[gentoo-dev] rfc: epause instead of sleep in ebuilds

[gentoo-dev] rfc: epause instead of sleep in ebuilds

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 1148 bytes --]

Currently, we have something like 200 ebuilds which call 'sleep'. How
would people feel about replacing this with an eutils function called
'epause' (or 'esleep', or 'ezzz', or esitaroundforabit)? This would give
us two advantages:

* For non-interactive builds (eg catalyst, chroot testing), an
environment variable could be set which makes epause not actually pause.
For certain packages on certain systems, the sleep takes longer than the
rest of the build...

* The behaviour of sleep isn't entirely standardised across different
UNIX implementations. It would be better to have one place to make any
changes needed by these new non-Linuxy platforms.

I'd suggest an implementation along the following lines:

# Wait for the supplied number of seconds. If no argument is supplied,
# defaults to five seconds. If the EPAUSE_IGNORE env var is set, don't
# wait.
epause() {
    if [ -z "$EPAUSE_IGNORE" ] ; then
        sleep ${1:-5}
    fi
}

Thoughts? I'd be happy to do all the work :)

-- 
Ciaran McCreesh : Gentoo Developer (Sparc, MIPS, Vim, Fluxbox)
Mail            : ciaranm at gentoo.org
Web             : http://dev.gentoo.org/~ciaranm


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Karl Trygve Kalleberg]

[-- Attachment #1: Type: text/plain, Size: 323 bytes --]

On Wed, Sep 01, 2004 at 09:21:44PM +0100, Ciaran McCreesh wrote:

> Thoughts? I'd be happy to do all the work :)

Sounds like a terrific idea.


One off-hand question, though: Are any of the sleeps here for synchronization
purposes, or are they all for actually displaying some information in a timely
fashion?


-- Karl T

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 863 bytes --]

On Wed, 1 Sep 2004 22:40:40 +0200 Karl Trygve Kalleberg
<karltk@gentoo.org> wrote:
| On Wed, Sep 01, 2004 at 09:21:44PM +0100, Ciaran McCreesh wrote:
| 
| > Thoughts? I'd be happy to do all the work :)
| 
| Sounds like a terrific idea.
| 
| One off-hand question, though: Are any of the sleeps here for
| synchronization purposes, or are they all for actually displaying some
| information in a timely fashion?

A quick glance suggests they're all for making sure an einfo/ewarn gets
read. I really can't think of a legitimate reason for using sleep to
sync something -- a sleep 1 might be fine to fix some really strange bug
on an amd64 box, for example, but it probably wouldn't be on an old sun
IPX...

-- 
Ciaran McCreesh : Gentoo Developer (Sparc, MIPS, Vim, Fluxbox)
Mail            : ciaranm at gentoo.org
Web             : http://dev.gentoo.org/~ciaranm


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Seemant Kulleen]

[-- Attachment #1: Type: text/plain, Size: 531 bytes --]

Ciaran,

That's a great idea.  Along these lines, perhaps something standardised
for the beeping einfo's would be nice too.  I know at least three
packages which make use of the beeping einfo messages (portage, gaim and
xorg-x11) -- it'd be nice to make it a function rather than the stuff
that's happening now.

Thanks,
-- 
Seemant Kulleen
http://dev.gentoo.org/~seemant

Public Key: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x3458780E
Key fingerprint = 23A9 7CB5 9BBB 4F8D 549B 6593 EDA2 65D8 3458 780E


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Chris Gianelloni]

[-- Attachment #1: Type: text/plain, Size: 1381 bytes --]

On Wed, 2004-09-01 at 16:21, Ciaran McCreesh wrote:
> Currently, we have something like 200 ebuilds which call 'sleep'. How
> would people feel about replacing this with an eutils function called
> 'epause' (or 'esleep', or 'ezzz', or esitaroundforabit)? This would give
> us two advantages:
> 
> * For non-interactive builds (eg catalyst, chroot testing), an
> environment variable could be set which makes epause not actually pause.
> For certain packages on certain systems, the sleep takes longer than the
> rest of the build...
> 
> * The behaviour of sleep isn't entirely standardised across different
> UNIX implementations. It would be better to have one place to make any
> changes needed by these new non-Linuxy platforms.
> 
> I'd suggest an implementation along the following lines:
> 
> # Wait for the supplied number of seconds. If no argument is supplied,
> # defaults to five seconds. If the EPAUSE_IGNORE env var is set, don't
> # wait.
> epause() {
>     if [ -z "$EPAUSE_IGNORE" ] ; then
>         sleep ${1:-5}
>     fi
> }
> 
> Thoughts? I'd be happy to do all the work :)

You'll do all the work?

