Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.lang.forth > #44147 > unrolled thread

lina32 and android

Started byAlbert van der Horst <albert@spenarnc.xs4all.nl>
First post2015-12-08 19:36 +0000
Last post2015-12-11 21:52 +0100
Articles 14 — 7 participants

Back to article view | Back to comp.lang.forth


Contents

  lina32 and android Albert van der Horst <albert@spenarnc.xs4all.nl> - 2015-12-08 19:36 +0000
    Re: lina32 and android Mark Wills <markwills1970@gmail.com> - 2015-12-09 01:27 -0800
      Re: lina32 and android Albert van der Horst <albert@spenarnc.xs4all.nl> - 2015-12-09 11:24 +0000
        Re: lina32 and android Ron Aaron <rambamist@gmail.com> - 2015-12-09 13:27 +0200
    Re: lina32 and android Julian Fondren <julian.fondren@gmail.com> - 2015-12-09 05:39 -0800
      Re: lina32 and android Albert van der Horst <albert@spenarnc.xs4all.nl> - 2015-12-09 15:00 +0000
      Re: lina32 and android Bernd Paysan <bernd.paysan@gmx.de> - 2015-12-09 18:06 +0100
        Re: lina32 and android Julian Fondren <julian.fondren@gmail.com> - 2015-12-09 10:12 -0800
        Re: lina32 and android Bernd Paysan <bernd.paysan@gmx.de> - 2015-12-09 23:06 +0100
          Re: lina32 and android Lars Brinkhoff <lars.spam@nocrew.org> - 2015-12-10 07:01 +0100
            Re: lina32 and android Lars Brinkhoff <lars.spam@nocrew.org> - 2015-12-10 09:32 +0100
              Re: lina32 and android anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2015-12-10 13:34 +0000
              Gforth automated build Lars Brinkhoff <lars.spam@nocrew.org> - 2015-12-11 08:54 +0100
                Re: Gforth automated build Bernd Paysan <bernd.paysan@gmx.de> - 2015-12-11 21:52 +0100

#44147 — lina32 and android

FromAlbert van der Horst <albert@spenarnc.xs4all.nl>
Date2015-12-08 19:36 +0000
Subjectlina32 and android
Message-ID<56673128$0$23866$e4fe514c@news.xs4all.nl>
ciforth on an android smartphone or tablet.
===========================================

Since some time I'm in the possesion of a tablet (an Asus Memopad) with
an Intel processor.  I was astonished to discover that ciforth
(lina32) runs on that tablet without any adaptation.
Getting it to run is somewhat involved though.

What is needed.
===============

To run lina, you must have it. Download
lina32 5.2 from the website of the Dutch Forth gg
http://www.forth.hcc.nl/w/Producten/Producten (ciforth stuff is
available in English, even if this site is predominantly Dutch.)
If you know that you want lina32-5.2.1.tgz , google will find
it for you.

Then we must unpack. You may not have already have a good
unpacker for tgz files. What certainly works is Zarchiver.
Get it from the app-store.

We need a command shell in order to type `` lina32 -e '.
Maybe you have one, maybe not. What certainly works is
``Terminal Emulator'', also from the app-store.

You can do without a decent keyboard, but not really.
With the default keyboard you'll have trouble with special
characters, spelling corrections and dead keys for double and single
quotes. Get `` Hacker's Keyboard'' from the store.

I don't quite understand the situation with so called rootkits. You
seem to need it to be able to run normal programs.
Get `` GNURoot ''. Run it once, that seems to do it.

Preparation.
============

Unpacking proceeds as follows:
  - go to the Downloads directory, using the file manager.
  - click lina32-5.2.1.tgz
  - Zarchiver proposes to unpack
     unpack it in the Download directory.
     You now have a tar
  - click on lina32-5.2.1.tar
  - let Zarchiver do its thing for the second time.
    You now have a directory tree, from the Download directory.

lina32 runs!
============

Now the running.
Optionally:
   click on the hackers keyboard.
     - enable it (click) *and also*
     - set it as the input keyboard (click)

Start the terminal emulator (click)
You go to your home directory and you can see what your home is
   - cd
   - pwd
   Leave the terminal emulator with ^D.
   If you don't have the hackers keyboard, good luck with the control
   characters.
   If pwd doesn't work then something went wrong with ``GNUroot''.
   You're on your own here...

