Alan McKinnon wrote: >> On Monday 22 March 2010 21:19:16 Mick wrote: >>> I thought that KDE4 uses KDEWM not KWin ... ? >>> >> Dunno where you got that name. There's no such thing, and I don't >> think there >> ever has been: > > There is such a thing but it appears to be a variable not a executable. > I found several references to it in the KDE mailing lists. All related > to KDE4 by the way. It's not KDE4-specific, but you are correct that it's the name of an environment variable, not an actual KDE component. Setting KDEWM was the standard way of getting compiz instead of KWin for your window manager in KDE3. Now that KWin composites all by itself, not really necessary, but the variable still works as before. --Mike
Re: [gentoo-user] f-spot depends on mono?
On 3/25/2010 11:31 AM, Xi Shen wrote: > hi, > > when i try to emerge f-spot, i found that it pull in dev-lang/mon, and > all its dependencies. why f-spot needs mono? since when mono becomes > stable, and is used in real linux application? F-Spot is written in C#. C# is a CIL-only language, and requires the Mono runtime to execute. If you consider F-Spot a "real" Linux application, then clearly Mono is used by "real" Linux applications. The author of F-Spot, at least, considered it stable enough to use. If you don't consider Mono stable, you'll have to use something other than F-Spot. Unfortunately I don't know of an application for Linux that does everything that F-Spot does, except maybe Picasa from Google? --Mike
Re: [gentoo-user] Can't get a DVD to burn.
On 3/30/2010 1:00 PM, Dale wrote: > status: 0x2 (CHECK CONDITION) > Sense Bytes: 70 00 03 00 00 00 00 0A 00 00 00 00 73 03 00 00 > Sense Key: 0x3 Medium Error, Segment 0 > Sense Code: 0x73 Qual 0x03 (power calibration area error) Fru 0x0 > Sense flags: Blk 0 (not valid) This part is claiming that your burner was unable to figure out what the correct power level is for writing to the media. This is entirely controlled in hardware, so there is no legitimate reason why your version of cdrecord should have any effect. It is almost always caused by: * Bad media * Bad hardware * Insufficient power supply You've basically eliminated #1 by using the same media successfully, and #3 isn't the kind of thing that just crops up, so I suspect your recorder is starting to fail. (It could be as simple as a dirty lens). The only question is why older versions of cdrecord continue to work. Can you post the output from a successful run on the same ISO as this? It would be helpful to see if cdrecord is using different settings for the different versions. In particular, if it defaulted to a lower burn speed, or didn't enable BURNFREE by default, those may change the power requirements. --K
Re: [gentoo-user] Can't get a DVD to burn.
On 4/1/2010 4:38 PM, Paul Hartman wrote: > On Thu, Apr 1, 2010 at 10:02 AM, Joerg Schilling > wrote: >> Your media was either accepted after several tries by the drive for unknown >> reasons and is now usable in general again or you are observing a problem >> caused by "hald". Note that hald does not care about the CD/DVD/BD Writing >> process and interrupts it. This is why hald may cause any strange result. > > Would a HAL preprobe FDI like this prevent it from interfering with burning? > > > > > > true > > > > Bad enough you mentioned HAL in a conversation with Dale. Now there's XML involved. This thread is officially never going to end. --Mike
Re: [gentoo-user] Forcing X.org video driver without xorg.config
On 4/5/2010 7:37 AM, Black Napalm wrote: > Hello! > > I am trying to set up X.org and the Gentoo X Server configuration > HOWTO says that HAL should be used and that xorg.conf should only be > used as a last resort. However, when running X without a configuration > file, it autodetects the video card and tries to load the wrong > drivers, which I haven't installed, and therefore fails. > Is there a way to force X to use a specific driver without using a > configuration file? I am not really familiar with HAL, but it seems as > though none of the provided policies have anything to do with video > cards. 1. Dale, preemptive shush. 2. HAL is going away, so I'd suggest you just ignore everything the HOWTO says about it and use the previous method of defining everything in xorg.conf. 3. If you want to keep HAL: the HOWTO isn't really referring to your video hardware when it says not to use a configuration file. It's mostly talking about your input devices and other hardware. If X isn't auto-detecting your hardware properly, adding configuration sections to xorg.conf is still correct.
Re: [gentoo-user] emerge update gcc downgrade
On 4/6/2010 6:40 AM, Kraus Philipp wrote: > Hi, > > I run in a virtual machine a gentoo (~x86) system. I synced the portage > tree at the weekend an run emerge --update > > The update runs without errors, but emerge installed the gcc 4.3.4, > but on the system is the 4.4.3 installed > > [ebuild NS ] sys-devel/gcc-4.3.4 [4.4.3] USE="hardened mudflap nls > nptl openmp (-altivec) -bootstrap -build -doc (-fixed-point) -fortran > -gcj -gtk (-libffi) (-multilib) -multislot (-n32) (-n64) -nocxx -nopie > -objc -objc++ -objc-gc -test -vanilla" You have the +hardened USE flag set. Did you just recently switch to the hardened profile? If so, make sure you followed the upgrade guide for hardened profiles, which should have warned you what was going to happen here. In the base system, gcc-4.4 is still masked off for hardened users because it's not quite ready for production use. If you want to use gcc 4.4 and all it's hardened features, you need to do two things: 1. Add the hardened-development overlay, for example using layman: basement ~ # layman -a hardened-development 2. Unmask gcc-4.4: basement ~ # mkdir /etc/portage/package.unmask basement ~ # echo '=sys-devel/gcc-4.4*' >> /etc/portage/package.unmask/hardened After that, you should get gcc 4.4.3 back.
Re: [gentoo-user] no text in certain flash app
On 4/14/2010 11:38 PM, Tony Miller wrote: > Hi, I've noticed that in linux(in all browsers), I cannot see the text > on this page. > > http://musictheory.net/lessons/html/id10_en.html > > There is supposed to be different text each time you hit the 'next' > slide, but I see none. Kinda sucks because this site is very useful. > > I have www-plugins/adobe-flash version 10.0.45.2 installed. Are you using Adobe's 64-bit Flash? If so, this is apparently a very common problem. I've seen the same effect on numerous Flash applets on Facebook, for example; thus far I've been unable to figure out which Flash element is being used that doesn't work, because it's not consistent even within a single applet. --Mike
Re: [gentoo-user] no text in certain flash app
On 4/15/2010 12:28 AM, Dale wrote: > Paul Hartman wrote: >> On Wed, Apr 14, 2010 at 10:38 PM, Tony Miller >> wrote: >> >>> Hi, I've noticed that in linux(in all browsers), I cannot see the text >>> on this page. >>> >>> http://musictheory.net/lessons/html/id10_en.html >>> >>> There is supposed to be different text each time you hit the 'next' >>> slide, but I see none. Kinda sucks because this site is very useful. >>> >>> I have www-plugins/adobe-flash version 10.0.45.2 installed. >>> >> Works for me using 64-bit adobe-flash-10.0.45.2-r1 in every browser I >> tried. I see things like: >> >> • Clefs assign individual notes to certain lines or spaces. >> • Two clefs are normally used: The Treble and Bass clefs. >> • The first clef we will discuss is the Treble Clef (also called the G >> Clef). >> • The staff line which the clef wraps around (shown above in red) is >> known as G. Any note placed on this line becomes G. >> • The note on the space above G is A. (Remember, there is not an "H" >> note). > Since you know more about music than me, I may be seeing the same thing > you described. I couldn't remember what some of the things were > called. I only took piano lessons when I was a kid. Based on you > description, it may be working the same here. No... he's actually quoting the text that should be on the page, not describing the musical elements. It's a series of bullet points describing in words what is displayed above it. So clearly you're not seeing what he is :) On the plus side: Paul Hartman is a genius; installing 'corefonts' has fixed a bunch of previously-broken Facebook games. But sadly, not this musictheory page, so I must still be missing the fonts. Paul, Is there a way to tell which font the Flash animation is trying to use? --Mike
Re: [gentoo-user] no text in certain flash app
On 4/15/2010 2:25 PM, Paul Hartman wrote: > I would also try media-fonts/freefont-ttf if you don't have that > installed. I decompiled the flash on that page and it's just using the > generic "_sans" font, which should let it use any sans-serif font on > your system. Maybe it can only use TTF fonts? I'm not really sure. Wierd. I have all the same fonts you do, plus a few extra, and still no text in that flash. Now that I know it's a font issue I might mess around with my fonts and see if I can figure out what it's problem is. --Mike
Re: [gentoo-user] How many ways are there for a user to increase their permissions?
On 4/18/2010 12:29 AM, Jonathan wrote: On Sun, 18 Apr 2010 00:46:25 +0100 David W Noon wrote: If any Joe Schmoe could imbue a program with capabilities, this might be true. But that's not the way the system works. Sorry, I think i'm missing your point. Only root can run the setcap program to add capabilities to a program, at least on a normal, UNIX-style security system. On a role-based security system, even root might not be permitted to do this. If I had the root password to own system(which I do...) and I wanted Wine to uses IPX without running as root. I would set "setcap cap_net_raw=ep /usr/bin/wine" as root. Then I could run Wine as my normal user. No one in there right mind would run Wine as root. If you did you may as well use Windows. You say "no one in their right mind" would run Wine as root. But if you did not have capabilities support available, and wanted Wine to use IPX, then you wouldn't have any other choice but to run Wine as root. By using capabilities, you aren't increasing Wines permissions, you are decreasing the permissions needed to support IPX. Trying to compare Wine without IPX to Wine with CAP_NET_RAW isn't a fair comparison, as the two don't have the same feature set and thus clearly don't have the same security needs. --Mike
Re: [gentoo-user] installation aborted due to poor programming practices
On 4/19/2010 10:19 AM, Arttu V. wrote:
> On 4/19/10, Helmut Jarausch wrote:
>> On 19 Apr, Neil Bothwick wrote:
>>> On Mon, 19 Apr 2010 12:47:17 +0200 (CEST), Helmut Jarausch wrote:
>>>
the installation of a package (unofficial gimp-gap) is 'aborted due to
poor programming practices' probably since there have been lots of
dereferencing type-punned pointers and similar warnings.
How can I forced portage to install it anyway?
>>>
>>> Possibly FEATURES="-strict" emerge --opts blah
>>>
>>
>> Unfortunately this didn't help.
>
> I wonder if strict and stricter FEATURES are somehow hardwired to each
> other inside portage code, or whether they're considered separate
> entities?
>
> Anyway, you could try FEATURES="-strict -stricter" to see which way it
> works. If that won't work then the next step could be to play around
> with the QA_STRICT_* variables mentioned in make.conf man page
> (haven't tried, so don't know if they'll help either).
>
The code that's causing the abort doesn't seem to have any options,
features, conditions, etc. that you can use to disable it. If it finds
any misused type pointers, and you're on a 64-bit arch, it aborts.
Note that the errors it's checking for aren't your normal sloppy pointer
usages. It's specifically looking for functions that are implicitly
declared as returning int, then later used as if they returned (void *).
That kind of thing can only happen if the developer just flat out
forgot to include the prototypes ("poor programming practice" doesn't
even come close) and are almost guaranteed to cause something to crash
eventually on a 64-bit arch.
The only thing I can see that may let you install this overtop of
portage's complaints is to unset PORTAGE_LOG_FILE so the gcc warnings
have nowhere to go and portage won't find them.
See also: http://bugs.gentoo.org/40023
Re: [gentoo-user] Can't get a DVD to burn.