In that case, I say go for it.  Anything that'll speed up building a
LiveCD is fine by me.

-- 
Chris Gianelloni
Release Engineering - Operations/QA Manager
Games - Developer
Gentoo Linux

Is your power animal a penguin?

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Tavis Ormandy]

On Wed, Sep 01, 2004 at 09:21:44PM +0100, Ciaran McCreesh wrote:
> I'd suggest an implementation along the following lines:
> 
> # Wait for the supplied number of seconds. If no argument is supplied,
> # defaults to five seconds. If the EPAUSE_IGNORE env var is set, don't
> # wait.
> epause() {
>     if [ -z "$EPAUSE_IGNORE" ] ; then
>         sleep ${1:-5}
>     fi
> }
> 
> Thoughts? I'd be happy to do all the work :)

sounds great, a test -t 1 in there could be handy as well.

-- 
-------------------------------------
taviso@sdf.lonestar.org | finger me for my gpg key.
-------------------------------------------------------

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 1084 bytes --]

On Wed, 01 Sep 2004 13:49:30 -0700 Seemant Kulleen <seemant@gentoo.org>
wrote:
| That's a great idea.  Along these lines, perhaps something
| standardised for the beeping einfo's would be nice too.  I know at
| least three packages which make use of the beeping einfo messages
| (portage, gaim and xorg-x11) -- it'd be nice to make it a function
| rather than the stuff that's happening now.

Ok, how's this look?

# Beep the specified number of times (defaults to five). If our output
# is not a terminal, or if seq (part of coreutils) is unavailable, don't
# beep.
ebeep() {
    local n
    if [ -t 1 ] && [ -x /bin/seq ] ; then
        for n in $(seq 1 ${1:-5}) ; do
            echo -ne "\a"
            sleep 0.1 &>/dev/null ; sleep 0,1 &>/dev/null
            echo -ne "\a"
            sleep 1
        done
    fi
}

I think this might not work on OSX, not sure that they have a seq
available (although they should...).

-- 
Ciaran McCreesh : Gentoo Developer (Sparc, MIPS, Vim, Fluxbox)
Mail            : ciaranm at gentoo.org
Web             : http://dev.gentoo.org/~ciaranm


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Tavis Ormandy]

On Wed, Sep 01, 2004 at 10:29:28PM +0100, Ciaran McCreesh wrote:
>         for n in $(seq 1 ${1:-5}) ; do
> 
> I think this might not work on OSX, not sure that they have a seq
> available (although they should...).

you can use this instead:

for ((n=1;n<=${1:-5};n++)); do

-- 
-------------------------------------
taviso@sdf.lonestar.org | finger me for my gpg key.
-------------------------------------------------------

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Mike Frysinger]

On Wednesday 01 September 2004 04:21 pm, Ciaran McCreesh wrote:
> Currently, we have something like 200 ebuilds which call 'sleep'. How
> would people feel about replacing this with an eutils function called
> 'epause' (or 'esleep', or 'ezzz', or esitaroundforabit)? This would give
> us two advantages:

i'm all for it :)
-mike

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Michael Sterrett -Mr. Bones.-]

Please wrap this is a NOBEEP variable or something so I can turn it all
off easily from make.conf.

Thanks,

Michael Sterrett
  -Mr. Bones.-
mr_bones_@gentoo.org


On Wed, 1 Sep 2004, Ciaran McCreesh wrote:

> On Wed, 01 Sep 2004 13:49:30 -0700 Seemant Kulleen <seemant@gentoo.org>
> wrote:
> | That's a great idea.  Along these lines, perhaps something
> | standardised for the beeping einfo's would be nice too.  I know at
> | least three packages which make use of the beeping einfo messages
> | (portage, gaim and xorg-x11) -- it'd be nice to make it a function
> | rather than the stuff that's happening now.
> 
> Ok, how's this look?
> 
> # Beep the specified number of times (defaults to five). If our output
> # is not a terminal, or if seq (part of coreutils) is unavailable, don't
> # beep.
> ebeep() {
>     local n
>     if [ -t 1 ] && [ -x /bin/seq ] ; then
>         for n in $(seq 1 ${1:-5}) ; do
>             echo -ne "\a"
>             sleep 0.1 &>/dev/null ; sleep 0,1 &>/dev/null
>             echo -ne "\a"
>             sleep 1
>         done
>     fi
> }
> 
> I think this might not work on OSX, not sure that they have a seq
> available (although they should...).
> 
> -- 
> Ciaran McCreesh : Gentoo Developer (Sparc, MIPS, Vim, Fluxbox)
> Mail            : ciaranm at gentoo.org
> Web             : http://dev.gentoo.org/~ciaranm
> 
> 

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 1239 bytes --]

On Wed, 1 Sep 2004 21:21:44 +0100 Ciaran McCreesh <ciaranm@gentoo.org>
wrote:
| Currently, we have something like 200 ebuilds which call 'sleep'. How
| would people feel about replacing this with an eutils function called
| 'epause' (or 'esleep', or 'ezzz', or esitaroundforabit)?

Ok, based upon feedback...

# Wait for the supplied number of seconds. If no argument is supplied,
# defaults to five seconds. If the EPAUSE_IGNORE env var is set, don't
# wait. If we're not outputting to a terminal, don't wait.
epause() {
    if [ -z "$EPAUSE_IGNORE" ] && [ -t 1 ] ; then
        sleep ${1:-5}
    fi
}

# Beep the specified number of times (defaults to five). If our output
# is not a terminal, don't beep. If the EBEEP_IGNORE env var is set,
# don't beep.
ebeep() {
    local n
    if [ -z "$EBEEP_IGNORE" ] && [ -t 1 ] ; then
        for ((n=1 ; n <= ${1:-5} ; n++)) ; do
            echo -ne "\a"
            sleep 0.1 &>/dev/null ; sleep 0,1 &>/dev/null
            echo -ne "\a"
            sleep 1
        done
    fi
}

If no-one screams, I'll go ahead with this tomorrow.

-- 
Ciaran McCreesh : Gentoo Developer (Sparc, MIPS, Vim, Fluxbox)
Mail            : ciaranm at gentoo.org
Web             : http://dev.gentoo.org/~ciaranm


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Brian Harring]

[-- Attachment #1: Type: text/plain, Size: 125 bytes --]

> If no-one screams, I'll go ahead with this tomorrow.
Go ahead with it how?
Aside from that question, I like it.
~Brian

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Mike Frysinger]

On Wednesday 01 September 2004 11:09 pm, Brian Harring wrote:
> > If no-one screams, I'll go ahead with this tomorrow.
>
> Go ahead with it how?

eutils.eclass ! :)
-mike

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Alexander Gretencord]

On Wednesday 01 September 2004 22:43, Ciaran McCreesh wrote:
> A quick glance suggests they're all for making sure an einfo/ewarn gets
> read.

If they really are, then we don't need an epause, as this is not at all 
adequate for making sure that important information gets read. My server is 
in the basement and I just don't hear any beeping and the sleeps don't 
interest me because I will just open a screen session, start the merge and 
close it. By the time I get back to it, the ewarn in merged ebuild number 3 
is not even in the buffer any more, while ebuild number 10 is building.

I'd propose something different. Have all ewarns/einfos put into some file, 
document where it is and display its contents and location _after_ the whole 
merge is over. This way I might see it, when I come to close the screen 
session/see if any build stopped.


Alex

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Dice R. Random]

Ah, you want the infamous bug 11359 ;) 
http://bugs.gentoo.org/show_bug.cgi?id=11359


On Thu, 2 Sep 2004 09:02:20 +0200, Alexander Gretencord <arutha@gmx.de> wrote:
> On Wednesday 01 September 2004 22:43, Ciaran McCreesh wrote:
> I'd propose something different. Have all ewarns/einfos put into some file,
> document where it is and display its contents and location _after_ the whole
> merge is over. This way I might see it, when I come to close the screen
> session/see if any build stopped.
> 
> Alex
> 
> 
> 
> --
> gentoo-dev@gentoo.org mailing list
> 
>

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Guillaume Destuynder]

Ciaran McCreesh wrote:

>
>I think this might not work on OSX, not sure that they have a seq
>available (although they should...).
>
>  
>
on macosx 10.3.5 (so i guess earlier alike) there is no seq program 
installed. (I just checked)

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Mike Frysinger]

On Thursday 02 September 2004 03:02 am, Alexander Gretencord wrote:
> My server is
> in the basement and I just don't hear any beeping

beeps goto the terminal, not to your system speaker ... so unless you have 
your sound muted on your client too ...
-mike

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Mamoru KOMACHI]

[-- Attachment #1: Type: text/plain, Size: 457 bytes --]

Hi all,

*BSD users have jot instead of seq ;)

--
Mamoru KOMACHI <usata@gentoo.org>

At Thu, 02 Sep 2004 15:05:42 +0200,
Guillaume Destuynder wrote:
> 
> Ciaran McCreesh wrote:
> 
> >
> >I think this might not work on OSX, not sure that they have a seq
> >available (although they should...).
> >
> >  
> >
> on macosx 10.3.5 (so i guess earlier alike) there is no seq program 
> installed. (I just checked)
> 
> --
> gentoo-dev@gentoo.org mailing list
> 

[-- Attachment #2: Type: application/pgp-signature, Size: 187 bytes --]

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 467 bytes --]

On Thu, 02 Sep 2004 22:49:10 +0900 Mamoru KOMACHI <usata@gentoo.org>
wrote:
| *BSD users have jot instead of seq ;)

Yeah, so much for hoping you'd all just use coreutils and have done with
it eh? :) I've switched to doing a bash for loop instead at Taviso's
suggestion, so we don't need to worry about this.

-- 
Ciaran McCreesh : Gentoo Developer (Sparc, MIPS, Vim, Fluxbox)
Mail            : ciaranm at gentoo.org
Web             : http://dev.gentoo.org/~ciaranm


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Alexander Gretencord]

On Thursday 02 September 2004 15:30, Mike Frysinger wrote:
> beeps goto the terminal, not to your system speaker ... so unless you have
> your sound muted on your client too ...

My terminal is the screen session, which gets distached after hitting enter to 
start the merge, so essentially they go nowhere. As others have said already 
in bug #11359, up to now, all ewarn/einfo messages and beeps get completely 
lost.


Alex

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Mike Doty]

why not have a emerge option, or when you use --pretend to display all
einfo/warn/errors.  before I install a package, I always do a emerge
-pv first, as I'd imagine most do.

On Thu, 2 Sep 2004 20:03:11 +0200, Alexander Gretencord <arutha@gmx.de> wrote:
> On Thursday 02 September 2004 15:30, Mike Frysinger wrote:
> > beeps goto the terminal, not to your system speaker ... so unless you have
> > your sound muted on your client too ...
> 
> My terminal is the screen session, which gets distached after hitting enter to
> start the merge, so essentially they go nowhere. As others have said already
> in bug #11359, up to now, all ewarn/einfo messages and beeps get completely
> lost.
> 
> Alex
> 
> 
> 
> --
> gentoo-dev@gentoo.org mailing list
> 
>

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[N. Owen Gunden]

On Thu, Sep 02, 2004 at 01:51:15PM -0500, Mike Doty wrote:
> why not have a emerge option, or when you use --pretend to display all
> einfo/warn/errors.  before I install a package, I always do a emerge
> -pv first, as I'd imagine most do.

Some if not most of the warnings and errors only pop up in certain
circumstances (e.g. something about your system means you have to do
additional configuration).  So you'd end up with a whole host of
warn/errors that may or may not apply to you.

 - O

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Paul de Vrieze]

On Thursday 02 September 2004 22:03, N. Owen Gunden wrote:
> On Thu, Sep 02, 2004 at 01:51:15PM -0500, Mike Doty wrote:
> > why not have a emerge option, or when you use --pretend to display
> > all einfo/warn/errors.  before I install a package, I always do a
> > emerge -pv first, as I'd imagine most do.
>
> Some if not most of the warnings and errors only pop up in certain
> circumstances (e.g. something about your system means you have to do
> additional configuration).  So you'd end up with a whole host of
> warn/errors that may or may not apply to you.

It should be possible to have an extra "optional" function in ebuilds that 
is called allready in the pretend stage, and has the restriction that it 
must only output warning messages and the like. Of course those messages 
can be optional.