   Now in the terminal emulator:
   - cd
   - cp -r /data/storage/legacy/Download/lina-5.2.1 .
   The whole directory tree will be copied from Download
   The hackers keyboard understands tab-completion, meaning
   that after a tab the remainder of the file name is filled in.
   (D<tab> -> Download).
   You can also get back and correct a command that you made a
   mistake in, using the arrow keys.

   Now you use lina:
   - cd lina32-5.2.1
   - ./lina32 -e
      You're in lina, with electives (a system dressed up
      with facilities)
      Show me some prime numbers:
       - WANT SIEVE 10 SIEVE
      Measure the speed of your systeem
      - CASE-SENSITIVE WANT MEASUREPRIME
      (This is pretty impressive, it load the assembler and handles
      the fast ticks of the RDTSC, instruction.)
    - ./lina -c hellow.frt
      ./hellow
      Hello world!
      Apparently the fasm generated elf files are compatible with
      android. This android's most lightweight programming system:
      lina + forth.lab

    What also works:
       importing commando's from a shell,
       such as to copy a file with cp from within Forth.
    (After further installation, read the pdf )

Groetjes Albert
-- 
Albert van der Horst, UTRECHT,THE NETHERLANDS
Economic growth -- being exponential -- ultimately falters.
albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst

[toc] | [next] | [standalone]


#44154

FromMark Wills <markwills1970@gmail.com>
Date2015-12-09 01:27 -0800
Message-ID<d5302a92-ba12-4498-b06f-552faaa4125a@googlegroups.com>
In reply to#44147
On Tuesday, 8 December 2015 19:36:10 UTC, Albert van der Horst  wrote:
> ciforth on an android smartphone or tablet.
> ===========================================
> 
> Since some time I'm in the possesion of a tablet (an Asus Memopad) with
> an Intel processor.  I was astonished to discover that ciforth
> (lina32) runs on that tablet without any adaptation.
> Getting it to run is somewhat involved though.
> 
> What is needed.
> ===============
> 
> To run lina, you must have it. Download
> lina32 5.2 from the website of the Dutch Forth gg
> http://www.forth.hcc.nl/w/Producten/Producten (ciforth stuff is
> available in English, even if this site is predominantly Dutch.)
> If you know that you want lina32-5.2.1.tgz , google will find
> it for you.
> 
> Then we must unpack. You may not have already have a good
> unpacker for tgz files. What certainly works is Zarchiver.
> Get it from the app-store.
> 
> We need a command shell in order to type `` lina32 -e '.
> Maybe you have one, maybe not. What certainly works is
> ``Terminal Emulator'', also from the app-store.
> 
> You can do without a decent keyboard, but not really.
> With the default keyboard you'll have trouble with special
> characters, spelling corrections and dead keys for double and single
> quotes. Get `` Hacker's Keyboard'' from the store.
> 
> I don't quite understand the situation with so called rootkits. You
> seem to need it to be able to run normal programs.
> Get `` GNURoot ''. Run it once, that seems to do it.
> 
> Preparation.
> ============
> 
> Unpacking proceeds as follows:
>   - go to the Downloads directory, using the file manager.
>   - click lina32-5.2.1.tgz
>   - Zarchiver proposes to unpack
>      unpack it in the Download directory.
>      You now have a tar
>   - click on lina32-5.2.1.tar
>   - let Zarchiver do its thing for the second time.
>     You now have a directory tree, from the Download directory.
> 
> lina32 runs!
> ============
> 
> Now the running.
> Optionally:
>    click on the hackers keyboard.
>      - enable it (click) *and also*
>      - set it as the input keyboard (click)
> 
> Start the terminal emulator (click)
> You go to your home directory and you can see what your home is
>    - cd
>    - pwd
>    Leave the terminal emulator with ^D.
>    If you don't have the hackers keyboard, good luck with the control
>    characters.
>    If pwd doesn't work then something went wrong with ``GNUroot''.
>    You're on your own here...
> 
>    Now in the terminal emulator:
>    - cd
>    - cp -r /data/storage/legacy/Download/lina-5.2.1 .
>    The whole directory tree will be copied from Download
>    The hackers keyboard understands tab-completion, meaning
>    that after a tab the remainder of the file name is filled in.
>    (D<tab> -> Download).
>    You can also get back and correct a command that you made a
>    mistake in, using the arrow keys.
> 
>    Now you use lina:
>    - cd lina32-5.2.1
>    - ./lina32 -e
>       You're in lina, with electives (a system dressed up
>       with facilities)
>       Show me some prime numbers:
>        - WANT SIEVE 10 SIEVE
>       Measure the speed of your systeem
>       - CASE-SENSITIVE WANT MEASUREPRIME
>       (This is pretty impressive, it load the assembler and handles
>       the fast ticks of the RDTSC, instruction.)
>     - ./lina -c hellow.frt
>       ./hellow
>       Hello world!
>       Apparently the fasm generated elf files are compatible with
>       android. This android's most lightweight programming system:
>       lina + forth.lab
> 
>     What also works:
>        importing commando's from a shell,
>        such as to copy a file with cp from within Forth.
>     (After further installation, read the pdf )
> 
> Groetjes Albert
> -- 
> Albert van der Horst, UTRECHT,THE NETHERLANDS
> Economic growth -- being exponential -- ultimately falters.
> albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst

I'm presuming then that your Intel based tablet is simulating an ARM
CPU? I could be wrong, but I don't think Google produces an Intel build
of Android. 

[toc] | [prev] | [next] | [standalone]


#44156

FromAlbert van der Horst <albert@spenarnc.xs4all.nl>
Date2015-12-09 11:24 +0000
Message-ID<56680f55$0$23730$e4fe514c@news.xs4all.nl>
In reply to#44154
Mark Wills <markwills1970@gmail.com> writes:

>On Tuesday, 8 December 2015 19:36:10 UTC, Albert van der Horst  wrote:
>> ciforth on an android smartphone or tablet.
>> ===========================================
>>
>> Since some time I'm in the possesion of a tablet (an Asus Memopad) with
>> an Intel processor.  I was astonished to discover that ciforth
>> (lina32) runs on that tablet without any adaptation.
>> Getting it to run is somewhat involved though.
>>
>> What is needed.
>> ===============
>>
>> To run lina, you must have it. Download
>> lina32 5.2 from the website of the Dutch Forth gg
>> http://www.forth.hcc.nl/w/Producten/Producten (ciforth stuff is
>> available in English, even if this site is predominantly Dutch.)
>> If you know that you want lina32-5.2.1.tgz , google will find
>> it for you.
<SNIP>
>>       - CASE-SENSITIVE WANT MEASUREPRIME

   CORRECTION! MAKE THAT
         - CASE-SENSITIVE WANT MEASURE-PRIME
<SNIP>

>>     What also works:
>>        importing commando's from a shell,
>>        such as to copy a file with cp from within Forth.
>>     (After further installation, read the pdf )
>>
>> Groetjes Albert

>I'm presuming then that your Intel based tablet is simulating an ARM
>CPU? I could be wrong, but I don't think Google produces an Intel build
>of Android.

How could simulating an ARM based CPU help an
Intel processor run an assembler-based ELF32-i386 binary?

Look up the source of lina32 and ask yourself how ARM could ever get
into the equation. Then there is of course the speed of
MEASURE-PRIME. That proves there is no layer between lina and android,
let alone the needed two layers (i386->ARM->i386),

I think the emperical evidence trumps any conception you may
have of what Google is doing.

Groetjes Albert
-- 
Albert van der Horst, UTRECHT,THE NETHERLANDS
Economic growth -- being exponential -- ultimately falters.
albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst

[toc] | [prev] | [next] | [standalone]


#44157

FromRon Aaron <rambamist@gmail.com>
Date2015-12-09 13:27 +0200
Message-ID<n49331$6mu$1@dont-email.me>
In reply to#44156

On 12/09/2015 13:24, Albert van der Horst wrote:

> I think the emperical evidence trumps any conception you may
> have of what Google is doing.

