Re: [gentoo-user] terminal spreadsheet - sc fork

2014-11-03 Thread Wang Xuerui
2014-11-03 7:16 GMT+08:00 Andrés Martinelli andma...@gmail.com:
 Hello there!!
 I am working on a terminal spreadsheet based on sc, but with some adds
 like undo/redo..
 you can find it here:

 https://github.com/andmarti1424/scim

 Any new ideas and/or contribution is always welcome!
 Thanks!

FYI: the package name collides with an IME framework (albeit largely
displaced by ibus and fcitx in recent years). I, as a Chinese user,
briefly wondered how you can transform an IME library into a
spreadsheet app! :-D



Re: [gentoo-user] Automounting USB drives

2014-08-06 Thread Wang Xuerui
2014-08-06 9:18 GMT+08:00 Chris Stankevitz chrisstankev...@gmail.com:
 [snip]
 10. Bonus: if you use words like COM/DDE/OLE

Just a side note... These 3 things don't play well with a Linux
ecosystem, as you might know. They're M$ technologies after all (-:

(actually they just don't exist in a native Linux install without
Wine, but AFAIK Wine doesn't participate in the automounting anyway)



Re: [gentoo-user] Kernel boot messages are no longer displayed

2014-05-25 Thread Wang Xuerui
2014-05-25 4:20 GMT+08:00 Hilco Wijbenga hilco.wijbe...@gmail.com:
 Hi all,

 Since kernel 3.12.13 (3.12.13-gentoo), the kernel boot messages have
 disappeared, i.e. they are no longer displayed at boot time. All I get
 is a line like Loading kernel 3.12.13. (I just upgraded to
 3.12.20-gentoo, so now it's something like Loading kernel 3.12.20.)

 I have no idea why. I checked my grub config and it does not add a
 quiet argument. I see nothing in /etc/rc.conf about quiet (or
 verbose). And X86_VERBOSE_BOOTUP is set to 'y'. I also looked for
 quiet, boot, and messages (while running make xconfig) and
 nothing untoward showed up.

 How do I get back to a normal and sane boot process?

Hi, I also experienced this change earlier this year, and it seems the
old behavior can be restored by setting loglevel=7 in kernel cmdline.
Actually, not printing that much message can lead to a small
performance gain, as some console drivers tend to not be very fast;
but you certainly can have the messages if you want.

Hope that helps~



Re: [gentoo-user] How to appoint python version for a package

2014-04-02 Thread Wang Xuerui
2014-04-02 19:56 GMT+08:00 林守磊 linxiu...@gmail.com:
 tell me why python3.2 case problem below


 return u'[%s-%s]' % (unichr(a), unichr(b))
 ^
 SyntaxError: invalid syntax


Well, it seems to be an upstream bug, as Python 3.2 does not recognize
the old (Python 2.x) style for Unicode literals. Python 3.3 adds the
syntax back, so the program will work in Python 3.3 but not 3.2. You
definitely should report this to upstream; it won't take much time to
fix bugs of this sort. :)



Re: [gentoo-user] How to appoint python version for a package

2014-04-02 Thread Wang Xuerui
2014-04-02 20:26 GMT+08:00 Wang Xuerui idontknw.w...@gmail.com:
 2014-04-02 19:56 GMT+08:00 林守磊 linxiu...@gmail.com:
 return u'[%s-%s]' % (unichr(a), unichr(b))
 [snip]
 so the program will work in Python 3.3 but not 3.2

Oops, there is also unichr. Seems the program is Python 2.x only, in
which case you can try using package.env (google it) to override the
PYTHON_TARGETS variable (or really, any variable in the build
environment). Reporting to upstream is also welcomed, though.

Sorry for the noise~



Re: [gentoo-user] How to appoint python version for a package

2014-04-02 Thread Wang Xuerui
2014-04-03 13:15 GMT+08:00 林守磊 linxiu...@gmail.com:
 @all

 I found that is a bug from package pygments-1.6_p20140324, and I downgrade
 to 1.6-r1. problem solved !

Oh, I overlooked the name of the offending file. Yes it's pygments
that's at fault, not retext :)



Re: [gentoo-user] out of disk space to compile webkit-gtk

2013-12-23 Thread Wang Xuerui
在 2013-12-23 下午3:54,Joseph syscon...@gmail.com写道:

 I'm upgrading the system and running out of disk space to compile
webkit-gtk

 How do I tell the system to use /home partition to use for compiling temp
tiles?

For such a change to apply only to webkit-gtk (or any other package you
want, actually), you can try package.env. Checkout make.conf's manpages to
see how it is used.