On 4/20/2010 11:01 AM, Paul Hartman wrote: > On Mon, Apr 19, 2010 at 8:44 PM, Dale wrote: >> /usr/bin/growisofs -Z /dev/hdd=/dev/fd/0 > And isn't /dev/fd/0 a floppy device? Are you burning from a floppy to > an IDE drive? /dev/fd0 is a floppy. /dev/fd/0 is file descriptor 0 for the current process -- stdin.
Re: [gentoo-user] Re: xorg-server upgrade
On 4/21/2010 1:41 AM, Graham Murray wrote: > walt writes: > >> That was true in the past, but no longer. The recent release of xorg 1.8 >> specifically says that hal will not be supported in any future xorg versions, >> so we should all start looking beyond hal. Don't spend a lot of effort now >> learning about hal because it's on the way out. (Not many people mourning >> it's impending demise, apparently.) > > And if you use udev then you need an (at least minimal) xorg.conf. > With the new modular xorg.conf.d setup, you really don't. I have no xorg.conf and the only thing I'm missing is the mouse wheel emulation on my touchpad. The xorg-server package includes configuration segments that handle the basic input devices for you, and the server itself is pretty good as figuring out your output devices. --Mike
Re: [gentoo-user] Updates = slow firefox
On 4/21/2010 3:07 PM, Grant wrote: > Could this be the problem? > > # grep ^\(EE /var/log/Xorg.0.log > (EE) Failed to load module "vesa" (module does not exist, 0) > (EE) Failed to load module "fbdev" (module does not exist, 0) > (EE) intel(0): [drm] Failed to open DRM device for : No such file or directory > (EE) intel(0): Failed to become DRM master. > (EE) intel(0): Failed to initialize kernel memory manager The first two errors are fine; Xorg defaults to trying vesa and fbdev as display drivers and you just don't have them. The last three are your problem. The intel video driver is unable to properly access the DRM subsystem, which will definitely cause X to slow to a crawl. The most likely cause of your errors is that the intel AGP driver (i810 or i915, depending on your hardware) isn't getting loaded. If that's the case, you should see an error such as: [drm] failed to load kernel module "i915" in Xorg.0.log just before the ones from intel. If the modules are being loaded, you'll likely see some other errors around that same area. The aren't tagged with (EE), unfortunately; try: # grep -5 'Failed to open DRM' Xorg.0.log You can also check your dmesg output to see if the devices are being initialized properly: platypus log # dmesg | grep agp Linux agpgart interface v0.103 agpgart-intel :00:00.0: Intel 965GM Chipset agpgart-intel :00:00.0: detected 7676K stolen memory agpgart-intel :00:00.0: AGP aperture is 256M @ 0xe000 platypus log # dmesg | grep drm [drm] Initialized drm 1.1.0 20060810 [drm] set up 7M of stolen space [drm] initialized overlay support fb0: inteldrmfb frame buffer device [drm] Initialized i915 1.6.0 20080730 for :00:02.0 on minor 0 If everything's working, you should have the following devices that the Xorg driver needs: platypus log # ls -l /dev/dri total 0 crw-rw 1 root video 226, 0 Apr 20 13:11 card0 crw-rw 1 root video 226, 64 Apr 20 13:11 controlD64
Re: [gentoo-user] Confusion with eix output
On 4/21/2010 3:48 PM, James Cunning wrote:
> I am having some trouble, I think, with my nvidia video driver, and eix
> produces some output I cannot decipher from information in the man page:
>
> jlc64 X11 # eix nvidia-drivers
> [D] x11-drivers/nvidia-drivers
> Available versions: [M]71.86.07!s [M]~71.86.09!s 96.43.09!s ~96.43.11!s
> 173.14.15!s ~173.14.18!s 180.29!s ~180.60!s {acpi custom-cflags gtk
> kernel_FreeBSD kernel_linux multilib userland_BSD}
> Installed versions: 190.42-r3!s(11:04:43 AM 04/21/2010)(acpi gtk
> kernel_linux multilib -custom-cflags)
> Homepage:http://www.nvidia.com/
> Description: NVIDIA X11 driver and GLX libraries
>
> The man page goes into great detail how to specify many things, but doesn't
> explain in simple terms the format of its default outputs. In particular, I
> don't understand what the [D] means, but would appreciate any clues to a more
> comprehensible explanation for all its output.
I feel your pain. The eix man page is a perfect example of why GNU made
the info system (which is otherwise overkill in most cases.)
At any rate: eix's output tries to mimic the emerge -v output when
possible. In this case, [D] means eix thinks your installed version is
higher than the latest unmasked version in the tree, and that you should
'D'owngrade. You could also see 'U', if eix thinks you need to upgrade.
You probably need to eix-update after your most recent sync. Use
eix-sync instead of emerge --sync if, like me, you tend to forget that step.
--Mike
Re: [gentoo-user] Updates = slow firefox
On 4/23/2010 1:37 PM, Grant wrote: >>> Can anyone confirm that as users we should be moving away from hal? >> >> http://www.gentoo.org/proj/en/desktop/x/x11/xorg-server-1.8-upgrade-guide.xml >> http://xorg.freedesktop.org/wiki/XorgHAL > > OK, and since xorg-server-1.7 doesn't have a udev USE flag, I should > probably stick with hal until 1.8. Please let me know if that isn't > the case. I'm on udev-149. If HAL is working for you, stick with it. If not, turn it off. Xorg 1.7 works equally well with or without HAL. The main difference is how much manually configuration you need to do. The relative stability of using/not using udev with Xorg 1.8 have yet to determined. :) --Mike
Re: [gentoo-user] Compiling 32 bit library on x86_64
On 4/30/2010 5:25 AM, Roger Mason wrote: > Hello, > > I need to compile a 32 bit version of libtermcap on an x86_64 (multilib) > system. Can someone tell me how to set up CFLAGS? This is what I have > at the moment: Have you tried using sys-devel/crossdev? It will set up the entire 32-bit cross-compiler environment for you; then it's just a matter of setting a couple of environment variables to switch compilers. --Mike
Re: [gentoo-user] Compiling 32 bit library on x86_64
On 4/30/2010 12:40 PM, Roger Mason wrote: > Mike Edenfield writes: > >> Have you tried using sys-devel/crossdev? > > Not in the present context. > >> It will set up the entire 32-bit cross-compiler environment for you; >> then it's just a matter of setting a couple of environment variables to >> switch compilers. > > Some time ago I tried setting up cross-compilation so that I could use a > rather heterogeneous collection of amd64, ppc and x86 machines in > icecream. Unfortunately I could not get cross-compilation to work. I > asked about it in this forum but did not get any replies. I have it set up on my laptop. I admit it's been a while since I used it, but I know it worked at one point. Though I was using it on a standard PC, the best source of information I found on the process was the Gentoo Embedded Handbook: http://www.gentoo.org/proj/en/base/embedded/handbook/?part=1 The whole first section is on setting up a cross-compiler, just substitute i686-pc-linux-gnu for your target architecture.
Re: [gentoo-user] Re: Safe to install libpng-1.2.44?
On 7/4/2010 10:05 PM, walt wrote: > On 07/04/2010 06:38 PM, Enrico Weigelt wrote: >> Hi folks, > >> * library imports should _always_ happen via pkg-config >> (dont use .la files) > > +1 (Am I allowed +100?) If so, allow me to +1billion
Re: [gentoo-user] Am I wrong?..
On 10/1/2009 10:44 AM, Arthur D. wrote: I just installed VIM with emerge, and removed nano because I considered it to be absolutely unnecessary in my system. Why I need nano? I am a VIM fan. And here the troubles begin... Run "sudo visudo" and you get this: ~ $ sudo visudo visudo: no editor found (editor path = /bin/nano) ~ $ env | grep -i edit EDITOR=/usr/bin/vim You have two options: 1. Tell sudo to preserve the EDITOR variable in /etc/sudoers: Defaultsenv_keep += "EDITOR VISUAL PAGER" Otherwise sudo will ignore your environment and use the defaults for the new user. 2. Change the default editor on your system by putting something in /etc/env.d: apollo ~ # cat /etc/env.d/99editor EDITOR="vim" --Mike
Re: [gentoo-user] Am I wrong?..
On 10/1/2009 1:34 PM, Arthur D. wrote: I'm using a 4 years old system, and if I change that line, log out and in again, it changes the env variable and everything works (that means the behavior is probably caused by your configuration). If visudo is still using that configuration, maybe that's because some configuration file has precedence over environment variables. In that case, you gotta find that file and change it. Not an easy task, anyway... I just did an "grep -r /bin/nano" in /etc. LOL, I know there's a better way, I'm just too lazy to look for it... Man, running "sudo visudo" and just running "visudo" is not the same. Be careful. Nano is hardcoded in sudo's ebuild. Normal users cannot run "visudo", so you must already be root to run it, or else use 'sudo visudo'. In the first case, it uses your EDITOR variable and there is no problem. In the second case, as already explained, it uses the first one of: * The EDITOR variable, if you've told sudo to keep it set * The default editor from the sudoers file, if you've set that * The default editor from the ebuild, which is nano. It is also not just visudo that has this behavior. Just run this: apollo ~ # sudo $EDITOR and if you haven't explicitly told sudo to preserve the EDITOR variable it will fail. As will any other program that reads EDITOR (or VISUAL, the other popular one). Point being, the behavior you're seeing isn't a bug in the sudo ebuild --- it's intended and intentional behavior of sudo itself. --Mike
Re: [gentoo-user] Am I wrong?..
On 10/1/2009 3:32 PM, forgottenwizard wrote: However, I'm also wondering why the ebuild doesn't make use of the EDITOR variable as was mentioned. This defaults to nano so it should work fine in a default install, and would avoid issues like this which seems to be an arguement that the dev(s) are trying to force specific programs on the users. I would be hesitant to use a user-specific variable like EDITOR to define the system-wide default on an ebuild. For example, what if my EDITOR was set to gvim or emacs when I installed sudo, then some other remote user tried to run visudo over ssh? --Mike
Re: [gentoo-user] Anyone using sys-devel/gcc-4.4.1
On 10/1/2009 6:26 PM, Dale wrote: It has finished the emerge -e system so far. Not a single failure that I can see. Do have to update a config file tho. ;-) In case anyone's keeping score, I've been using gcc-4.4 with the hardened profile (from the hardened-development overlay, of course) for a good while, and the only problems I've run into are hardened related. It has built OOo, Gnome, Gimp, Firefox, Emacs, and even CLISP with no apparent problems so far. --Mike
Re: [gentoo-user] Am I wrong?..