One could, minimally, Google it:  https://01.org/android-IA

[toc] | [prev] | [next] | [standalone]


#44158

FromJulian Fondren <julian.fondren@gmail.com>
Date2015-12-09 05:39 -0800
Message-ID<d051d6b8-4502-417d-a320-256c8ae4cc55@googlegroups.com>
In reply to#44147
On Tuesday, December 8, 2015 at 1:36:10 PM UTC-6, Albert van der Horst wrote:
> ciforth on an android smartphone or tablet.
> ===========================================
> 
> Since some time I'm in the possesion of a tablet (an Asus Memopad) with
> an Intel processor.  I was astonished to discover that ciforth
> (lina32) runs on that tablet without any adaptation.
> Getting it to run is somewhat involved though.

Yep, Android is more Linux than many people seem to be willing to
believe.  It's also quite easy to get ARM-Linux Forths to run under
it.  Compiling an ARM jonesforth f.e. only required that I install
gcc.  Root was not needed.  The ARM assembler work that I did a few
months back was all in AirTerm on a KitKat laptop with vim and gas.

It's a tragedy that there's no QuartusForth for this generation.
gforth is fun, but lacks a reasonable build system or any intent
towards having an Android-hosted build system.  VFX Forth's graphics
are X, for other ARM-Linuxes.  For Forth Inc, ARM is SwiftX-only,
with a tether for development.

Maybe one of these parties will get a Pixel C for Christmas.
https://store.google.com/product/pixel_c


lina32 on Android is nice, but Intel-Android products barely exist.
I've had a succession of Kindle Fires and Samsung products and that
HP Android laptop... and may well get a Pixel C.  App market aside,
and even making-apps-for-others aside, there is also a huge device
quality difference between these and what you can find with an Intel
CPU.


-- Julian

[toc] | [prev] | [next] | [standalone]


#44160

FromAlbert van der Horst <albert@spenarnc.xs4all.nl>
Date2015-12-09 15:00 +0000
Message-ID<56684203$0$23804$e4fe514c@news.xs4all.nl>
In reply to#44158
Julian Fondren <julian.fondren@gmail.com> writes:

>On Tuesday, December 8, 2015 at 1:36:10 PM UTC-6, Albert van der Horst wrote:
>> ciforth on an android smartphone or tablet.
>> ===========================================
>>
>> Since some time I'm in the possesion of a tablet (an Asus Memopad) with
>> an Intel processor.  I was astonished to discover that ciforth
>> (lina32) runs on that tablet without any adaptation.
>> Getting it to run is somewhat involved though.

>Yep, Android is more Linux than many people seem to be willing to
>believe.  It's also quite easy to get ARM-Linux Forths to run under
>it.  Compiling an ARM jonesforth f.e. only required that I install
>gcc.  Root was not needed.  The ARM assembler work that I did a few
>months back was all in AirTerm on a KitKat laptop with vim and gas.

>It's a tragedy that there's no QuartusForth for this generation.
>gforth is fun, but lacks a reasonable build system or any intent
>towards having an Android-hosted build system.  VFX Forth's graphics
>are X, for other ARM-Linuxes.  For Forth Inc, ARM is SwiftX-only,
>with a tether for development.

I wouldn't like to try to build ciforth's on android, as especially
for the documentation I want tools like M4,texinfo,sorting program's
Make,and fasm. Typing is a bitch too on a tablet.
I like the powerful environment on linux 64 bits that allows me to
build everything from MS DOS windows stand alone 32 and 64 bits.

>Maybe one of these parties will get a Pixel C for Christmas.
>https://store.google.com/product/pixel_c

>lina32 on Android is nice, but Intel-Android products barely exist.
>I've had a succession of Kindle Fires and Samsung products and that
>HP Android laptop... and may well get a Pixel C.  App market aside,
>and even making-apps-for-others aside, there is also a huge device
>quality difference between these and what you can find with an Intel
>CPU.

lina32 for ARM is on its way. Frans van de Markt is working on it.
It passes all tests and can generate the documentation. What's
missing is the compilation option -c / SAVE-SYSTEM. That is
sufficiently technical that he needs my assistance.
The targets for the moment are the Bear Blackboard and the Raspberry
Pi.