Hope that helps!


Re: [gentoo-user] Switching off a long list of USE flags

2013-09-19 Thread Wang Xuerui
2013/9/19 gevisz gev...@gmail.com:
 I would like to build sane-backend only for one specific scanner.

 However, there are a huge lot of different USE flags for other scanners
 that are already switched on by default.

 Is there any easy way to switch them off?

 Something like:

   media-gfx/sane-backends -sane_backends* sane_backends_hp

 Exactly that way does not work, but may be something like that...

 ...instead of directly switching off every one of them.

According to the Package Manager Specification, the variable should
not be incremental, i.e. any value set would override *any* defaults
or previously set values. So just do SANE_BACKENDS=xxx in your
make.conf, same as what you'd do for VIDEO_CARDS and the like.

Hope it helps :)



Re: [gentoo-user] bios+gpt+grub2 boot windows

2013-09-10 Thread Wang Xuerui
2013/9/10 东方巽雷 dongfangxun...@gmail.com:
 I format my hard disk to gpt.Then create a 1M partition at the header of
 hard disk ,add bios_grub flag.I can install and boot gentoo,debain and
 archlinux,but have no idea how to boot windows7 and 8.When I boot win7 or
 win8 from usb,it always complain the gpt partition and cannot install.What
 should I do?

It's impossible to install Windows into a GPT partition if the
firmware isn't EFI. This is a limitation artificially set by
Microsoft; here's Microsoft's FAQ on this.
http://msdn.microsoft.com/en-us/library/windows/hardware/gg463525.aspx

In this case, you must either revert to an msdos partition table (can
be done w/o data loss IIRC), or trick Windows into believing it's
running on EFI, i.e. set up UEFI DUET. Google that and you'll find an
excellent HOWTO on this. Happy hacking~



Re: [gentoo-user] bios+gpt+grub2 boot windows

2013-09-10 Thread Wang Xuerui
2013/9/11 Yuri K. Shatroff yks-...@yandex.ru:
 Well, actually it *is* possible using hybrid MBR+GPT partitioning.

 http://www.rodsbooks.com/gdisk/hybrid.html

 Also google for 'hybrid mbr'.

 I am using this setup for multibooting several OS incl. Win 7 without
 problems for a couple of years.

I almost forgot about hybrid MBRs... also it's fragile indeed, so I
probably won't recommend it even if I thought about it in the first
place.

 In this case, you must either revert to an msdos partition table (can
 be done w/o data loss IIRC), or trick Windows into believing it's
 running on EFI, i.e. set up UEFI DUET. Google that and you'll find an
 excellent HOWTO on this. Happy hacking~


 An interesting idea, yet, probably, a little more complicated than hybrid
 MBR for the sole purpose of multibooting.

I've tried that on a VirtualBox VM last year, and I'd like to point
out that describing the process as a little more complicated is
serious underestimation. :D



Re: [gentoo-user] Where is the file which masks glibc-2.18?

2013-09-05 Thread Wang Xuerui
2013/9/5 东方巽雷 dongfangxun...@gmail.com:
 I cannot find glibc-2.18 in /usr/portage/profile/package.mask,but glibc-2.18
 is being masked.I would like to know which file masks glibc-2.18.

Isn't it KEYWORD=? I noticed its ebuild's KEYWORDS line is commented out.



Re: [gentoo-user] grub2 or kernel config - unable to properly boot

2013-08-27 Thread Wang Xuerui
2013/8/27 Francisco Ares fra...@gmail.com:
 In regard of file systems,
 it only loads a ext2 module, and the root partition is ext4 formated.

 Although this kernel has ext2, ext3 and ext4 built in (not modules), can
 this be a cue?  Right now I am preparing to format the root partition as
 ext2, just to check this out.

Well, GRUB modules are *GRUB* modules, that is, they're there only for
GRUB to be able to understand your partition table and read your
filesystems. After successfully reading the kernel into memory and
passing control to it, they're not relevant any more, so you really
don't have to reformat /.

Instead, focus on your initramfs as the error shows some inconsistency
between the expected and actual initramfs content. Also that last line
seems to come from initramfs, based on its appearance (unlike dmesg
lines). You may understand your problem better there.

Hope that helps.



Re: [gentoo-user] How can I unsubscribe from gentoo-news?

2013-08-27 Thread Wang Xuerui
在 2013-8-28 上午4:03,Alan McKinnon alan.mckin...@gmail.com写道:

 On 27/08/2013 18:35, Jarry wrote:
  Hi Gentoo-users,
 
  I'd like to stop receiving news, how can I do that?
 
