Re: [9fans] Making available a pre-compiled go binary for Miller's plan-9 Pi image

2016-10-28 Thread David du Colombier
> you have to have an n-1 version of Go to compile n

You have to bootstrap from Go >= 1.4.

If you want to build the current Go source from scratch on plan9/386,
you can just do:

cd /tmp
git clone -b go1.4.3 https://go.googlesource.com/go go1.4
cd go1.4
hget http://9legacy.org/go/patch/syscall-exec.diff | ape/patch -p1
cd src
make.rc

GOROOT_BOOTSTRAP=/tmp/go1.4
cd /tmp
git clone https://go.googlesource.com/go
cd go/src
make.rc

This is documented on https://github.com/golang/go/wiki/Plan9

-- 
David du Colombier



[9fans] acme tag bars stacking

2016-10-28 Thread Mathieu Lonjaret
Hi,

I regularly feel slightly tempted to modify (p9p) acme so that, when a
window is maximized with button 2, the other windows tags would stack
to the top of the column, instead of to the bottom. I find that, them
being on the bottom makes it harder for me to find back a particular
window. Probably just because I'm more comfortable looking at things
at the top half of my screen rather than at the top bottom. It's also
how other GUIs behave: e.g. your browser tags are at the top of the
window.

Anyway, does anyone know what the rationale was for choosing to stack
them at the bottom? Or why it would be a a bad idea to make them stack
at the top instead?

tia,
Mathieu



Re: [9fans] acme tag bars stacking

2016-10-28 Thread Mark van Atten
On Fri, Oct 28, 2016 at 4:23 PM, Mathieu Lonjaret
 wrote:
>Probably just because I'm more comfortable looking at things
> at the top half of my screen rather than at the top bottom.

Perhaps that was the very design consideration, but with preference given
to the window in which one is working over tags of other windows.

Mark.



Re: [9fans] acme tag bars stacking

2016-10-28 Thread James A. Robinson
I suspect one reason for placing the stack at the bottom is that it gives
you a consistent area to target, at the top, when going back to work on your
primary window.  If they were stacked top down then, depending on how many
files you had open, you might have to more carefully target how to get back
to your primary edit window.

Jim


Re: [9fans] acme tag bars stacking

2016-10-28 Thread Mathieu Lonjaret
I thought about that. you can explicitly give that preference to the
working window by fully maximizing it (i.e. making all the tags
disappear with button 3), so it didn't seem justification enough to
me. But maybe it is.


On 28 October 2016 at 16:37, Mark van Atten  wrote:
> On Fri, Oct 28, 2016 at 4:23 PM, Mathieu Lonjaret
>  wrote:
>>Probably just because I'm more comfortable looking at things
>> at the top half of my screen rather than at the top bottom.
>
> Perhaps that was the very design consideration, but with preference given
> to the window in which one is working over tags of other windows.
>
> Mark.
>



Re: [9fans] Making available a pre-compiled go binary for Miller's plan-9 Pi image

2016-10-28 Thread Chris McGee
Thanks David,

I will give this a try.

Chris

On Oct 28, 2016, at 2:58 AM, David du Colombier <0in...@gmail.com> wrote:

>> you have to have an n-1 version of Go to compile n
> 
> You have to bootstrap from Go >= 1.4.
> 
> If you want to build the current Go source from scratch on plan9/386,
> you can just do:
> 
> cd /tmp
> git clone -b go1.4.3 https://go.googlesource.com/go go1.4
> cd go1.4
> hget http://9legacy.org/go/patch/syscall-exec.diff | ape/patch -p1
> cd src
> make.rc
> 
> GOROOT_BOOTSTRAP=/tmp/go1.4
> cd /tmp
> git clone https://go.googlesource.com/go
> cd go/src
> make.rc
> 
> This is documented on https://github.com/golang/go/wiki/Plan9
> 
> -- 
> David du Colombier
> 



Re: [9fans] Making available a pre-compiled go binary for Miller's plan-9 Pi image

2016-10-28 Thread Skip Tavakkolian
An alternative to building a Go 1.4.3 bootstrap environment, is to build a
Go 1.7 bootstrap environment for Plan 9 in a mainstream environment like
Linux or Mac OS X, then use it to build Go 1.7 on Plan 9.   General outline
is to:

* follow the instructions on how to clone and build Go 1.7 on, say, Linux.
* once Go is built  and installed, build a Plan 9 bootstrap
cd $GOROOT/src
GOOS=plan9 GOARCH=386 ./boostrap.bash

this produces the bootstrap hierarchy in ../../go-plan9-386-bootstrap.
 copy the hierarchy over to Plan  9 and set GOROOT_BOOTSTRAP to that
location, then build Go 1.7 on Plan 9.