>-- Julian

Groetjes Albert
-- 
Albert van der Horst, UTRECHT,THE NETHERLANDS
Economic growth -- being exponential -- ultimately falters.
albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst

[toc] | [prev] | [next] | [standalone]


#44163

FromBernd Paysan <bernd.paysan@gmx.de>
Date2015-12-09 18:06 +0100
Message-ID<n49mtc$km6$1@dont-email.me>
In reply to#44158
Julian Fondren wrote:
> It's a tragedy that there's no QuartusForth for this generation.
> gforth is fun, but lacks a reasonable build system or any intent
> towards having an Android-hosted build system.

I'm using cross compilation with the NDK build system.  As Gforth comes as 
real Android app, not as command line program for a terminal app, this is 
the only way.  Yes, it's complicated, especially since Gforth needs several 
other third-party libraries to work, like SOIL to load the PNG for the 
terminal font, or Gerald's patched-for-Forth swig to generate the C library 
bindings.

When I started, I first compiled Gforth as command line program, that was 
much easier, and instant success.  I've also compiled Gforth as command line 
program for iOS, also instant success (however requires rooting iOS, and 
only runs as process when you ssh to the device).

IMHO you can't have both: usefulness and easy to build.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
net2o ID: kQusJzA;7*?t=uy@X}1GWr!+0qqp_Cn176t4(dQ*
http://bernd-paysan.de/

[toc] | [prev] | [next] | [standalone]


#44164

FromJulian Fondren <julian.fondren@gmail.com>
Date2015-12-09 10:12 -0800
Message-ID<4f71106b-bd73-461b-9cb5-4267a08aa66e@googlegroups.com>
In reply to#44163
On Wednesday, December 9, 2015 at 9:00:21 AM UTC-6, Albert van der Horst wrote:
> I wouldn't like to try to build ciforth's on android, as especially
> for the documentation I want tools like M4,texinfo,sorting program's
> Make,and fasm.

make, m4, and texinfo are available.  fasm is not: a version of it
can emit ARM, but there is no fasm-in-ARM.

Android has rsync, git, and vi.

> Typing is a bitch too on a tablet.

I've never once typed on a tablet.  I've used bluetooth keyboards.
There are also hybrid and laptop Android devices.  I found that I
don't like bluetooth or hybrid devices as they don't support in-bed
propped-up hacking, but the laptop works fine and the Pixel C's
keyboard may well be bed-friendly.

> I like the powerful environment on linux 64 bits that allows me to
> build everything from MS DOS windows stand alone 32 and 64 bits.

An environment that completely wastes the multiuser potential of
Linux, which Android spends on app security and per-app firewalling.
An environment where the unfree bits are also dodgy as hell (my
Yoga 3 Pro has me cursing the next three generations of Broadcom's
employees.  How dare they pretend to make WiFi components?)  The
HiDPI screen is halfway supported - everything else that's
distinctive about the laptop is broken.  It can't even go to sleep
and reawaken reliably.  I underclock it by half to ensure that Linux
doesn't actually burn the CPU out because Linux doesn't know about
the hidden fan.

And it's all downhill (better-supported in some directions, but then
the hardware's not worth the support) from here.

If desktop Linux has any power, it's just that people respect it as a
platform, but think of Android as a crippled media consumption shell.
Every once in a while someone takes Android seriously and you get
things like Terminal IDE, an open source project that lets you create
Android apps, even the Java UI kind, on Android, from scratch.  Or
GNUroot, which is actually not a 'rootkit' at all, but just ptrace
containerization with arm-linux root filesystems.

> lina32 for ARM is on its way. Frans van de Markt is working on it.
> It passes all tests and can generate the documentation. What's
> missing is the compilation option -c / SAVE-SYSTEM. That is
> sufficiently technical that he needs my assistance.
> The targets for the moment are the Bear Blackboard and the Raspberry
> Pi.