On 10/2/2009 1:29 AM, Arthur D. wrote: Agree. There's no need in making vim as depends. But in other hand in vanilla sudo package there's VI hardcoded by default. And MOST if not ALL users who have VIM So basically, you're entire silly argument boils down to "I don't like nano, make it go away." And yes, I also don't like nano, don't have it installed, and use vim for everything. Shockingly enough, visudo works *exactly the way I want*. So lets not go lumping "most if not all users who have vim" into your little rant. If I had to venture a guess, I'd say most, if not all, users who managed to get vim installed and nano removed are more than capable of configuring sudo appropriately. --Mike
Re: [gentoo-user] What is a "packet"? Was: Checksum error
On Mon, 2009-10-12 at 00:04 +0200, KH wrote: > Peter Ruskin schrieb: > > > > I'm 71 ... is that old enough? > > > To use a trendy idiom: That's cool. I believe the current trendy idiom (with the identical meaning) is "That's hot". Thus portraying exactly the problem with our language. :x --K
Re: [gentoo-user] Re: Library file formats
On 10/16/2009 12:54 PM, walt wrote: On 10/16/2009 04:37 AM, Neil Bothwick wrote: On Fri, 16 Oct 2009 12:20:30 +0100, Peter Humphrey wrote: I thought I'd have a play with swami, but the emerge fails with "/bin/sed: can't read /usr/lib64/libogg.la: No such file or directory", and indeed there is none such. The elog message from the last libogg install explains this. Run lafilefixer --justfixit. There is no manpage for lafilefixer, but the --help flag prints: --justfixit Choose some reasonable dirs, such as /usr/lib*, etc. , find all .la files and fix them to not use .la files for linking I can't make sense out of that -- one of the major uses of libtool (I thought) is for linking. Can anyone 'splain that to me? One of the major problems with .la files (besides being mostly useless on Linux/FreeBSD/any other ELF-based OS) is when they refer to other .la files instead of linking directly to the binaries. This breaks things when they get removed from a ebuild for a library that other libraries depend on. The only time that libtool archives provide a real benefit is when there's a need to link in static libraries that have external dependencies -- the libtool archive defines the dependency information that can't be stored in the static archive format. If you don't have any publicly-consumed static libraries, .la files are just pointless clutter, so package maintainers very often remove them. Any build system depending on libtool for its linking then breaks because the dependency chain is broken. Take, for example, this very real example from my system. I have both hal and dbus installed; hal depends on dbus, and both packages install libtool archives. In /usr/lib/libhal.la, there is the following: # Libraries that this one depends upon. dependency_libs=' /usr/lib/libdbus-1.la -lcap -lpthread -lrt' Thus, whenever a package that uses hal tells libtool to link to libhal.la, libtool recursively links to libdbus-1.la as well. Now, say the dbus maintainer suddenly doesn't like .la files (perhaps a bad break-up in Los Angeles), and removes them from the ebuild. The next time I true to use libtool to link in hal, it will fail because the latest dbus does not include /usr/lib/libdbus-1.la. The fix is to run lafilefixer, which changes the above line to say: # Libraries that this one depends upon. dependency_libs=' -L/usr/lib -ldbus-1 -lcap -lpthread -lrt' So that hal no longer cares whether or not the dbus package installed its libtool archive, and all is well. --Mike
Re: [gentoo-user] ">=net-libs/libsoup-2.25.1:2.4[gnome]" Huh?
On 10/19/2009 9:44 AM, Dale wrote: I just added -eds and this is what I get: r...@smoker / # emerge -uvDNa world These are the packages that would be merged, in order: Calculating dependencies ... done! emerge: there are no ebuilds built with USE flags to satisfy ">=net-libs/libsoup-2.25.1:2.4[gnome]". !!! One of the following packages is required to complete your request: - net-libs/libsoup-2.26.3-r3 (Change USE: +gnome) (dependency required by "dev-libs/libgweather-2.26.2.1" [ebuild]) (dependency required by "gnome-extra/evolution-data-server-2.26.3" [ebuild]) (dependency required by "app-office/openoffice-3.1.1" [installed]) (dependency required by "@world" [argument]) Did this emerge actually show that openoffice was being built without the "eds" USE flag? Did you possibly add "eds" to a local /etc/portage/package.use file for openoffice that's overriding your global USE settings? The eds USE flag is the root of your problem: openoffice[eds] requires evolution-data-server, which requires libgweather, which requires libsoup[gnome]. Try this: USE="-eds" emerge -upvtDN openoffice It should show openoffice being built without the "eds" flag and evolution-data-server not included in the dependency tree at all. --Mike
Re: [gentoo-user] Has MySQL become compulsory?
On 10/25/2009 8:10 PM, Dale wrote: Well, I put -semantic-desktop in my USE line and ran emerge -uvDN world. It recompiled several things and told me it had some @preserved-rebuild packages to build. So, I ran that and got this little message: r...@smoker / # emerge @preserved-rebuild -a These are the packages that would be merged, in order: Calculating dependencies ... done! emerge: there are no ebuilds built with USE flags to satisfy ">=kde-base/kdelibs-4.3.2:4.3[semantic-desktop,-kdeprefix]". !!! One of the following packages is required to complete your request: - kde-base/kdelibs-4.3.2-r3 (Change USE: +semantic-desktop) (dependency required by "kde-base/nepomuk-4.3.2" [ebuild]) (dependency required by "@preserved-rebuild" [argument]) The semantic desktop is the whole point of nepomuk. If you don't want all the semantic desktop KDE stuff, then you don't want nepomuk either. Try removing it instead of rebuilding it. --Mike
Re: [gentoo-user] Re: Linux Magazine tests Gentoo performance
On 10/29/2009 5:13 PM, Nikos Chantziaras wrote: On 10/29/2009 10:45 PM, Nikos Chantziaras wrote: I suppose this is interesting to most Gentoo users. Linux Magazine performed a detailed benchmark of Gentoo, comparing it to Ubuntu 9.04: http://www.linux-mag.com/id/7574/1 Btw, I think this is a very nice example of why per-package CFLAGS would have been very useful. Some application largely benefit from -Os, others from -O2. If you're willing to put in a bit of effort, I believe you can set up a per-package environment (including custom CFLAGS) in a number of ways. The one that seems to be most popular is described here: http://www.mail-archive.com/gentoo-portage-...@lists.gentoo.org/msg00585.html
Re: [gentoo-user] hal vs. devicekit (was: more about hal)
On 10/31/2009 4:59 PM, Alan McKinnon wrote: Even it's author knows this (but apparently many distros do not) which is why he deprecated hal and started over with devicekit. Speaking of which... Has the switchover to devicekit officially started and I missed it? And if so, is there some migration guide of some sort I can peruse for assistance? I just upgraded Xorg and Gnome yesterday and my previously more-or-less-working with hal system went belly up. It no longer recognizes my synaptics touchpad, and in fact refused to start X until I re-merged the kdb/mouse drivers, and gnome-power-manager is totally useless. For the record, I'm all about moving to devicekit from hal if that's where the dev(s) are moving, but I'd rather do so while continuing to use my laptop :) --Mike
Re: [gentoo-user] Kernel upgrading and linux symlink
On 10/31/2009 7:29 PM, Mark Knecht wrote: So I think this thread addresses a question I've had about the kernel installation process over the years. I only copy bzImage to /boot with a rename to whatever this kernel is. I don't do anything with the other files - System.map and something else - which I don't even have on most of my systems anymore. They don't seem to be needed. Are they just things used in the old days but now too outdated or replaced by other stuff? (Like config.gz in the kernel, etc.) Having a copy of the config file is useful, for me anyway, when I have multiple types of kernel (one without PaX, one without SELinux, whatever) and a new version comes out, I can copy its config from /boot to .config and run make oldconfig. The System.map file is probably the least useful of the three for the average user. It's main use is for address resolution in oops messages. I think ps uses it for something as well. --Mike
Re: [gentoo-user] When masked pkg not in [...]profiles/package.mask, where is it
On Sun, 2009-11-01 at 20:17 -0600, Harry Putnam wrote:
> When a package comes up as masked in an eix search, they are usually
> found in /usr/portage/profiles/package.mask, but if a particular
> masked package is not listed there... where else would it be.
>
> I see libtool is masked above version 1.5.26-r1, but is not in
> /usr/portage/profiles/package.mask...
According to my eix output, libtool isn't masked, just keyworded (and
only 1.5.26-r1 specifically):
[I] sys-devel/libtool
Available versions:
(1.3) 1.3.5
(1.5) 1.5.26 (~)1.5.26-r1 2.2.6a **
{test vanilla}
That information is per-ebuild, found in the KEYWORDS line in
the .ebuild file; it will have keyword ~x86 or ~amd64 instead of simply
x86 or amd64.
--Mike
Re: [gentoo-user] Firefox 3.5 without changing entire system to ~arch?
On 11/2/2009 12:16 PM, Marcus Wanner wrote: Could anyone tell me how to install firefox 3.5.x without changing the entire installation to ~arch? I have been looking around on the web for how to do this, but can't find anything that doesn't require being (afaict) very invasive to the rest of the system. Thanks! For mozilla specifically, see below. In general, you would create a file called /etc/portage/package.keywords (or as I prefer, a directory /etc/portage/package.keywords/ and then a file within that directory), and put the package name in it. However, it's generally considered a bad idea to mix and match arch and ~arch on a single system, since the dependencies cascade pretty quickly. You'll eventually end up with the core packages on your system keyworded anyway, which defeats the whole point of running stable. P.S. Mozilla considers this version to be stable/mature and is pushing out it as the version most people should use. Why hasn't it been marked as stable in the portage tree yet? It may just be an oversight; check bugs.gentoo.org to see if there's already a bug report asking it to be stabilized. If there isn't already one, just file a new one. (It may help to mention that it's stable on amd64 already.) --Mike
[gentoo-user] NetworkManager and/or WICD
Does anyone have any experience getting either NetworkManager or WICD to work properly under Gentoo? When I attempt to use either of those utilities to get onto my wireless network, the NIC refuses to stay connected to the base station for more than a few seconds at a time. Instead, it continually disassociates and deauthenticates, only for nm/wicd to hop right back on. However, if I manually configure wpa_supplicant for a given SSID and start it via the init script directly, I don't have any such problems. Additionally, I can't get either applet to actually save settings (like known networks, passwords, etc.) which means that I'm continually interrupted by a prompt for the wireless password. What I'm really looking for is a graphical utility that will let me connect to specific newly-detected wireless networks without having to edit the WPA configuration and restart it; NetworkManager and WICD seem to be the most popular. Any other suggestions would also be welcome. --Mike
Re: [gentoo-user] Re: kernel build - back in the soup.