  Or better said, I'd like to receive gentoo-news only
  on one of my 6 servers, and turn this feature off
  on remaining 5...
 
  Jarry



 It's in the mail headers:

 mailto:gentoo-user+unsubscr...@lists.gentoo.org

 you would obviously use -news and not -user


 Normal practice (at least with Mailman) is to have such links in the
 email footer. I see Gentoo lists don't do that; I believe because they
 run Encartis?

 --
 Alan McKinnon
 alan.mckin...@gmail.com



Ah, I think he actually means those news items as seen in eselect news
read, as some keywords like server and turn off suggest...


Re: [gentoo-user] how can I pause emerge after it finish running configure or cmake and before it do any compilation?

2013-08-19 Thread Wang Xuerui
2013/8/19 东方巽雷 dongfangxun...@gmail.com:
 good!  Thank you.

Manually executing phases of ebuild is certainly not a long-term
solution; changes won't persist on the next build.

You'd better patch the ebuild's src_configure section (before you
forget your exact modifications), then put them into your overlay; If
you don't have an overlay already, it's a good opportunity to create
one.



Re: [gentoo-user] libreoffice cannot use oxygen theme

2013-08-19 Thread Wang Xuerui
2013/8/19 Alan McKinnon alan.mckin...@gmail.com:
 NOT upgrading icu on a whim also comes with massive user benefits:

 such as, for example, NOT having to rebuild every damn huge piece of
 software on the box every other week just coz icu decided to change how
 something is done and shove it into a point release. Again.

 /rant over

Sure...

And if you choose to not rebuild things you end up putting a whole lot
of packages into @preserved-rebuild, which shows up *every* time
emerge completes. Well, I basically get over this by simply ignoring
them, but the length of @preserved-rebuild output usually forces me
into emerging it every 2 weeks or so.

And just consider how much CPU cycles are lost in frequent qtwebkit,
webkit-gtk and chromium rebuilds...



Re: [gentoo-user] libreoffice cannot use oxygen theme

2013-08-19 Thread Wang Xuerui
2013/8/19 Alan McKinnon alan.mckin...@gmail.com:
 webkit* updates have slowed down recently so I'm not too bothered about
 that. But chromium, that thing drive me to tears, now I just use
 google-chrome.

Well, I choose to temporarily mask chromium if I don't have the time.
I have a /etc/portage/package.mask/11temp for (mainly) that (-:

I've switched yesterday from firefox to firefox-bin after some
mysterious crashes in the graphic backend (with HW acceleration force
enabled), only to find my compiler is not at fault... But the binary
version indeed fixed another bug, that's the page loading spinner
not animating, so I'd stick with that.

I don't have to use google-chrome because chromium is usually *still*
masked when I feel like upgrading, so it actually doesn't cost me much
time. In fact, I usually go to sleep right after starting emerge (-:

 I don't always need the latest icu - 7-bit ASCII satisfies all my daily
 needs :-) So I can afford to mask anything after the current version and
 just leave it as-is until it suits me to rebuild things.

As a Chinese, living with Unicode support is a must. But since CJK
codepoints are well supported for ages, and most of Chinese users are
not researchers who need access to the latest glyphs or complex-layout
writing systems, we don't have to upgrade icu that often either.

Actually, AFAICT all my icu upgrades are forced on me by the version
bump of big things like chromium or libreoffice.



Re: [gentoo-user] libreoffice cannot use oxygen theme

2013-08-19 Thread Wang Xuerui
2013/8/18 东方巽雷 dongfangxun...@gmail.com:
 I download LibreOffice_4.1.0_Linux_x86-64_deb.tar.gz from
 https://www.libreoffice.org/download/ and extract all the deb files.My
 desktop is KDE,but libreoffice only uses gtk+ theme.
 How should I find out the problem?

Oh... upon further reading of your question it seems we're all misled
by your description.

Your real question is Why a GTK application won't pick up KDE's theme
settings, and the answer is it simply doesn't know about Qt in the
first place. I downloaded the archive you used and ldd'd all the
libraries, none seems to depend on Qt. Gentoo's libreoffice-bin
packages are build by Gentoo developers (I remember a blog post
explaining this), so they're different from the upstream provided
ones. I haven't used them, so I can't say if they would help, but it's
certain that the official builds does NOT support Qt widgets.

Solution: Install oxygen-gtk for Oxygen-looking GTK2 widgets.