Cool.  I'll look forward to it.


On Wednesday, December 9, 2015 at 11:06:09 AM UTC-6, Bernd Paysan wrote:
> Julian Fondren wrote:
> > It's a tragedy that there's no QuartusForth for this generation.
> > gforth is fun, but lacks a reasonable build system or any intent
> > towards having an Android-hosted build system.
> 
> I'm using cross compilation with the NDK build system.  As Gforth comes as 
> real Android app, not as command line program for a terminal app, this is 
> the only way.  Yes, it's complicated, especially since Gforth needs several 
> other third-party libraries to work, like SOIL to load the PNG for the 
> terminal font, or Gerald's patched-for-Forth swig to generate the C library 
> bindings.

You have to enter a series of commands to assemble and build and
prepare those libraries.  Any series of commands that you do once,
and then want to be able to repeat, and then want to not have to
repeat when no longer necessary, you can put into a makefile.  Even
when the bits rot, makefiles like this are better documentation, and
easier to update, than instructions to the same effect.  They are
really just documentation that you can test.

wget; tar; (cd; make install); touch .have_swig.  It's hackish but
far less the long adventure into insanity that you get with perfectly
normal and proper autoconf usage (which begins with: don't put
'configure' in your repo.)  And you don't have to make autoconf care
about it.  Just stick it in a build-required-libs.mk

> When I started, I first compiled Gforth as command line program, that was 
> much easier, and instant success.  I've also compiled Gforth as command line 
> program for iOS, also instant success (however requires rooting iOS, and 
> only runs as process when you ssh to the device).
> 
> IMHO you can't have both: usefulness and easy to build.

Well you can have multiple build steps.

1. build gforth.so and the other libraries.

2. build the native NDK launcher that loads these libraries.

3. given #1 and #2 and a bunch of static content, package an Android
app that will fire things up and then look at one of those static
files for an actual application to compile and run -- like
gl-terminal, or Disoriented Bird, or whatever.

#1 gives you ultimate customization, including removing what you
don't need, but requires gathering all of these other libraries
together.

#2 you can probably do pretty easily after a simple NDK install but
it gives you maximum control over the Android details.

#3 you can do on the Android device itself, although perhaps only
with debugging signing.  Good enough for development, and using
gforth as a scripting language with Android-launched scripts, and
passing apps to friends, but not for deployment to a market.  For
deployment you can go back to step #2.

The hard parts are precisely the parts that will change the least,
and need to be rebuilt the least often.  Step #3 meanwhile is all
Forth source files and application-specific assets (and XML), and
most of the 'building' time would be spent watching the Android
installer.

Gentoo had stage1, stage2, and stage3 installs like this.  The
Android NDK itself comes with an already compiled GCC toolchain.


-- Julian

[toc] | [prev] | [next] | [standalone]


#44165

FromBernd Paysan <bernd.paysan@gmx.de>
Date2015-12-09 23:06 +0100
Message-ID<n4a8h4$v8r$1@dont-email.me>
In reply to#44163
Bernd Paysan wrote:
> IMHO you can't have both: usefulness and easy to build.

BTW: I still have to write up all the steps to get a successful build.  
However, this requires third parties to test it, people who don't have 
everything readily installed like I do.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
net2o ID: kQusJzA;7*?t=uy@X}1GWr!+0qqp_Cn176t4(dQ*
http://bernd-paysan.de/

[toc] | [prev] | [next] | [standalone]


#44168

FromLars Brinkhoff <lars.spam@nocrew.org>
Date2015-12-10 07:01 +0100
Message-ID<86fuzag7ve.fsf@vps34351.public.cloudvps.com>
In reply to#44165
Bernd Paysan wrote:
> BTW: I still have to write up all the steps to get a successful build.  
> However, this requires third parties to test it, people who don't have 
> everything readily installed like I do.

How about setting up a continuous integration build?  If you send the
instructions to me, I can do that for you.