On 11/3/2009 11:10 PM, Harry Putnam wrote: hamilton writes: Just checking - but you didn't mention: did you copy the .config to the new kernel src directory? If not, that would certainly explain the disparity in configuration settings you're seeing. I think you can say make `oldconfig' and the `old config' is supposed to be incorporated so no I didn't If I had put .confg into the new sources, then plain make menuconfig is what I would have used. Do you know where the man pages or docs for that stuff is .. its not in `man make' The 'make' man page wouldn't know anything about the kernel's makefile. You want the README file that's included in the top of the kernel source folder. That file says, among other things: "make oldconfig" Default all questions based on the contents of your existing ./.config file and asking about new config symbols. You need to already have a .config file in the source tree in order for 'make oldconfig' to work; otherwise you are going to get the default answers to just about every question. The benefit of this is that you don't have to search through the entire menu tree in the UI to find what's new. When you're ready to build a new kernel version, you should copy the .config file from your current kernel into the new source tree. For example, if you use 'make install' it will copy .config to /boot/config-; from there you can copy it back to /usr/src/linux/.config for the next version. When you run 'oldconfig' you should rarely get more than a few dozen questions, and it should all be on truly new items that didn't exist in your previous kernel. The hardware drivers you selected should all carry over as-is. --Mike
Re: [gentoo-user] Another angle on hal/xorg thread
On 11/4/2009 10:51 AM, Harry Putnam wrote: I didn't want to derail the ongoing thread about hal/xorg with this question there. Far as I remember I haven't done anything special concerning hal but at some point hal disappeared. And is not on my system anymore. I believe that some packages in portage recently masked off the "hal" USE flag (GNOME stuff, maybe?), so if those were the only packages relying on hal it might have gone away. I've always used and /etc/X11/xorg.conf file for starting X. What I'm wondering from seeing this kind of topic frequently here is if I'm running in some deprecated mode? If my setup using no hal, and xorg.conf is going to become outdated and stop working anytime soon? The answer is a solid "who the heck knows". If it works for you now, don't mess with it. Wait for the Xorg/hal/devkit/whatever situation to settle down before you go making any drastic changes. Some people, like myself, are running X with hal and no .conf file and it works like a champ. I get better hardware detection with hal, especially on my laptop, than I ever got manually. Other people have had problems with hal and Xorg not detecting their hardware at all. What you are "frequently" seeing is those people reminding everyone, every time the topic come up, that you don't *need* to use the new hal-ified way if it doesn't work for you. All of this is probably moot because hal itself is going away and being replaced by devicekit, but not yet because devicekit isn't quite ready. What the configuration situation will be under devicekit I have no idea, though I would hope having no configuration file would still be a goal for the devkit team. --Mike
Re: [gentoo-user] NetworkManager and/or WICD
On 11/3/2009 11:16 PM, Iain Buchanan wrote: On Tue, 2009-11-03 at 21:52 -0500, Mike Edenfield wrote: When I attempt to use either of those utilities to get onto my wireless network, the NIC refuses to stay connected to the base station for more than a few seconds at a time. Instead, it continually disassociates and deauthenticates, only for nm/wicd to hop right back on. However, if I manually configure wpa_supplicant for a given SSID and start it via the init script directly, I don't have any such problems. Have you stopped your net.wlan0 script? Also remove it from the default runlevel, and set "rc_hotplug="!net.wlan0 !net.eth0" in /etc/rc.conf (if you're using openrc). I hadn't done that initially, but I did this time, and still get the same issue. I'm having trouble determining if the problem is NetworkManager telling the NIC to disassociate, or the NIC telling NetworkManager to disassociate. As long as NetworkManager is running, my dmesg output shows this happening about once every 15 seconds or so (I don't know what any of those status codes mean, or they may help troubleshoot the problem :) ) wlan0: authenticate with AP 00:1e:58:04:1e:ec wlan0: authenticated wlan0: associate with AP 00:1e:58:04:1e:ec wlan0: RX AssocResp from 00:1e:58:04:1e:ec (capab=0x411 status=12 aid=1) wlan0: AP denied association (code=12) wlan0: associate with AP 00:1e:58:04:1e:ec wlan0: RX AssocResp from 00:1e:58:04:1e:ec (capab=0x411 status=12 aid=1) wlan0: AP denied association (code=12) wlan0: associate with AP 00:1e:58:04:1e:ec wlan0: RX AssocResp from 00:1e:58:04:1e:ec (capab=0x411 status=12 aid=1) wlan0: AP denied association (code=12) wlan0: association with AP 00:1e:58:04:1e:ec timed out wlan0: authenticate with AP 00:1e:58:04:1e:ec wlan0: authenticated wlan0: associate with AP 00:1e:58:04:1e:ec wlan0: RX AssocResp from 00:1e:58:04:1e:ec (capab=0x411 status=0 aid=1) wlan0: associated wlan0: disassociated (Reason: 14) wlan0: deauthenticated (Reason: 6) The syslog output from NetworkManager shows basically the same cycle: Nov 04 14:12:11 [NetworkManager] Activation (wlan0) successful, device activated._ Nov 04 14:12:11 [NetworkManager] Activation (wlan0) Stage 5 of 5 (IP Configure Commit) complete._ Nov 04 14:12:38 [NetworkManager] (wlan0): supplicant connection state: completed -> disconnected_ Nov 04 14:12:38 [NetworkManager] (wlan0): supplicant connection state: disconnected -> scanning_ Nov 04 14:12:38 [NetworkManager] (wlan0): supplicant connection state: scanning -> disconnected_ Nov 04 14:12:40 [NetworkManager] (wlan0): supplicant connection state: disconnected -> associating_ Nov 04 14:12:41 [NetworkManager] (wlan0): supplicant connection state: associating -> disconnected_ Nov 04 14:12:50 [NetworkManager] (wlan0): supplicant connection state: disconnected -> scanning_ Nov 04 14:12:52 [NetworkManager] (wlan0): supplicant connection state: scanning -> associating_ Nov 04 14:12:52 [NetworkManager] (wlan0): supplicant connection state: associating -> disconnected_ Nov 04 14:12:53 [NetworkManager] (wlan0): device state change: 8 -> 3 (reason 11)_ Nov 04 14:12:53 [NetworkManager] (wlan0): deactivating device (reason: 11)._ As soon as I stop NetworkManager and start net.wlan0, it jumps onto the AP and stays there for good: wlan0: authenticate with AP 00:1e:58:04:1e:ec wlan0: authenticated wlan0: associate with AP 00:1e:58:04:1e:ec wlan0: RX AssocResp from 00:1e:58:04:1e:ec (capab=0x411 status=0 aid=2) wlan0: associated Additionally, I can't get either applet to actually save settings (like known networks, passwords, etc.) which means that I'm continually interrupted by a prompt for the wireless password. Do you have seahorse (gnome)? Depending on how it's setup, you should only have to provide the master password once for nm to access all your keys. I do have seahorse, and when I first boot up it asks me for the default keyring password and connects to the AP automatically. Every time after that, when the connection bounces, syslog shows: Nov 04 14:13:21 [NetworkManager] Activation (wlan0/wireless): connection 'Auto Informagration' has security, and secrets exist. No new secrets needed._ but NetworkManager presents the WPA pass phrase dialog anyway. --Mike
Re: [gentoo-user] Strange iwl3945 behavior (possibly wpa_supplicant related?)
On Sat, 2009-11-07 at 12:09 +0200, Alan McKinnon wrote: > On Saturday 07 November 2009 04:20:09 Mike Edenfield wrote: > > When using NetworkManager on my work network, however, things go > > horribly wrong. I get tons of this in my kernel logs: > > > > wlan0: authenticate with AP 00:1e:58:04:1e:ec > > wlan0: authenticated > > wlan0: associate with AP 00:1e:58:04:1e:ec > > wlan0: RX AssocResp from 00:1e:58:04:1e:ec (capab=0x411 status=12 aid=1) > > wlan0: AP denied association (code=12) > > wlan0: associate with AP 00:1e:58:04:1e:ec > > wlan0: RX AssocResp from 00:1e:58:04:1e:ec (capab=0x411 status=12 aid=1) > > wlan0: AP denied association (code=12) > > wlan0: associate with AP 00:1e:58:04:1e:ec > > wlan0: RX AssocResp from 00:1e:58:04:1e:ec (capab=0x411 status=12 aid=1) > > wlan0: AP denied association (code=12) > > wlan0: association with AP 00:1e:58:04:1e:ec timed out > > wlan0: authenticate with AP 00:1e:58:04:1e:ec > > wlan0: authenticated > > wlan0: associate with AP 00:1e:58:04:1e:ec > > wlan0: RX AssocResp from 00:1e:58:04:1e:ec (capab=0x411 status=0 aid=1) > > wlan0: associated > If it works everywhere else and not at home, the difference is obviously with > your home router. > > What config does it have and how does it differ from what everywhere else has? Well, it works at home but not at work, so I don't have much information beyond what they can tell me. I'll try to find someone who knows more, but as far as I can tell it's nearly identical to what I have at home: a single WAP with a broadcast SSID using WPA Personal, even using the same (cheap) Linksys hardware. What really confuses me is that the NIC works fine at work *if* I run wpa_supplicant manually; it only seems to fail when NetworkManager is controlling the NIC. So, yeah, it seems like the difference is with NetworkManager and/or wpa_supplicant, but I have no idea what that difference is. (Also, to head off the upcoming "just don't use NetworkManager": this laptop is eventually going to someone who'll be roaming a lot more than I do, for whom constantly editing wpa_supplicant.conf isn't really an option. Wicd doesn't support VPN connections, so NetworkManager seems to be my only option :\) --Mike
Re: [gentoo-user] sun-jre-bin end-of-life - what now?
On 11/9/2009 9:41 PM, Mark Knecht wrote: If I remove the new ACCEPT_LICENSE="dlj-1.1" I added to make.conf So don't do that? !!! All ebuilds that could satisfy "dev-java/sun-jre-bin" have been masked. !!! One of the following masked packages is required to complete your request: - dev-java/sun-jre-bin-1.6.0.17 (masked by: dlj-1.1 license(s)) - dev-java/sun-jre-bin-1.6.0.16 (masked by: dlj-1.1 license(s), ~x86 keyword) - dev-java/sun-jre-bin-1.6.0.15 (masked by: dlj-1.1 license(s)) - dev-java/sun-jre-bin-1.5.0.22 (masked by: dlj-1.1 license(s)) - dev-java/sun-jre-bin-1.5.0.21 (masked by: dlj-1.1 license(s), ~x86 keyword) - dev-java/sun-jre-bin-1.5.0.20 (masked by: dlj-1.1 license(s)) - dev-java/sun-jre-bin-1.4.2.19 (masked by: package.mask) One very old version is masked off because of security problems. Siz different subsequent versions are available as long as you accept the new license. --Mike
Re: [gentoo-user] Downgrade glibc-2.11 to 2.10
On 11/17/2009 11:06 AM, Alan McKinnon wrote:
Yes, I have read this in
/var/portage/sys-libs/glibc/files/eblits/pkg_setup.eblit
and I understand the risks.
if has_version '>'${CATEGORY}/${PF} ; then
eerror "Sanity check to keep you from breaking your system:"
eerror " Downgrading glibc is not supported and a sure way to destruction"
die "aborting to save your system"
fi
I want to do it anyway.
I have never done this on a package-managed system but I did have to
downgrade glibc on a slackware system once. It's very VERY ugly; all
kinds of implementation details get exposed as special symbols, or new
symbol versions, or whatever, that make going backwards a mess.
The way I was told to do it was to get a compiled copy of lower glibc
version into an alternative install path, like /usr/local/glibc, and
rebuild everything against that copy. You could have emerge build 2.10
into an alternative --root, or go get a binary package of 2.10 and
uncompress it somewhere. Then update CFLAGS and LDFLAGS to include the
correct -L and -I parameters pointing to
/usr/local/glibc/{lib,usr/lib,include}.
When I did it, I rebuilt everything on my system twice, just to be safe.
First time through, you build against the extra copy of glibc,
including building a downgraded glibc in the proper system location, and
having the build tools link to the correct lower version. Then you
remove the CFLAGS/LDFLAGS and rebuild everything again, this time
against the downgraded version in the correct location, and then you can
remove your extra copy. I'm not entire positive that second one was
strictly necessary but it worked.
Or you can just back up your data and reinstall :)
A multitude of apps that used to run just fine now give "free(): invalid
pointer" errors since I upgraded to glibc-2.11
Make sure you file bug reports on these. The programs are probably
doing buggy things that glibc used to be rather forgiving about. I
believe in 2.11 they added extra checks to the memory management used by
C++ programs, though I don't know specifics. This would catch things
like using delete where they meant delete[], or free() on something
allocated with new.
The standard says the behavior of this type of operation is undefined,
so glibc is technically free to do "anything it wants". Unfortunately,
when glibc is nice and make that "anything" be "what you wanted it to do
anyway" it encourages people do keep doing bad things, thus the
ever-increasing strictness of the library.