Re: [gentoo-user] libreoffice cannot use oxygen theme

2013-08-18 Thread Wang Xuerui
2013/8/19 东方巽雷 dongfangxun...@gmail.com:
 libreoffice-bin needs older icu version,libreoffice need so much time to
 compile

libreoffice-bin (at this time only 4.0.4.2 is present) requests
icu/51.1, while the latest is slotted 51.2. I don't know of any
package that specifically asks for such a new version of icu; a quick
equery on my system shows this:

equery depends icu
 * These packages depend on icu:
app-i18n/fcitx-4.2.8.1 (icu ? dev-libs/icu)
app-office/libreoffice-4.1.0.4 (=dev-libs/icu-4.8.1.1)
app-text/libmspub-0.0.6 (dev-libs/icu)
app-text/texlive-core-2013 (xetex ? =dev-libs/icu-50)
dev-db/sqlite-3.7.17 (icu ? dev-libs/icu)
dev-lang/php-5.4.18 (intl ? dev-libs/icu)
dev-lang/php-5.5.1-r1 (intl ? dev-libs/icu)
dev-libs/boost-1.53.0 (icu ? =dev-libs/icu-3.6)
dev-libs/libxml2-2.9.1-r1 (icu ? dev-libs/icu)
dev-qt/qtcore-4.8.5 (icu ? =dev-libs/icu-49)
dev-qt/qtwebkit-4.8.5 (icu ? dev-libs/icu)
dev-tex/bibtexu-3.71_p20130530 (=dev-libs/icu-4.4)
media-libs/harfbuzz-0.9.18-r1 (icu ? dev-libs/icu)
media-libs/libcdr-0.0.14 (dev-libs/icu)
media-libs/libvisio-0.0.30 (dev-libs/icu)
media-libs/raptor-2.0.9 (unicode ? dev-libs/icu)
net-libs/webkit-gtk-1.8.3-r201 (=dev-libs/icu-3.8.1-r1)
net-libs/webkit-gtk-2.0.4 (=dev-libs/icu-3.8.1-r1)
net-nds/openldap-2.4.35 (icu ? dev-libs/icu)
sys-apps/gptfdisk-0.8.6 (icu ? dev-libs/icu)
(icu ? dev-libs/icu[static-libs(+)])
www-client/chromium-29.0.1547.41 (=dev-libs/icu-49.1.1-r1)

So at least nothing installed on my system requires a recent icu to
run. Thus, you shouldn't have problems setting up libreoffice-bin; if
you indeed have to stick with the latest icu, share with us your
specific setup and we'll be glad to help.



Re: [gentoo-user] What (free) remote desktop do you use

2013-08-15 Thread Wang Xuerui
2013/8/15 Helmut Jarausch jarau...@igpm.rwth-aachen.de:
 Hi,

 some recent updates of X11, gtk++ or glib has broken the possibility to use
 some gnome3 applications (like a recent balsa)
 via nxclient.

 Since NX3's internal X-server has some known deficiencies and Nomachine has
 stopped the development of NX3 and
 has closed sources of NX4, I am in need for some replacement since   'ssh
 -Y' is dead slow even on an 15 megabit/sec connection.

 What remote desktop do you use?

 Many thanks for your hints,
 Helmut


Hello, you can try adding -C (enable gzip compression) to ssh
commandline first to see if the speed is acceptable. I once used it on
a 10Mbps LAN connection, and the decrease in bandwidth consumption is
impressive while not causing too much CPU load.



Re: [gentoo-user] How to determine 'Least Common Denominator' between Xen(Server) Hosts

2013-08-14 Thread Wang Xuerui
2013/8/14 Helmut Jarausch jarau...@igpm.rwth-aachen.de:
 Why not compute it yourself?

 Do   cat /proc/cpuinfo on all machines and compute the intersection of all
 flags
 Then enter something like the following into /etc/portage/make.conf

 CFLAGS=-O3 -pipe -msse -msse2 -msse3 -msse4a -m3dnow


Or even better, directly intersect the GCC flags which can be obtained
in this way:

echo | gcc -O2 -pipe -march=native -E -v - 21 | grep cc1

Also, according to the Gentoo Handbook going -O3 for the entire system
may cause instability and other problems. Has the situation changed
over the years?



Re: [gentoo-user] hostapd2.0 not work

2013-08-14 Thread Wang Xuerui
2013/8/14 东方巽雷 dongfangxun...@gmail.com:
 Hello everyone!hostapd1.1 work well in my system,but when I upgrade to
 2.0,it doesn't work any more.And I have found that 1.1 is no longer in the
 gentoo portage.What should I do?