[toc] | [prev] | [next] | [standalone]


#44170

FromLars Brinkhoff <lars.spam@nocrew.org>
Date2015-12-10 09:32 +0100
Message-ID<866106g0vr.fsf@vps34351.public.cloudvps.com>
In reply to#44168
Lars Brinkhoff <lars.spam@nocrew.org> writes:
> Bernd Paysan wrote:
> > BTW: I still have to write up all the steps to get a successful
> > build.  However, this requires third parties to test it, people who
> > don't have everything readily installed like I do.
>
> How about setting up a continuous integration build?  If you send the
> instructions to me, I can do that for you.

I made an attemp myself, and created a very simple build.  Running gfgen
failed unless I installed an earlier version of Gforth.  Is there
another way to boostrap?

https://travis-ci.org/larsbrinkhoff/gforth/builds/95980624

[toc] | [prev] | [next] | [standalone]


#44177

Fromanton@mips.complang.tuwien.ac.at (Anton Ertl)
Date2015-12-10 13:34 +0000
Message-ID<2015Dec10.143439@mips.complang.tuwien.ac.at>
In reply to#44170
Lars Brinkhoff <lars.spam@nocrew.org> writes:
>Lars Brinkhoff <lars.spam@nocrew.org> writes:
>> Bernd Paysan wrote:
>> > BTW: I still have to write up all the steps to get a successful
>> > build.  However, this requires third parties to test it, people who
>> > don't have everything readily installed like I do.
>>
>> How about setting up a continuous integration build?  If you send the
>> instructions to me, I can do that for you.
>
>I made an attemp myself, and created a very simple build.  Running gfgen
>failed unless I installed an earlier version of Gforth.  Is there
>another way to boostrap?

Yes, provide the results of gfgen (and give them timestamps that are
newer than the files they depend on).  That's what we do when we
distribute tarballs.

There is no way to build Gforth from the repository without having a
working Gforth.

- anton
-- 
M. Anton Ertl  http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
     New standard: http://www.forth200x.org/forth200x.html
   EuroForth 2015: http://www.rigwit.co.uk/EuroForth2015/

[toc] | [prev] | [next] | [standalone]


#44183 — Gforth automated build

FromLars Brinkhoff <lars.spam@nocrew.org>
Date2015-12-11 08:54 +0100
SubjectGforth automated build
Message-ID<86y4d1e7z4.fsf_-_@vps34351.public.cloudvps.com>
In reply to#44170
Lars Brinkhoff wrote:
> Bernd Paysan wrote:
>> BTW: I still have to write up all the steps to get a successful
>> build.  However, this requires third parties to test it, people who
>> don't have everything readily installed like I do.
>
> How about setting up a continuous integration build?  If you send the
> instructions to me, I can do that for you.

Looks great!

https://travis-ci.org/forthy42/gforth

There are also ways to do 32-bit builds, OS X, Windows, iOS, Android,
etc.

[toc] | [prev] | [next] | [standalone]


#44197 — Re: Gforth automated build

FromBernd Paysan <bernd.paysan@gmx.de>
Date2015-12-11 21:52 +0100
SubjectRe: Gforth automated build
Message-ID<n4fcun$1dq$1@dont-email.me>
In reply to#44183
Lars Brinkhoff wrote:

> Lars Brinkhoff wrote:
>> Bernd Paysan wrote:
>>> BTW: I still have to write up all the steps to get a successful
>>> build.  However, this requires third parties to test it, people who
>>> don't have everything readily installed like I do.
>>
>> How about setting up a continuous integration build?  If you send the
>> instructions to me, I can do that for you.
> 
> Looks great!
> 
> https://travis-ci.org/forthy42/gforth
> 
> There are also ways to do 32-bit builds, OS X, Windows, iOS, Android,
> etc.

The Android build is pretty demanding (native code compiling for all 5 
architectures and also building the required libraries), and unlike the 
normal, non-cross-compilation build (which just runs "make check" for self-
testing), I don't have anything like an automated test yet.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
net2o ID: kQusJzA;7*?t=uy@X}1GWr!+0qqp_Cn176t4(dQ*
http://bernd-paysan.de/

[toc] | [prev] | [standalone]


Back to top | Article view | comp.lang.forth


csiph-web