--Mike
[gentoo-user] overlays.gentoo.org having git issues?
None of my git-based overlays from overlays.gentoo.org have been able to sync this morning. Is the server having problems or is something wrong with my git? --Mike
Re: [gentoo-user] Wireless...
On 12/2/2009 9:17 PM, BRM wrote: I have wireless working (b43legacy driver for the Dell Wireless Broadcom) through a static configuration in /etc/conf.d/net - basically: essid_wlan0="myWLAN" key_MYWLAN="somekey" config_MYWLAN=( "dhcp" ) preferred_APS= ( "myWLAN" ) I would like to use a tool like WPA Supplicant instead so I can have a more dynamic configuration. I've tried to setup WPA supplicant but haven't been able to get it to work. Probably not what you wanted to hear, but I had the exact same problem with the Dell bcm-based adapter in my Inspiron laptop. It would work fine for open wireless and WEP-secured wireless, but wouldn't associated with a WPA-secured access point. Eventually I spent about $30 to purchase an iwl3945 replacement from Dell, which worked fine, and never looked back. --Mike
Re: [gentoo-user] What magic does portage use?
On 12/11/2009 9:38 AM, Dale wrote: Mickaël Bucas wrote: From the process name, you can deduce the service and restart it. I've never needed a reboot for this kind of problem. You may have to switch to run level 1 to restart some important services like udev. Actually, you can kill udev and restart it. Kill the process and then run "/sbin/udevd --daemon" and it will be started again. Yeah, or you could, you know, just reboot. Frankly I have never figured out the irrational fear Linux people have about rebooting their machines after a big upgrade. It takes my laptop way less time to shutdown and restart than it does for me to manually stop and restart everything that just got updated, and I can go grab a soda in the meantime. --Mike
Re: [gentoo-user] Is rc.conf no longer used by Gentoo (baselayout-1.12.13)?
On 12/14/2009 3:50 PM, Mick wrote: When I am looking for my XSESSION I get nothing: $ echo $SESSION How do I set this up, other than Mike's suggestion of '/etc/env.d/90xsession'? With baselayout-2, setting it in /etc/env.d is the correct method; if you want per-user sessions you can also set it in your local .bashrc file. Also, note: $XSESSION != $SESSION, just in case it's already set and you missed it. --Mike
Re: [gentoo-user] The current correct way to start kde 4
On 12/28/2009 1:49 AM, Andrew Lowe wrote: On 28/12/2009 1:03 PM, Volker Armin Hemmann wrote: On Montag 28 Dezember 2009, Kirill Lipatov wrote: well yeah. I just think that setting the kdm as the login manager is the easiest way to automatically start kde4 session after loging in. However, kdm is of course not mandatory and it can do more than just start kde4 OP wrote he wanted 'kde' as login manager. Which implies he wants kdm. Thank you gentlemen for your replies. Setting DISPLAYMANAGER to kdm in /etc/conf.d/xmd it will be then. My confusion comes from there being posts suggesting this method, others setting XSESSION in /etc/env.d/90xsession, things mentioning /etc/X11/sessions and so on. I've still got my training wheels on so it can be a bit confusing at times. You are probably confusing KDE - the entire Qt-based desktop environment, with kdm - a Qt-based display manager and an (optional) component of KDE. The first option will set up the KDE display manager to run when you boot, but it won't actually start "KDE". The second will make KDE the default when you log in through the generic xdm display manager. The last option will get KDE to show up in the list of known session types (for example, what you see in the drop-down list on gdm). You don't *need* to use KDE's display manager to launch KDE, you just need to tell whatever display manager you have to start the kde4 session. Similarly, just because you run kdm doesn't mean you have to launch KDE when you log in; you could launch any of the sessions in kdm's list. But if you are only installing KDE on your machine, then there's really no good reason not to use kdm as well, and allow it to default to KDE4, so you should be all set. --Mike
Re: [gentoo-user] Understanding sets
On 1/5/2010 12:53 PM, Zeerak Waseem wrote: Ah well, having the sets in the wrong place would give me an error. And regarding portage version, weren't sets included in v. 2.x.x? My portage version is 2.1.7.16, anything above seems to be hardmasked... No, sets appear starting in portage v2.2. And yes, the whole 2.2 series is hard-masked. The stated reason is so that the 2.1 test versions will actually get some testing. Opinions over precisely how silly it is to test old, being-obsoleted versions in deference to new, enhanced versions vary and can be found in the list archives. Most unstable users have probably unmasked sys-apps/portage long ago and moved into the future. --Mike
Re: [gentoo-user] Devicekit - especially just for Dale
On 1/18/2010 5:10 AM, Dale wrote: +1 I do OK with plain text but no clue on the new xml stuff. Why not just keep it simple? Is xml REALLY needed? XML allows you to generate complex, structured, hierarchical data that can be read, changed, and stored by well-tested third party libraries that don't need to know anything about the contents or meaning of your configuration data beforehand. This means I, as a developer, don't need to write any code to read and parse configurations, validate the syntax or structure (only the content), or persist it back out. In simpler terms: less time spent on the configuration parser, more time spent being productive. If there was a less-verbose alternative that was as easy to implement, with known stable parsing libraries, that had the same expressiveness as XML, I'd probably use that instead. But when you're talking about data that goes beyond a simple list of name/value pairs, anything attempt to stream it to a flat-file format is going to result in something that is either 1) redundant, or 2) hard to read. I'd go with 2 over 1 any day. In my opinion, if the worst thing you can come up with to complain about is "they used XML for their configuration files", then I'd say that software is in pretty good shape. On the other hand, even I can see that HAL has plenty of problems (besides its XML configuration). The fact that it completely fails to work for you being a good example :) --Mike
Re: [gentoo-user] correct way for unmasking several packages
On 1/22/2010 6:37 AM, Arnau Bria wrote: Hi all, I'm wondering which is the correct way for unmasking several depending packages when you want to unmask only want. Let me explain, I want to unmask xorg-server, and every time I try to update, it complains about one package, I unmask it, then it complains about other, and so on... You can try autounmask, which appears to scan the dependency tree and package.keyword everything it finds. Though I've never used it personally, I've seen plenty of reports of success. However, this cascade of unstable dependencies is the reason so many people discourage mixing arch and ~arch. Especially if you're using an unstable version of something with as many dependencies as xorg-server, by the time you're done you'll likely have half of your system unmasked. If you're willing to run an unstable Xorg anyway, you might want to consider running fully ~arch. --Mike
Re: [gentoo-user] baselayout2/openrc question
On 2/2/2010 3:48 PM, Alan McKinnon wrote: No, you completely misunderstand what stable, unstable and masked mean. You are using stable (and call it unstable which is wrong). What you call masked is actually called unstable. Masked is something else entirely. Do not confuse these terms. They have *exact* meaning. Has there ever been any discussion on coming up with more precise wording for portage's error messages? I suspect a lot of confusion between masked/keyworded comes from the fact that portage calls them all "Masked", e.g.: !!! All ebuilds that could satisfy "=app-editors/vim-7.2.303" have been masked. !!! One of the following masked packages is required to complete your request: - app-editors/vim-7.2.303 (masked by: ~amd64 keyword) Not that I came up with any better wording off the top of my head, but is the portage team open to suggestions? Or has this issue been beaten to death already? --K
Re: [gentoo-user] baselayout2/openrc question
On 2/4/2010 6:05 AM, Alan McKinnon wrote: How about a portage feature request? The contents of @system can have dependencies. Put a setting in a conf file which means the system uses portage, therefore python is in @system. Without the setting, python does not get included in @system. Since the system package set is package manageragnostic, perhaps that's the wrong place to deal with dependencies for portage or whatever? The package manager itself should know enough to keep itself functional. It just seems odd that portage lets you use portage commands to break portage. If I wrote a package manager in C# (as a ridiculous example) I would probably make it smart enough not to let you remove mono. --K
Re: [gentoo-user] baselayout2/openrc question
On 2/4/2010 10:43 AM, Neil Bothwick wrote: On Thu, 4 Feb 2010 16:14:25 +0200, Alan McKinnon wrote: How about giving the same warning when unmerging a dependency of @system as you do when unmerging a package directly in there. Either way, you risk breaking the system. Aren't all deps of packages in @system themselves already in @system? No, otherwise portage would complain if you tried to unmerge python. Anyway, deps are USE-dependent. Try USE="X" emerge @system on a headless server to see jut how much @system can pull in. Portage isn't in @system, either. "virtual/portage" is, but paludis also provides that. Python isn't a dependency of any other system package (except "file", but that's only enabled by the USE flag). It appears that portage's refusal to unmerge itself is hard-coded into portage; that reinforces my belief that portage should be responsible for refusing to unmerge it's own dependencies. --Mike
Re: [gentoo-user] Bizarre SSH connection reset
Dan Farrell wrote: On Sun, 9 Mar 2008 20:16:09 -0400 "Mark Shields" <[EMAIL PROTECTED]> wrote: Are you thinking his ISP is doing port-based connection filtering? What kind of connection filtering allows a connection to go through for 5 seconds, then resets it? Comcast? -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Bizarre SSH connection reset
Mick wrote: On Tuesday 11 March 2008, Dan Farrell wrote: On Mon, 10 Mar 2008 22:51:42 + Mick <[EMAIL PROTECTED]> wrote: On Monday 10 March 2008, Dan Farrell wrote: On Mon, 10 Mar 2008 15:43:55 -0400 Mike Edenfield <[EMAIL PROTECTED]> wrote: Comcast? I was on comcast for a long time (2.5 yrs) and never had a problem like this. They might have blocked port 25 and squelched my bittorrenting at times, but never anything like this. Of course, ymmv. IIRC they also block port 80 for sure on their retail accounts. They don't want the average punter to run a webserver at home. Even when they blocked port 25 for me bidirectionally (evidently sending 6 gigs through that port made me look like a spammer, even if it was all to the same address ;) ), and I called security assurance and they listed that among all the open ports I wasn't allowed on a residential account, even then, they still didn't block port 80 (or 26, 22, 21, 110, 993, or any other port!). Hmm, I don't know . . . The particular address I was trying to connect was definitely blocked. Other than not beeing able to connect with a browser, nc, httping and tcptraceroute confirmed it). Could it be an area/account specific block perhaps? When I questioned the owner he said that this was common practice and that his ISP does not allow webservers to run. When I was on Comcast, the only ports they blocked outright, that I found, were mail related. Presumably this was a spam prevention measure more than anything else. However, they did *monitor* other common ports for traffic. Occasionally I'd put some local service or another on my firewall during development, or for testing, or whatnot. If it happened to be on port 80, 443, or 21, I'd usually get a nasty-gram from then within a day reminding me of their AUP. --Mike -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] crontab entry
Kaushal Shriyan wrote: Hi I have the following entry in the crontab [EMAIL PROTECTED] 0 18 * * * /home/kaushal/rsync_mysql.sh I want my subject line to be "hostxx:yyDB refresh daily" is there a way to do it Thanks and Regards Kaushal The easiest way is to write a wrapper script; I have a few of them that do something like: ( echo "From: songbird.jungle <[EMAIL PROTECTED]>" echo "To: Michael Edenfield <[EMAIL PROTECTED]>" echo "Subject: Portage Update Report" echo "" # do stuff here. ) | sendmail [EMAIL PROTECTED] -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] emerge -aDNvu world fails
Mark Knecht wrote: On Sat, Mar 29, 2008 at 10:46 AM, Norberto Bensa <[EMAIL PROTECTED]> wrote: Mark Knecht wrote: > I never emerge everything I see in the list since most are > dependencies themselves and I don't want to add them to my world file. > emerge -1 what-ever-dep-you-dont-want-in-world Of course, if I Want to figure out which 10 items shouldn't be in my world file. I'm doing that right now with an openmotif emerge. However I was saying that if I see a big list of items, do equery depends one of the items, see something like k3b or even gnome, then emerging gnome picks up the dependencies even easier than using -1. At least I feel that way, but that's just me. - Mark I usually do this too, unless the number of ports to upgrade is very small. One tip that might save you time, add --tree to your emerge call and you can tell at a glance when an ebuild with a lot of dependencies is in your list. -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] scp escape characters
Mick wrote: On 16/04/2008, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: Am Mittwoch, den 16.04.2008, 15:06 +0100 schrieb ext Mick: > I was trying to scp a file which had spaces in its name; e.g. > > This\ is\ the\ name\ of\ it.txt I tried it in my zsh, with TAB-completion (means: I typed "scp remote_host:~/Fil"), and it gives this: scp remote_host:~/File\\\ with\\\ blanks And, of course, it copies the file just fine :-) Thanks guys, I'll try it next time. Just one thing: if the spaces need to be escaped at both ends, shouldn't it have just 2 \\, why are you showing it with 3 \\\ ? It's escaping the first slash and the space: '\\ \ ' -> '\ ' -> ' ' --K -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] What overwrites resolv.conf
Michael Higgins wrote: So, what overwrites it, when, how, and how to stop it? Is there a definitive guide to the syntax of the various config files? Or, BETTER YET, is there anyone who has a smoothly-functioning configuration to switch between wireless DHCP and connected hard-wired net setups and would like to share? It's being overwritten by your DHCP client, which is the expected behavior. As long as your setup is such that your wireless card always uses DHCP, and your wired card never uses DHCP, you can configure the Gentoo networking script to do the right thing depending on which interface you're starting up. In your /etc/conf.d/net setup, add variables for: config_eth0 = ( "w.x.y.z/nn" ) dns_servers_eth0 = ( "w.x.y.z", "w.x.y.z" ) dns_domain_eth0 = "my.domain" This will work as long as you manually stop and start the interfaces when you switch adapters. The net.eth0 startup script will write out a new resolv.conf, etc. --Mike -- gentoo-user@lists.gentoo.org mailing list
[gentoo-user] New b43 driver can't find my AP
Hi, After upgrading to the 2.6.24 kernel and switching from the bcm driver to the new b43 driver, I can no longer attach to my AP. If I downgrade to 2.6.23 it starts working again, so I'm confident that the hardware setup is all fine. Using the new driver, "iwscan" does not locate any APs and running wpa_supplicant hangs scanning for an AP to attach to. The driver seems to find the card ok and I followed the instructions here: http://wireless.kernel.org/en/users/Drivers/b43 to get the correct firmware. I also compiled both b43 and b43-legacy (since I have a 4306 rev 3 and wasn't sure which one worked) but the driver attaches b43 to my card. Does anyone know what I can do to track down the problem? --Mike -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] Re: emerge nano ?