Simple, grab a copy of the 1.1 ebuild and its accompanying files (if
necessary) from Internet and put them into a local overlay. Then mask
the =2.0 versions in your /etc/portage/package.mask/xxx (create an
appropriately named file there, you may have to mkdir first).



Re: [gentoo-user] about LIBRARY_PATH

2013-08-11 Thread Wang Xuerui
2013/8/12 东方巽雷 dongfangxun...@gmail.com:
 It seems that this variable is hard-code by gcc.I cannot change it any
 more.When I use gcc -m32 to compile a 32bit program,gcc is looking for
 /usr/lib rather than /usr/lib32.But in my system,/usr/lib is a symlink to
 /usr/lib64.The real 32bit librarys is in /usr/lib32.The linker is always
 complaining about i386:x86-64 architecture of input file
 `/usr/lib/gcc/x86_64-pc-linux-gnux32/4.7.3/../../../../lib/crt1.o' is
 incompatible with i386 output..Why does it not search /usr/lib32?

Well, you can always emerge crossdev and crossdev i686-pc-linux-gnu then... (-:

PS: A quick check on my system shows this:

# [08/12 13:27:52] xenon@ribosome ~ $
gcc -m32 -v --version 21 | grep LIBRARY_PATH
LIBRARY_PATH=/usr/lib/gcc/x86_64-pc-linux-gnu/4.7.3/32/:/usr/lib/gcc/x86_64-pc-linux-gnu/4.7.3/../../../../lib32/:/lib/../lib32/:/usr/lib/../lib32/:/usr/lib/gcc/x86_64-pc-linux-gnu/4.7.3/:/usr/lib/gcc/x86_64-pc-linux-gnu/4.7.3/../../../../x86_64-pc-linux-gnu/lib/:/usr/lib/gcc/x86_64-pc-linux-gnu/4.7.3/../../../:/lib/:/usr/lib/
# [08/12 13:28:00] xenon@ribosome ~ $
gcc -v --version 21 | grep LIBRARY_PATH
LIBRARY_PATH=/usr/lib/gcc/x86_64-pc-linux-gnu/4.7.3/:/usr/lib/gcc/x86_64-pc-linux-gnu/4.7.3/../../../../lib64/:/lib/../lib64/:/usr/lib/../lib64/:/usr/lib/gcc/x86_64-pc-linux-gnu/4.7.3/../../../../x86_64-pc-linux-gnu/lib/:/usr/lib/gcc/x86_64-pc-linux-gnu/4.7.3/../../../:/lib/:/usr/lib/

So the lib32 paths are actually included in the library search path,
and correctly placed *before* the lib counterparts. I'm not sure how
your toolchain ends up in the state you described, but it's certainly
weird... Can you please provide the output of emerge --info and gcc
-v --version? This way people will have more clue to help you...



Re: [gentoo-user] Re: Do I really need 194 pkgs to install git?

2013-08-03 Thread Wang Xuerui
2013/8/3 Harry Putnam rea...@newsguy.com:
 Hehe, alright, now we're talking that reduced dependancies to just
 1 lonesome cpio.

 Now, is it reasonable to install that way?  Will I run into some
 horrible unsightly mess using git, when installed this way.

IIRC you only have to do USE=-perl and most dependencies will be
gone... I distantly recall the last time I installed git on one of my
server nodes most packages pulled in was in the form virtual/perl-*
and perl-*/*, i.e. CPAN packages.

However, according to the ebuild, you need USE=perl set if you want
to enable subversion support or something funnier like CGI, so you'd
probably have to consider this whole thing a little bit more before
drawing your conclusion...



Re: [gentoo-user] how to tell you are using systemd?

2013-07-31 Thread Wang Xuerui
在 2013-8-1 上午10:26, cov...@ccs.covici.com写道:

 Can a shell script tell if systemd is the init?  I have a couple of
 places where it would be nice to know this.

 Thanks in advance for any suggestions.

Check /proc/1/comm or something like that, IIRC...


Re: [gentoo-user] Question about qemu QEMU_SOFTMMU_TARGETS and QEMU_USER_TARGETS

2013-07-22 Thread Wang Xuerui
2013/7/22 Walter Dnes waltd...@waltdnes.org:
   I'm usually pretty good a Google, but I've run into a brick wall with
 qemu's QEMU_SOFTMMU_TARGETS and QEMU_USER_TARGETS settings.  I find that
 wine on a 64-bit-only machine does not support 32-bit Windows programs.
 Years ago, I was able to build a 32-bit qemu Gentoo guest, and run wine
 32-bit mode on that.  I need to try it again, but I have no clue what
 QEMU_SOFTMMU_TARGETS and QEMU_USER_TARGETS settings to use.  I repeat,
 I'm on a 64-bit Intel machine, and I want to emulate Intel 32-bit.  Do
 these variables refer to the guest architecture or the host
 architecture?