Paul

-- 
Paul de Vrieze
Gentoo Developer
Mail: pauldv@gentoo.org
Homepage: http://www.devrieze.net

--
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Simon Stelling]

[-- Attachment #1.1: Type: text/plain, Size: 2039 bytes --]

Alexander Gretencord wrote:
>>A quick glance suggests they're all for making sure an einfo/ewarn gets
>>read.
> 
> 
> If they really are, then we don't need an epause, as this is not at all 
> adequate for making sure that important information gets read. My server is 
> in the basement and I just don't hear any beeping and the sleeps don't 
> interest me because I will just open a screen session, start the merge and 
> close it. By the time I get back to it, the ewarn in merged ebuild number 3 
> is not even in the buffer any more, while ebuild number 10 is building.
> 
> I'd propose something different. Have all ewarns/einfos put into some file, 
> document where it is and display its contents and location _after_ the whole 
> merge is over. This way I might see it, when I come to close the screen 
> session/see if any build stopped.

I think it's much easier and more user-friendly to make emerge less
verbose, the only thing the user would see is (in case of a successful
emerge) this:

  >>> Unpacking ncurses-5.4.tar.gz to /var/tmp/portage/ncurses-5.4-r5/work
* Applying ncurses-5.4-xterm.patch...
* Applying ncurses-5.4-share-sed.patch...

* Using GNU config files from /usr/share/libtool
*  Updating config.sub
*  Updating config.guess

 >>> Source unpacked.
 >>> Install ncurses-5.4-r5 into /var/tmp/portage/ncurses-5.4-r5/image/
category sys-libs
  * Installing basic terminfo files in /etc...
 >>> Completed installing into /var/tmp/portage/ncurses-5.4-r5/image/
 >>> Merging sys-libs/ncurses-5.4-r5 to /
 >>> Auto-cleaning packages ...
 >>> No outdated packages were found on your system.

  * GNU info directory index is up-to-date.

or so ;)

of course this is not fine if you have a package that failed to emerge,
but caching the last X lines and printing it in case of an error would
be a nice workaround so if there is an error you don't have to reemerge
the package with verbose option and if there is no one, you only see
what really is interesting. -v could be used to get the old-fashioned style.

blubb



[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 190 bytes --]

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Lisa Seelye]

[-- Attachment #1: Type: text/plain, Size: 555 bytes --]

On Wed, 2004-09-01 at 16:21, Ciaran McCreesh wrote:
> Currently, we have something like 200 ebuilds which call 'sleep'. How
> would people feel about replacing this with an eutils function called
> 'epause' (or 'esleep', or 'ezzz', or esitaroundforabit)? This would give
> us two advantages:

+1 for the idea.

Would be neat to have some options:

epause --beep 1s --sleep 5s

To beep one second for 5 seconds while sleeping... Just a thought.
-- 
Regards,
Lisa Seelye
Key fingerprint = 09CF 52D6 B82B 72B9 97A7  601B CB46 B556 1E49 6FC5

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] rfc: epause instead of sleep in ebuilds

[Thomas de Grenier de Latour]

On Fri, 03 Sep 2004 20:17:13 +0200
Simon Stelling <blubb@gentoo.org> wrote:

> I think it's much easier and more user-friendly to make emerge
> less verbose

You can give a try to my patch in bug #37491, I've updated it for
last portage pre-release, and making emerge a bit quiet was its
exact purpose.

> * Applying ncurses-5.4-xterm.patch...
> * Applying ncurses-5.4-share-sed.patch...

Imho, this ones should not be shown to user. With my patch, such
messages could be changed from "einfo" to "veinfo", which looks
the same but only display in verbose output/log. Since they are
mostly in a few eclasses, that would not be to much work to make
this cleanup. 

> of course this is not fine if you have a package that failed to
> emerge, but caching the last X lines and printing it in case of
> an error would be a nice workaround

What I do is that if emerge is in quiet mode, and the user don't
uses PORT_LOGDIR, then there will still be a verbose log in ${T}.
And "die" messages point to this log. (see the 2nd screenshot in
the bug report)

> -v could be used to get the old-fashioned style.

With my patch, default behavior is still to be verbose, but if
"quiet" is in feature flags (or on command line options). And yes,
in all case, -v forces verbose mode.

-- 
TGL.

--
gentoo-dev@gentoo.org mailing list