This has been reported to bugs.gentoo.org -- it seems to be a bug in nano. For the time being you can get it to build if you enable the "spell" USE flag: echo "app-editors/nano spell" >> /etc/portage/package.use emerge nano -- gentoo-user@lists.gentoo.org mailing list
Re: [gentoo-user] protect running kernel fomr depclean?
On Sun, 2009-04-12 at 17:11 -0500, Dale wrote: > Alan McKinnon wrote: > > If it's already installed but not in world use -1 to tell portage not to go > > through the install process, but just add it to world. Or just edit the > > world > > file by hand. > > > > You sure about the -1 option? I thought that would emerge foo and not > add foo to the world file which --depclean would certainly focus on. I think Alan meant to say -n (--noreplace) instead of -1 (--oneshot). -n does exactly what he describes: it will not actually build packages that are already installed, but will add the explicitly specified packages to world. -1, as you said, does the opposite: it *will* recompile and reinstall the package, but won't update world. --K
Re: [gentoo-user] xorg-server-1.5.3 no mouse [solved]
Michael P. Soulier wrote: On 12/04/09 Michael P. Soulier said: And for me, I had the exact same problem. I'm not using HAL, and the x11-drivers/xf86-input-mouse-1.4.0 had to be rebuilt. This was not detected automatically. And, I missed the ebuild note to rebuild drivers. I did have to comment out the same line in my xorg.conf. Run "qlist -I -C x11-drivers/" for a list of what should be rebuilt, which is in portage-utils. At some point I should just enable hal and be done with it. Full agreement here, but just for the record, having hal enabled would not have saved you from the mismatched driver versions. That was a problem entirely between the xorg-server and xf86-* driver modules. --Mike
Re: [gentoo-user] Re: I don't like xorg-server 1.5.3
On 4/13/2009 12:55 PM, Nikos Chantziaras wrote: Paul Hartman wrote: On Mon, Apr 13, 2009 at 11:25 AM, Mark Knecht wrote: There's a lot of us voting 1 today I think. How do things like this go stable when they aren't stable, tested and not causing problems. (rhetorical...) I must be lucky because I've been using it since it hit ~amd64 and using the HAL/fdi way and it works fine for me. :) Same here. All is working perfectly (or almost perfectly; see crappy ATI Catalyst drivers) for months. Linux is getting better and the X.Org updates are playing a major part. I've never had any problems with HAL or the new X that I didn't cause myself. (e.g. enabling modesetting in the kernel by accident, blindly copying FDI files from the intarwebs without noticing that the hal package already included then, completely failing to read the update-your-drivers warning, etc). In other words, exactly the same thing that happened to old-X when you didn't pay attention to what you are doing, happens to new-X when you don't pay attention to what you are doing. On the other hand, for the first time since I started putting Linux on my laptops, I have (with zero effort on my part) a working Synaptics touchpad with actual Synaptics features AND X recognizes my hot-plugged USB mouse. HAL++. --Mike
Re: [gentoo-user] Re: I don't like xorg-server 1.5.3
On 4/13/2009 3:50 PM, 7v5w7go9ub0o wrote: I have not yet added hal; seems like unnecessary complexity at this point - I don't know how it will make life better. The major benefit of hal is for people who don't actually *have* an "old" xorg.conf. In most cases, the X server can do a better job than a user at detecting and configuring the video hardware, while hal can do a better job than a user at detecting and configuring input hardware. This means that, "out of the box" a clean X installation has a decent chance of working with zero configuration, at least sufficiently to get you running well enough to tweak. It also means that, as the specifics of the hal FDI database improves, you will automatically pick up improvements from upstream without needing to modify your configuration. If you have a working X setup, then "replacing" it with hal may not provide much in the way of benefits. But the same argument can be made just on the decision to upgrade to 1.5.3 -- if your current version "works", why the motivation to upgrade? If its just to have the "latest and greatest" -- well that's hal. --Mike
Re: [gentoo-user] Re: I don't like xorg-server 1.5.3
gibbo...@gmail.com wrote:
1) I don't want hal, one more daemon running only to... spot /dev/input/*,
from what I understand xf86-input-* does this pretty well. I won't
unplug my mouse and so want to keep my xorg simple conf.
Hal does a lot more than just monitor /dev/input for you.
It's a framework for providing consolidated and consistent
access to *all* hardware information on your system.
Gnome's automounting, for example, relies on hal to let it
know when a new device was plugged in, what mount point its
on, what type of device it is, etc. The vendor database of
FDI files includes information about everything from
batteries to power management to keyboards to rf kill
switches.
And really, how much simpler an xorg.conf can you get than
by deleting 2/3 of it?
2) Anyway, I tried to make use of evdev instead of the *deprecated*
mouse and kbd drivers but...
3) evdev without hal replaced well my mouse driver (for the moment I
just replaced /dev/input/mice by /dev/input/event2 in the mouse section)
4) for the keyboard it's far less simple : if I switch to evdev, I
cannot define the Xkb{Variant,Model,..} in xorg.conf so :
stuck with the 'kbd' driver.
There's nothing wrong with continuing to use the old
drivers. Nothing about hal requires you to switch to
evdev... and nothing about evdev requires to you use hal :)
The various non-Linux OS's supported by Xorg won't even have
the proper kernel support for evdev, so the "old" keyboard
and mouse drivers will probably be around for a long time.
They are only "deprecated" in the sense that the Linux
generic input layer exists at all.
I believed gentoo users would be more sceptic when it comes to make a
new daemon mandatory ;)
Well, it's clearly not *mandatory* because you can just
turn it off with -hal :)
Having said that, hal is exactly the kind of thing I would
expect Gentoo users to flock to: its powerful, flexible,
extensible, configurable, and it's the new cutting-edge
stuff from the upstream vendors. Before it went offline,
the Gentoo wiki was easily the most informative place on the
web to find information about hal. I would have predicted
hal going mainsteam on Gentoo years ahead of Red Hat or
Debian.
Also, just for the record, hal isn't by any stretch of the
imagination a "new" daemon. Its been a USE option for
Gentoo's gnome-vfs package since Gnome 2.8, in 2004.
(Any advice to use evdev, define a keyboard layout,model,variant without
having to install hal and its r>ation>files and daemon ?)
Unless you have a general aversion to using XML for anything
(which I understand, if tend to disagree with), the FDI
syntax is pretty straightfoward. That's even without the
abundance of sample code that ships with hal. You basically
need to know two tags: match and merge. In this case:
type="string">keyboard
goes into /etc/hal/fdi/policy/keyboard_driver_ftw.fdi and
restart hal. Since hal lets you "merge" arbitrary keys to
its database, Xorg will also look for any
input.x11_options.foo keys, replacing everything that went
into xorg.conf.
---Mike
Re: [gentoo-user] rc-update add hald default or boot?