   In plain English, given host and guest architectures which of the
 following combinations do I use...

 QEMU_SOFTMMU_TARGETS=host  QEMU_USER_TARGETS=host
 QEMU_SOFTMMU_TARGETS=host  QEMU_USER_TARGETS=guest
 QEMU_SOFTMMU_TARGETS=guest QEMU_USER_TARGETS=host
 QEMU_SOFTMMU_TARGETS=guest QEMU_USER_TARGETS=guest

Well, if all you want to do is run a 32-bit Wine on an amd64 box, full
hardware virtualization is not needed. According to this answer:

http://askubuntu.com/a/231605

All you need to do is ensure the proper WINEARCH and WINEPREFIX is
used when invoking the 32-bit program. Wine must be built with
ABI_X86=64 32, though.

As for the QEMU_{SOFTMMU,USER}_TARGETS variables, those USER targets
are for user-space emulation only, i.e. used to execute Linux binaries
from a different architecture, thus not what you need. Something like
this in make.conf may suffice:

QEMU_SOFTMMU_TARGETS=i386 x86_64  # only enable x86 and x86_64
QEMU_USER_TARGETS=  # prevent user-space emulation from being built

Note the variables are not expanded, so an empty value is used to
clear QEMU_USER_TARGETS instead of -*, as you might do otherwise.



Re: [gentoo-user] Networkmanager and gnome-3.6 (negating a default USE)

2013-07-16 Thread Wang Xuerui
在 2013-7-16 下午9:50, gottl...@nyu.edu写道:

 (This is not a msg about networkmanager's quality.)

 My main system is ~amd64 and runs gnome-3.6.

 I wanted to experiment with removing networkmanager.
 I do not have networkmanager in world and did not have it in make.conf.
 However gnome-3.6 by default pulls it in.  For example
 gnome-control-center-3.6.3-r1.ebuild contains

   IUSE=+bluetooth +colord +cups +gnome-online-accounts +i18n
   input_devices_wacom kerberos +networkmanager +socialweb systemd v4l

 and

   COMMON_DEPEND=
 ...
 networkmanager? (
 =gnome-extra/nm-applet-0.9.1.90
 =net-misc/networkmanager-0.8.997 )

 I believe the first says that if the networkmanager use flag is neither
 set nor unset, treat it as set and the second says
 if the use flag is set (or treated as set), require
 net-misc/networkmanager.

 I don't see how to have the flag unset since its default is set.
 For example, editing make.conf to add

   USE=-networkmanager

 (I had no USE in make.conf previously) has no effect
 since networkmanager was not in USE before.

 How do I specify that the networkmanager USE flag is explicitly unset
 (rather than just not explicitly set).

 thanks,
 allan


Just write a package.use file to handle that.


Re: [gentoo-user] pop up windows with text message

2013-07-10 Thread Wang Xuerui
2013/7/11 Joseph syscon...@gmail.com:
 I just want to monitor one file in a directory.
 I usually copy the file over the network to another computer (same file
 name) I'll overwrite the file.

 Once I copy the file, the system should notify a use that new file exist (a
 message pop-up).  If the use closes the message and will not open the file,
 after few ours I want to remind the user that the file has not been open
 yet.

 Will it be very complicated to write a bash script like this?

I once wrote a small Python script to help me with automating
Git-based deployments, but it's simple enough to be suitable for a
variety of use cases:

https://github.com/xen0n/touchmon

It allows you to specify a hook for each file observed, just make sure
the watched file exists before starting the watcher (a known
limitation). Inside the hook script (any language is OK) you can
pop-up a dialog and store a  timestamp somewhere, if the user has the
file opened for him/her. Then for the periodic notofication, maybe a
cronjob would suffice? Simply check the stored timestamp with the
current time. If polling is not what you want, you can extend the
inotify part to also monitor events like IN_OPEN or IN_ACCESS and
manipulate your timer accordingly. But be aware of false positives
like some background daemon accessing the file, as ALL opens and reads
are watched.