The general explanation of bootstrapping Go is here:
https://golang.org/doc/install/source#go14

FYI, when i need to 'git clone' something on Plan 9, I mount the fs from
Linux (p9p) and do the git stuff there.

On Fri, Oct 28, 2016 at 6:13 AM David du Colombier <0in...@gmail.com> wrote:

> > you have to have an n-1 version of Go to compile n
>
> You have to bootstrap from Go >= 1.4.
>
> If you want to build the current Go source from scratch on plan9/386,
> you can just do:
>
> cd /tmp
> git clone -b go1.4.3 https://go.googlesource.com/go go1.4
> cd go1.4
> hget http://9legacy.org/go/patch/syscall-exec.diff | ape/patch -p1
> cd src
> make.rc
>
> GOROOT_BOOTSTRAP=/tmp/go1.4
> cd /tmp
> git clone https://go.googlesource.com/go
> cd go/src
> make.rc
>
> This is documented on https://github.com/golang/go/wiki/Plan9
>
> --
> David du Colombier
>
>


Re: [9fans] Making available a pre-compiled go binary for Miller's plan-9 Pi image (Chris McGee)

2016-10-28 Thread Marshall Conover
> An alternative to building a Go 1.4.3 bootstrap environment, is to build a
Go 1.7 bootstrap environment for Plan 9 in a mainstream environment ...

> If you want to build the current Go source from scratch on plan9/386,
you can just do...

The binary I built was specifically for arm (or more accurately, for the
raspi), on which I initially spent a chunk of time trying to get 1.4
building; eventually, as suggested here, cross-compiling on my Linux box
was how I ended up getting things working. However, that process wasn't
really intuitive, and felt a good chunk like wasted time.

I figured having a binary available would save people some hassle; there's
no real reason to make everyone cross-compile the binary, especially with
as small a target platform as the pi, and for people like myself who
weren't immediately aware of exactly what to do, it's more things to figure
out for no real reason. Bandwidth is cheap, the binary's small, so I
figured having it around would help other people who install plan 9 on
their Pi to check it out.

I'm already hosting it myself, but it seems like my link got me
spam-filtered. I'm more just wondering where to make it more obvious that
it's available, so noobies like myself can just grab it and play instead of
having to spin the same wheels I did.

Thanks!

mars


Re: [9fans] Making available a pre-compiled go binary for Miller's plan-9 Pi image (Chris McGee)

2016-10-28 Thread David du Colombier
As you wish:

http://www.9legacy.org/download/go/go1.7.3-plan9-386-bootstrap.tbz
http://www.9legacy.org/download/go/go1.7.3-plan9-amd64-bootstrap.tbz
http://www.9legacy.org/download/go/go1.7.3-plan9-arm6-bootstrap.tbz
http://www.9legacy.org/download/go/go1.7.3-plan9-arm7-bootstrap.tbz

-- 
David du Colombier



Re: [9fans] Question re source trees

2016-10-28 Thread Kurt H Maier
On Thu, Sep 29, 2016 at 11:54:03AM +0100, Richard Miller wrote:
>
> > > I can also supply bluetooth code for the pi3 if anyone is interested.
> > >
> >
> > I am interested.  Will it be part of your next pi release?
>
> Probably not for a while. I would have to write some man pages first.

Is your contrib directory on sources the place this code will turn up
when you release it?  Or is there another release mechanism I should be
monitoring?

Thanks,
khm



Re: [9fans] acme tag bars stacking

2016-10-28 Thread yy
On 28 October 2016 at 16:23, Mathieu Lonjaret
 wrote:
> Anyway, does anyone know what the rationale was for choosing to stack
> them at the bottom? Or why it would be a a bad idea to make them stack
> at the top instead?

Let's suppose you have many windows in a column. When you work in one
of them, you B2 it and put it on the top of the stack. Then you work
on another one and it goes to the top, moving the previous one to the
second position, and so on. This way, your most recently used windows
are always on top, the least used ones go to the bottom of the stack.
I would find counterintuitive that the windows you used a longer time
ago stayed at the top, between your "working windows" and the column
and main tag lines.

But I would guess the main reason it works this way is that it seemed
more natural to move a window to the head than to the tail of a linked
list, and it just worked well enough.

I see how it may be more practical to stack them at the top when
working only with two or three windows, but it would be kind of weird
if you have ten. If you feel it will fit your workflow better, it is
probably not too difficult to get it done.


-- 
- yiyus || JGL .