Mark Knecht wrote: Hi, I've wondered for a while whether the design intention of hald was to be run at boot time or as a default level process? Googling for It's intended that you'll be running it at the default level. Hal requires D-BUS, which means it needs to be run at the earliest after your local mounts, clock, etc. are started. If you put them both into the boot runlevel I don't see why it would necessarily hurt, but it doesn't need to be there to work properly. --Mike
Re: [gentoo-user] intelfb usage in grub
On 4/20/2009 2:47 PM, Valmor de Almeida wrote: but no luck (that is I get the 80x25 console). The only working format appears to be kernel /boot/vmlinuz root=/dev/sda3 vga=xxx This is the only correct syntax for the intelfb device, because of Apr 20 14:27:54 dcpl-lpt1 [ 0.173678] intelfb: Non-CRT device is enabled ( LVDS port ). Disabling mode switching. Apr 20 14:27:54 dcpl-lpt1 [ 0.173703] intelfb: Initial video mode is 1024x768...@60. Apr 20 14:27:54 dcpl-lpt1 [ 0.174157] intelfb: Changing the video mode is not supported. these log messages. I assume this is a laptop (or else your intelfb is very confused). For some reason that I don't fully understand, but assume is a good one, the intel fb device cannot change the video mode on a laptop display. The vga parameter actually gets the kernel to change the video mode much earlier, so the intelfb driver can keep using the same one. I typically include both vga= and video= lines in my grub command to get this to work. --Mike
Re: [gentoo-user] Synaptics and HAL Device Information Files
On 5/26/2009 5:58 AM, Redouane Boumghar wrote: First of all where can I find information about the file names of FDI ? NUMBER-NAME-NAME.fdi Where are the specification of the nomenclature ? I have found different names possible : 11-x11-synaptics.fdi 99-x11-synaptics.fdi > Why the donkey would it be 99 or 11 ? Since (as I see below) you've added "hal" to your USE flags for the xf86-input-synaptics driver, you should already have the synaptics HAL data installed -- the post-ebuild messages would tell you where those are. Currently, the stock HAL rules are installed into /usr/share/hal/fdi/policy/10osvendor/11-x11-synaptics.fdi The reason it's number 11 is because the synaptics Xorg driver exposes itself to HAL in a way that also looks like a mouse, so it matches all on the HAL rules for standard pointer device behavior. The synaptics rules need to override the normal rules, which are in the file 10-x11-pointer.fdi, so the synaptics rules go into 11-x11-synaptics.fdi and get loaded second. The file named 99- is most likely a suggestion from someone for a local customization to the policy, since it will be loaded after everything else. Since the synaptics rules are included in the base HAL policy now, there's no need for the 99- file. The thing is that I don't know how to match my touchpad with the fdi policy or that I have another unknown problem. You should not have to do anything to get your touchpad recognized by HAL as a synaptics device, since you already have the HAL policy file locally. You can use lshal(1) to ask HAL what devices it found, for example: lshal | grep -9 input.x11_driver You should see a result which includes: input.x11_driver = 'synaptics' If you back up a few lines you will see the info.capabilities set, which should include items like "input", "input.mouse", and input.touchpad. The default settings are in the FDI file in /usr/share, which will also show you how to override any of those settings. Basically, anything that used to be an xorg.conf option can be set using an "input.x11_options.OPTIONNAME" key. For example, to turn on SHMConfig so you can use the synaptics utilities: On Put this in an FDI file inside your /etc/hal/fdi/policy folder, such as /etc/hal/fdi/policy/10osvendor/11-x11-synaptics.fdi
Re: [gentoo-user] USE="mmx mmxext sse sse2 ssse3 3dnow 3dnowext"
On 5/27/2009 4:08 PM, Ward Poelmans wrote: On Wed, May 27, 2009 at 21:57, Wyatt Epp wrote: Reading this thread, though, it seems like it would be useful to have a FEATURES=cpudetection for Portage. I honestly don't understand why the user should have to be arsed to set those USEs manually if they don't actually care about moving binary packages around. Thoughts? W're gentoo users, we like being in control and don't like it when a program thinks for us. Even if this feature were to appear (and it seems to me like a complete pain in the ass to implement), no one is gonna force your fingers to type "FEATURES=cpudetection" into your make.conf file.
Re: [gentoo-user] USE="mmx mmxext sse sse2 ssse3 3dnow 3dnowext"
On 5/27/2009 4:40 PM, Wyatt Epp wrote: On Wed, May 27, 2009 at 4:09 PM, Alan McKinnon mailto:alan.mckin...@gmail.com>> wrote: On Wednesday 27 May 2009 22:04:06 Volker Armin Hemmann wrote: > > thoughts: there are distros that hold your hand already. There's also Windows. For just in case Ubuntu doesn't hold enough of your hand. Oh right, what was I thinking, giving the users an optional tool to make their lives marginally easier after they've tired of trying to remember for their thirtieth install just which of these inane x86 SIMD instructions are supported by cpu x? I'm rolling my eyes over here, guys; that was uncalled for and you know it. Clearly you haven't seen any of the Gentoo Installer threads.
Re: [gentoo-user] USE="mmx mmxext sse sse2 ssse3 3dnow 3dnowext"
Stroller writes: But, surely "-march=" also instructs gcc to support the additional instructions. Suggest you re-read Daniel's post that I was replying to. What's the difference between supporting the "certain set of instructions" with "-march=" and doing so with USEs? One is for telling gcc how to compile C code, the other is for telling packages what inline assembler code is legal. In the case of mplayer, just setting mmx/sse/etc does absolutely nothing. You also have to enable custom-cpuopts, which then tells mplayer to ignore its own build-time detection of CPU features and only use the ones you tell it. Either way, if mplayer does it automatically or you pass in USE flags, all they do is enable selected asm files in the codecs that use those opcodes. --K
Re: [gentoo-user] Can I get a tip for reading eix output?
On 6/15/2009 8:28 AM, Harry Putnam wrote: Can I get a likely search string for the massive man page of eix to understand all the info contained in its ouput. A quick search on `output' seems to miss it. The eix man page is way too long, and about 90% of it is only useful to the people trying to customize it beyond recognition, so it always takes me a while to find what I want to know, too. In this case, the information you want was snuck into the examples (!!!) in the "Slots" section of output. You'll notice that one of the examples has a series of !flags, while the description says the output indicates RESTRICT= options, and you get to extrapolate from there. So, !t means RESTRICT="test"
Re: [gentoo-user] Can I get a tip for reading eix output?
On Tuesday 16 June 2009 14:46:53 bn wrote: Alan McKinnon ha scritto: What do the `!t' entries following the versions available mean? RESTRICT="test" You can find a clue in sub-section "Slots" under main heading "OUTPUT" Which in turn, means? m. If you're asking what the meaning of the RESTRICT= options are, they're listed in the man page for ebuild(5), including: test do not run src_test even if user has FEATURES=test. which likely means that the package in question has a test suite that is known to be broken, on Gentoo or in general. (Side note: the man page only has one entry for PROPERTIES although eix knows of at least 4. Is there somewhere else this kind of portage information is documented?) --Mike
Re: [gentoo-user] gcc update failed...
On 7/6/2009 2:23 PM, Jarry wrote:
-I/var/tmp/portage/sys-devel/gcc-4.3.2-r3/work/gcc-4.3.2/gcc/../libdecnu
mber/bid -I../libdecnumber insn-recog.c -o insn-recog.o
{standard input}: Assembler messages:
{standard input}:36051: Warning: end of file not at end of a line;
newline inserted
{standard input}:37071: Error: bad register name `%e'
xgcc: Internal error: Killed (program cc1)
Please submit a full bug report.
See
Re: [gentoo-user] gcc update failed...
On 7/6/2009 4:23 PM, Jarry wrote: Mike Edenfield wrote: On 7/6/2009 2:23 PM, Jarry wrote: MAKEOPTS="-j2" Can you try this again without -j2 in the make opts? The gcc build process is enough of a pain to debug when you can see output sequentially, running parallel makes makes it worse. Now this is strange: I commented out that MAKEOPTS="-j2" in make.conf and gcc compilled without any complaint. I did not change anything else! Can't believe this. I will check it tomorrow again, with -j2 and without... I can't explain why this happens but I do know that I've seen a number of complex builds (especially gcc/glibc) that just randomly fail when using a parallel make. I see this effect on FreeBSD moreso than Gentoo, but I suspect that's only a question of scale. --Mike
[gentoo-user] emerge and running tests as non-root
In trying to merge the most recent mysql, I am getting this notice at the start of the ebuild: * Testing with FEATURES=-userpriv is no longer supported by upstream. Tests MUST be run as non-root. Is there a way to actually do what it suggests using just emerge, or would I need to ebuild this package in two steps (ebuild test, then ebuild merge as root)? --K
[gentoo-user] bash stopped running python scripts...
I dunno what I did, but I've managed to break python shell scripts, which of course is playing havoc with portage. Bash no longer wants to execute the scripts with python as the interpreter, but insists on executing them as bash scripts. Python itself is still functioning properly, when invoked directly. Since I happen to have ImageMagick on this machine, the result is this: kut...@apollo ~ $ cat test.py #!/usr/bin/python import sys print "Python Ok." kut...@apollo ~ $ ./test.py X connection to localhost:11.0 broken (explicit kill or server shutdown). ./test.py: line 3: print: command not found kut...@apollo ~ $ python ./test.py Python Ok. kut...@apollo ~ $ Perl and Tcl both still work as expected, and by invoking python directly I can still get portage to merge things. For example: r...@apollo ~ $ python -O /usr/lib/portage/bin/ebuild /usr/portage/dev-lang/python/python-2.6.2-r1.ebuild install will re-merge python successfully, which I have now done for both python and bash, to no effect. Any ideas? --Mike
Re: [gentoo-user] bash stopped running python scripts...
On 8/3/2009 5:03 PM, Alan McKinnon wrote:
On Monday 03 August 2009 22:56:51 Mike Edenfield wrote:
kut...@apollo ~ $ cat test.py
#!/usr/bin/python
import sys
print "Python Ok."
kut...@apollo ~ $ ./test.py
X connection to localhost:11.0 broken (explicit kill or server shutdown).
./test.py: line 3: print: command not found
kut...@apollo ~ $ python ./test.py
Python Ok.
kut...@apollo ~ $
Did you recently merge python-3 and were so foolish as to make it the
default?
I did emerge python-3, but then unmerged it almost immediately, and it
was never the default. Python was already broken when I merged python
3.1, which I did to see if it fixed anything, which of course it didn't.
What is /usr/bin/python? and what version is it (-V)?
r...@apollo ~ # /usr/bin/python -V
Python 2.6.2
r...@apollo ~ # cat /usr/bin/python
#!/bin/bash
# Gentoo Python wrapper script
[[ "${EPYTHON}" =~ (/|^python$) ]] && EPYTHON="python2.6"
"${0%/*}/${EPYTHON:-python2.6}" "$@"
Is that supposed to be that way? I vaguely recall from my Tcl days that
tclsh used to cause problems with the #! lines when it was a shell
script, and that you had to use some odd exec trick to get tcl shell
scripts to run. Is that still true?
Looking back through my emerge.log it appears that the last thing to
successfully run through emerge was eselect-python, if that makes a
difference.
--Mike
Re: [gentoo-user] Re: bash stopped running python scripts...
On 8/3/2009 5:14 PM, Remy Blank wrote: Mike Edenfield wrote: I dunno what I did, but I've managed to break python shell scripts, which of course is playing havoc with portage. http://bugs.gentoo.org/show_bug.cgi?id=279915 The whole issue seems to be handled quite strangely IMO. You would think breaking Python for all ~x86 is a major offense... Ah, thanks. Usually my google searches pick up stuff from bugs.gentoo.org but this time it didn't, guess it was too new :\ And yes, reverting /usr/bin/python to a symlink instead of a shell script does solve the problem. I'll just follow the bug for now and mask off eselect-python. --Mike
Re: [gentoo-user] bash stopped running python scripts...
On 8/3/2009 5:48 PM, Alan McKinnon wrote:
On Monday 03 August 2009 23:22:08 Mike Edenfield wrote:
On 8/3/2009 5:03 PM, Alan McKinnon wrote:
On Monday 03 August 2009 22:56:51 Mike Edenfield wrote:
kut...@apollo ~ $ cat test.py
#!/usr/bin/python
import sys
print "Python Ok."
kut...@apollo ~ $ ./test.py
X connection to localhost:11.0 broken (explicit kill or server
shutdown). ./test.py: line 3: print: command not found
kut...@apollo ~ $ python ./test.py
Python Ok.
kut...@apollo ~ $
Did you recently merge python-3 and were so foolish as to make it the
default?
I did emerge python-3, but then unmerged it almost immediately, and it
was never the default. Python was already broken when I merged python
3.1, which I did to see if it fixed anything, which of course it didn't.