Hope this helps~



Re: [gentoo-user] Re: Can't find init due to inconsistent drive order

2013-07-02 Thread Wang Xuerui
2013/7/2 Grant Edwards grant.b.edwa...@gmail.com:
 It looks like my options are:

 snip

  5) For the drive with the root parition on it switch from a DOS
 parition table to a GPT partition table and use the
 root=PARTUUID=whatever kernel option.

 Switching to a GPT partition table sounds like the cleanest solution,
 but I need to figure out if the grub (legacy) ebuild includes GPT
 support or not.  I know it's supported by grub2, but I don't really
 feel like switching to grub2 ATM.

I'm not exactly sure about your system's setup, but msdos partition
tables can also hold partition UUIDs, which can be seen by running
blkid. So, such a solution does not necessarily imply a GPT-based
setup... you can experiment with it a bit, having fun tweaking kernel
parameters :)



Re: [gentoo-user] Gentoo installation, network adapter not supported

2013-06-29 Thread Wang Xuerui
2013/6/29 Zind wzmind...@gmail.com:
 hi all,
 I am new to Gentoo. I came across some problems in the Gentoo installation
 process.

 I got a relatively new laptop: Lenovo IdeaPad Y400.
 The Gentoo iso I choose is: admincd-amd64-20130620.iso, from:
 http://distfiles.gentoo.org/releases/amd64/autobuilds/20130620/hardened/
I once failed to install Hardened Gentoo to a server using the
admincd, so I'm not sure if you can proceed much further even if the
network is working.

Since after chroot the environment of LiveCD is not important any more
(except the running kernel, of course), you could simply use the
installation CD and a hardened stage3. If you are not trying to use
SELinux, this should be enough.

 After I made the LiveUSB, disabling the UEFI boot, I finally boot into the
 Gentoo kernel. But I could not connect to the network.
 Using the `ifconfig -a` command, I can only see the loopback interface: lo.
 Using the `lspci` command, I can see the two network adapters:
 (1) the Ethernet adapter
 Ethernet controller: Atheros Communications Inc. AR8161 Gigabit Ethernet
 (2) the wireless adapter
 Network controller: Intel Corporation Centrino Wireless-N 2230
 Subsystem: Intel Corporation Centrino Wireless-N 2230 BGN
 (The two network adapters both work well under Windows 8.)

 After several times of STFW, I can confirm it's a network adapter driver
 support issue: lacks of coresponding network adapter drivers.
 For the AR8161 network adapter, it requires the alx kernal module: alx.ko.
 Previously, I thought I could at least compile this driver by myself, but
 soon I found it lacks of build-essential packages, too. :-(
The toolchain is present on the installation CD so it must be the admincd...

 For the wireless network adapter, I can find
 iwlwifi-2030-6.ucode(coresponding to Intel Centrino Wireless-N 2230 BGN
 adapter) under /lib/firmware, I tried the `modprobe iwlwifi` command, still
 doesn't work. Weird.

 Uh... I don't know what to do next to continue the installation.
 Any suggestion or advice is appreciated.
You can get the firmware files from the git tree of linux-firmware
project (browsable online). Then you can transfer them into the LiveCD
environment's /lib/firmware by means of a USB stick, for example,
after which you simply rmmod iwlagn  modprobe iwlagn.

Also, before you finally reboot to finish the installation make sure
you emerge linux-firmware, so that the required firmware files are
installed into the target system.

Hope this information helps~



Re: [gentoo-user] Ext3 FS File Size Limits

2013-06-02 Thread Wang Xuerui
2013/6/2 Fast Turtle ftur...@gmail.com:
 One of the possible causes I've thought of was running out of innodes but 
 don't know how to check that or any of the other options used to create the 
 file system on - anyone want to help there?

You can try `touch`ing a file on that partition and see if it fails.
If so, you've run out of inodes; no need to peek at fs internals :)



Re: [gentoo-user] system time 6-hr. ahead

2013-05-11 Thread Wang Xuerui
2013/5/11 Joseph syscon...@gmail.com:
 My computer time reported by asterisk server or php database is 6hr ahead.
 Why?

 My desktop clock is correct, and command line date reporting local time as
 well.
 My /etc/conf.d/hwclock
 clock=local
 clock_hctosys=YES
 clock_systohc=YES

 Should I set clock=UTC? I'm running Windows via VirtualBox and I know
 windows is using local time.


You can make Windows use UTC by editing registry:
http://superuser.com/questions/185773/does-windows-7-support-utc-as-bios-time

However, do note that Windows would NOT update the hardware clock if
this is done. So make sure Linux would properly set the HW clock at
shutdown or your clock may drift over time.



Re: [gentoo-user] fdisk warnings during install; questions

2013-05-10 Thread Wang Xuerui
2013/5/10 Thanasis thana...@asyr.hopto.org:
 Where do we find gdisk (in portage)?

It's sys-apps/gptfdisk. Took a bit of time finding the package when I
first deployed a Gentoo server some time ago (:



Re: [gentoo-user] evince - Error printing to PDF

2013-04-18 Thread Wang Xuerui
2013/4/18 Joseph syscon...@gmail.com:
 Well, I finally narrow it down to one build in functionality it that is
 build into some library, though nobody can tell me which library is it.
 I'm talking about the option when you go to Print from for example Firefox
 or evince, when print window pop-up there is option: Print to File (this
 has nothing to do with cups-pdf filter)
 This option is not working as before. Prior to upgrade I could print from
 Firefox to PDF to any user sub-directory now I can only print to user home
 folder. Evince will not print at all to pdf or ps file.
 So that library screw me up, but I don't know which one is it.


Is that library ghostscript-gpl? The package provides quite a bit of
functionality regarding PS/PDF file generation. On my system this
gives:

$ equery depends ghostscript-gpl
 * These packages depend on ghostscript-gpl:
app-text/epspdf-0.5.3 (app-text/ghostscript-gpl)
app-text/libspectre-0.2.7 (=app-text/ghostscript-gpl-8.62)
app-text/ps2eps-1.64 (app-text/ghostscript-gpl)
dev-lang/nasm-2.10.07 (doc ? app-text/ghostscript-gpl)
dev-python/matplotlib-1.2.0-r2 (latex ? app-text/ghostscript-gpl)
dev-tex/tex4ht-20090611_p1038-r1 (app-text/ghostscript-gpl)
media-gfx/gimp-2.8.4 (postscript ? app-text/ghostscript-gpl)
media-gfx/imagemagick-6.7.8.7 (postscript ? app-text/ghostscript-gpl)
media-gfx/inkscape-0.48.4 (postscript ? app-text/ghostscript-gpl)
media-gfx/uniconvertor-1.1.5 (app-text/ghostscript-gpl)
media-libs/sk1libs-0.9.1 (app-text/ghostscript-gpl)
net-print/cups-1.5.2-r4 (app-text/ghostscript-gpl[cups])
net-print/foomatic-filters-4.0.17 (app-text/ghostscript-gpl)
sci-mathematics/octave-3.4.3-r1 (app-text/ghostscript-gpl)
sci-visualization/gnuplot-4.6.1 (doc ? app-text/ghostscript-gpl)

A lot of packages. So I suspect that is the problem.



Re: [gentoo-user] HD7870 AMD Catalyst no framebuffer console

2013-04-18 Thread Wang Xuerui
2013/4/18 Amankwah amankw...@gmail.com:
 Hi all,
 I have a video card HD7870, I tried install the AMD Catalyst driver 
 ~amd64 in portage, it works well under X. But when I press ctrl+alt+F1, the 
 screen is blank??
 Although I googled this problem I add nomodeset and radeon.modeset=0 
 video=uvesa to the kernel cmdline in /boot/grub/menu.lst, it didn't work at 
 all. What can I do to solve this?

IIRC fglrx does not support KMS, hence no framebuffer.

The radeon parameters are of no use in this case; in order for the
proprietary driver to work properly, the open source module must be
blacklisted. So instructing the radeon module makes little sense.

Please correct me if I got any of these wrong; haven't used the binary
blob for about 2 years myself, the memory is slowly fading.



Re: [gentoo-user] fcitx crash----something to do with libpng cairo?

2013-04-12 Thread Wang Xuerui
2013/4/13 Jackie jiangjun12...@gmail.com:
 I am use fcitx for Chinese  after updating few packages today,fcitx just
 crash on startup and won't work even I start it in a terminal.I searched
 through Internet and figured it that this may have some thing to do with
 update libpng and cairo.One way out,however,may be just downgrade libpng 
 cairo,which seem to involve tons of recompilation of related packages(got
 that after mask the update  emerge libpng cairo -pv). So,anyone met this 
 got a better solution?
 BTY,using fcitx 4.2.7,libpng-1.6.1:0/16 and cairo-1.12.14. HELP ME OUT:(


I'm also a fcitx user, but my box has the stable versions of libpng
and cairo as I checked just now. So I suppose your crashes are related
to the ~KEYWORDed dependencies.

If you have enough free time, you may try to debug the crash with gdb
and report the issue to upstream... this way the developers will know
about the breakage and come up with a patch, which benefits other
fcitx users running ~KEYWORD systems as well.