What is /usr/bin/python? and what version is it (-V)?
r...@apollo ~ # /usr/bin/python -V
Python 2.6.2
r...@apollo ~ # cat /usr/bin/python
#!/bin/bash
# Gentoo Python wrapper script
[[ "${EPYTHON}" =~ (/|^python$) ]]&& EPYTHON="python2.6"
"${0%/*}/${EPYTHON:-python2.6}" "$@"
Is that supposed to be that way? I vaguely recall from my Tcl days that
tclsh used to cause problems with the #! lines when it was a shell
script, and that you had to use some odd exec trick to get tcl shell
scripts to run. Is that still true?
I have the identical file, it works here.
That's very odd, as everything I've read over the past hour indicates
that it's not *supposed* to work if you put a shell script in that line,
but that's clearly not always true.
What's the #! line in your /usr/bin/emerge file?
Your original post has a "X connection to localhost:11.0 broken" error,
which is mighty unusual. The error is common enough, but has nothing to do
with python.
I know where that particular error is coming from. The first
non-comment line in /usr/bin/emerge is "import sys". Since bash is
interpreting the script instead of python, when it gets to that line, it
runs ImageMagick's "import" command. It just indicates that bash
ignored the #! line and kept going.
The bug Remy sent me to did have two working solutions: replacing the
shell script with a symlink, or replacing "#!/usr/bin/python" with
"#!/usr/bin/env python" in emerge/ebuild, both of which point the #!
line at a real executable. I'm now curious why that shell script works
for anyone.
--Mike
Re: [gentoo-user] Re: bash stopped running python scripts...
On 8/4/2009 7:13 AM, Graham Murray wrote: Remy Blank writes: The whole issue seems to be handled quite strangely IMO. You would think breaking Python for all ~x86 is a major offense... It did not break for all ~x86. I have 2 systems both running ~x86, both have emerged (but not made active) python-3.1, /usr/bin/python is a bash script on both yet emerge works with no problems. The only non-standard things are that I have unmasked gcc-4.4.1 and am using git 2.6.31-rc kernels. Aha. I think I've discovered the "problem". There was a change in the kernel execl() call between 2.6.27 and 2.6.28: --- linux-2.6.27-hardened-r3/fs/binfmt_script.c 2008-10-09 18:13:53.0 -0400 +++ linux-2.6.28-hardened/fs/binfmt_script.c2008-12-24 18:26:37.0 -0500 @@ -22,14 +22,15 @@ char interp[BINPRM_BUF_SIZE]; int retval; - if ((bprm->buf[0] != '#') || (bprm->buf[1] != '!') || (bprm->sh_bang)) + if ((bprm->buf[0] != '#') || (bprm->buf[1] != '!') || + (bprm->recursion_depth > BINPRM_MAX_RECURSION)) return -ENOEXEC; /* * This section does the #! interpretation. * Sorta complicated, but hopefully it will work. -TYT */ - bprm->sh_bang = 1; + bprm->recursion_depth++; allow_write_access(bprm->file); fput(bprm->file); bprm->file = NULL; The kernel >= 2.6.28 now supports nesting up to 4 levels of script in the #! lines, if I'm reading that right, whereas < 2.6.28 it only supported 1 level. I'll go update the b.g.o entry and try upgrading my kernel. Though I dunno what that means for Gentoo/FreeBSD. --Mike
Re: [gentoo-user] Knock on wood
On 8/12/2009 4:19 PM, Stroller wrote: On 12 Aug 2009, at 15:20, Dale wrote: ... maske install does that for you, it also sets up the vmlinuz and vmlinuz.old symlinks so you don't need to mess with your GRUB config. But it doesn't do it the way that I do. I have used it a few times but it didn't work like I do manually. +1 I don't know why anyone would choose to compile & install the kernel any way other than manually. It's only a handful of commands, after all. $ make && make modules_install $ cp arch/x86/boot/bzImage /boot/vmlinuz-2.6.28-hardened-r1-wombat-4 $ cp .config /boot/config-2.6.28-hardened-r1-wombat-4 $ cp System.map /boot/System.map-2.6.28-hardened-r1-wombat-4 $ cd /boot $ mv vmlinuz vmlinuz.old $ mv System.map System.map.old $ mv config config.old $ ln -sf vmlinuz-2.6.28-hardened-r1-wombat-4 vmlinuz $ ln -sf config-2.6.28-hardened-r1-wombat-4 config $ ln -sf System.map-2.6.28-hardened-r1-wombat-4 System.map vs. $ make && make modules_install && make install I can't see any reason *not* to use make install, especially when its just doing exactly the same thing you are doing manually with about 100 times as much typing. --K
Re: [gentoo-user] Knock on wood
On 8/12/2009 5:08 PM, Volker Armin Hemmann wrote: On Mittwoch 12 August 2009, Mike Edenfield wrote: $ make&& make modules_install&& make install too much to type. make all modules_install install is much better. I always forget that the 'all' target (typically) does the same thing as just typing 'make'. --K
Re: [gentoo-user] lm_sensors: dependency problem...
On 8/23/2009 8:45 AM, Jarry wrote: Albert Hopkins wrote: On Sun, 2009-08-23 at 09:42 +0200, Jarry wrote: # emerge --pretend lm_sensors These are the packages that would be merged, in order: Calculating dependencies... done! [ebuild N ] sys-fs/sysfsutils-2.1.0 [ebuild N ] x11-misc/read-edid-1.4.2 [ebuild N ] sys-apps/lm_sensors-2.10.7 USE="-sensord" What is that x11-misc/read-edid good for? I do not want to pull the whole x11 stuff with it, as this is a server with no graphics. I agree with the original poster of this "resolved" bug: pulling in x11-stuff to server just to be able to monitor its temperatures is simply dangerous and risky... read-edid is only tangentially related to X. It doesn't use any X libraries, or even require X to be installed to run. It perhaps shouldn't even be in the x11-misc category, but that's entirely cosmetic. Also, I noted that lm_sensors-3.0 doesn't have this dependency anymore, so it looks like it was handled upstream. Next, I want to use sensord for monitoring, so I modified package.use and tried it once more: # echo "sys-apps/lm_sensors sensord" >> /etc/portage/package.use # emerge --pretend lm_sensors These are the packages that would be merged, in order: Calculating dependencies... done! emerge: there are no ebuilds built with USE flags to satisfy ">=x11-libs/cairo-1.4.6[svg]". !!! One of the following packages is required to complete your request: - x11-libs/cairo-1.8.6-r1 (Change USE: +svg) (dependency required by "net-analyzer/rrdtool-1.3.8" [ebuild]) (dependency required by "sys-apps/lm_sensors-2.10.7" [ebuild]) (dependency required by "lm_sensors" [argument]) Now I do not understand it at all. Why rrdtool? All I want to use sensord for is to have some status messages recorded to syslog. Why should I install that x11-libs/cairo graphics library and rrdtool stuff??? Likely sensord does a lot more than recording to syslog. man sensord: DESCRIPTION Sensord is a daemon that can be used to periodically log sensor readings from hardware health-monitoring chips to syslog(3) ***OR*** a round-robin database (RRD) and to alert when a sensor alarm is signalled; for example, if a fan fails, a temperature limit is exceeded, etc. You see that big fat ***OR*** there? syslog or RRD. And I think even gentoo-user should have the right to choose what he wants to do with those sensor-readings. Definitelly I'm against any x11-libs on my server, be it x11-misc/read-edid or x11-libs/cairo. BTW, sensord can write data to round-robin database even if no rrd-tool is installed. There's no need to put net-analyzer/rrdtool as a dependency... Tell that to the upstream: make: *** No rule to make target `rrd.h', needed by `prog/sensord/rrd.rd'. Stop. --Mike
Re: [gentoo-user] [OT] filesystems
Dale wrote: I'm not expecting a answer but along the lines of a viewpoint in a question form. Why is it that smart, I mean seriously smart, people have the worst social skills? They can invent a super fast CPU, memory http://en.wikipedia.org/wiki/Asperger_syndrome
Re: [gentoo-user] locale issue to clean up
Mark Knecht wrote: lightning ~ # cat /etc/locale.gen en_US ISO-8859-1 en_US.UTF-8 UTF-8 Just to be safe, try running locale-gen again. The glibc ebuild does this automatically, but if you've changed locale.gen since the last time that ebuild ran, you need to run locale-gen to pick up the changes. lightning ~ # cat /etc/env.d/02locale LANG=en_US I think that adding LC_ALL="en_US" in this file should be all you need. Basically, your "locale" is made up of about a dozen different settings that describe how to display things like dates, money, big numbers, sorted lists, etc. Each of these options has an LC_* variable associated with it. Additionally, there is a LANG variable that specifies which language translation to use (LANGUAGES is a GNU-specific alternative to LC_MESSAGES, IIRC, and I rarely specify it). LC_ALL is used if you know you want the use one locale for everything, and specifying LC_ALL="en_US" will automatically set all of the others for you. Setting that end rerunning env-update ; source /etc/profile should get rid of those locale errors. Looking above I don't have LANG or LC_ALL entries and the LINGUAS is suspect. If I was to convert them to your settings do I then need to rebuild any apps that use them? You don't need LANG/LC_ALL in your make.conf if they are set properly in your environment, e.g. in env.d/02locale, since they are both run-time environment settings that you want to be the same everywhere. LINGUAS is a build-time setting that tells any localization-aware autoconf scripts to install the language translation files for the set of languages you listed. Having LINGUAS="en en_US" in make.conf is correct, it makes sure that only English-language translations are installed. (Not all applications are LINGUAS-aware, so you will still see a lot of other languages you didn't ask for.).
Re: [gentoo-user] locale issue to clean up
Mark Knecht wrote: lightning ~ # locale -a locale: Cannot set LC_CTYPE to default locale: No such file or directory locale: Cannot set LC_MESSAGES to default locale: No such file or directory locale: Cannot set LC_COLLATE to default locale: No such file or directory C POSIX This looks like the problem: you have specified that you want to use en_US as your locale, but you don't have that locale installed. You should have a directory in /usr/share/locale for every locale you want available on your system. The source files for the locales should be in /usr/share/i18n/locales and /usr/share/i18n/charsets. That is, you should have all of the following: /usr/share/i18n/locales/en_US /usr/share/i18n/charsets/ISO8859-1 /usr/share/i18n/charsets/UTF-8 /usr/share/locale/en_US.ISO8859-1 /usr/share/locale/en_US.UTF-8 If you are missing the last two, rerunning locale-gen should create them. If you're missing any of the first three, then you will probably need to emerge -1 glibc to get everything back. --Mike
Re: [gentoo-user] locale issue to clean up
Dale wrote: I'm not sure if the -N feature will catch those changes or not. It may even depend on the version of portage you are using too. The newer --newuse will pick up changes to LINGUAS since portage treats that like an expandable variable (like VIDEO_CARDS etc). The other settings have no direct effect on the USE flags, so portage won't notice anything. You also don't need to reinstall an application to switch locales, as long as the language packs for your new locale are installed via LINGUAS and/or by default. --Mike
Re: [gentoo-user] locale issue to clean up
smallnow wrote: Mike Edenfield wrote: Um, on my system, i have /usr/share/i18n/charmaps/UTF-8.gz /usr/share/i18n/charmaps/IS8859-1.gz notice charmaps vs charsets the other folders all have en_US files and folders, no utf8 extensions. And my locale stuff seems to work fine. Do you actually have those files on your computer or did you just type them from memory and get them wrong? I was working from memory, and yes, I meant charmaps instead of charsets (I've seen both names used, but it's the same data.)
