Re: [racket-users] Re: preferences file

[Geoffrey Knauth]

William, that's a neat idea, I'll try your package.  Thanks!  --Geoff

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] raco planet open tries to install?

[Neil Van Dyke]

If it ever helps, the ancient Quack package for Emacs has a convenient 
and safe mode for quickly inspecting ".plt" files (page through them, 
isearch):


http://www.neilvandyke.org/temporary/20171102-quack-plt-screenshot.png
http://www.neilvandyke.org/quack/

(So ancient, I just noticed, that I have it using non-antialiased fonts.)

BTW, it should probably be noted, for the benefit of non-old-timers on 
the list, that PLaneT is the old package system, and is deprecated.  So 
no need to go looking at it as something you should possibly be using.  
Everyone should be using the new package system whenever possible.


Regarding untrusted code: that's an unsolved problem in general.  I was 
hoping for some kind of distributed trust and auditing model for the 
evolution of PLaneT, addressing what I'd say was one of the two main 
drawbacks of PLaneT for third-party developers.  And it would be really 
helpful for some kinds of third-party contributors if we restore, to the 
new package system, some support for low-friction 
non-backward-compatible versions.


--
You received this message because you are subscribed to the Google Groups "Racket 
Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] raco planet open tries to install?

[Robby Findler]

I think that, at some point, we changed things so that planet declines
to run arbitrary code in the .plt file. I'm not sure how this works,
tho, it could just check that the code is the known code and then runs
it. (Or I could misremember and we merely planned to do that.)

Robby

On Thu, Nov 2, 2017 at 5:49 PM, 'John Clements' via Racket Users
 wrote:
> Ooh… yikes. Looking at the code, it looks like PLT files include their own 
> unpackers? Is this correct? In this case, I understand that this could be a 
> problem. This also seems like a bad idea; I wouldn’t have expected unpacking 
> to run potentially untrusted code.
>
> Anyhow, that’s all in the past, IIUC. Probably easiest just to dig up an 
> older version of Racket to install this .plt file.
>
> John
>
>
>
>> On Nov 2, 2017, at 3:46 PM, 'John Clements' via Racket Users 
>>  wrote:
>>
>> I’m surprised by this interaction with raco planet:
>>
>> % raco planet open ../margrave3-dev-0203.plt .
>> cannot install; version (400) of collection ("mzscheme") is required, but 
>> version  is installed
>>  context...:
>>   /Users/clements/racket/racket/collects/setup/unpack.rkt:408:31: loop
>>   
>> /Users/clements/racket/racket/collects/setup/../racket/private/map.rkt:57:19:
>>  loop
>>   /Users/clements/racket/racket/collects/setup/unpack.rkt:274:7
>>   /Users/clements/racket/racket/share/pkgs/planet-lib/planet/raco.rkt: 
>> [running body]
>>   /Users/clements/racket/racket/collects/raco/raco.rkt: [running body]
>>   /Users/clements/racket/racket/collects/raco/main.rkt: [running body]
>> %
>>
>> Specifically, it appears to me that “open” shouldn’t perform checks 
>> associated with installation. Before diving into the code for PLaneT…is this 
>> a bug? Or am I misunderstanding how PLaneT is (was) supposed to work?
>>
>> Thanks!
>>
>> John
>>
>>
>>
>> --
>> You received this message because you are subscribed to the Google Groups 
>> "Racket Users" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to racket-users+unsubscr...@googlegroups.com.
>> For more options, visit https://groups.google.com/d/optout.
>
>
>
> --
> You received this message because you are subscribed to the Google Groups 
> "Racket Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to racket-users+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] raco planet open tries to install?

['John Clements' via Racket Users]

Ooh… yikes. Looking at the code, it looks like PLT files include their own 
unpackers? Is this correct? In this case, I understand that this could be a 
problem. This also seems like a bad idea; I wouldn’t have expected unpacking to 
run potentially untrusted code.

Anyhow, that’s all in the past, IIUC. Probably easiest just to dig up an older 
version of Racket to install this .plt file.

John



> On Nov 2, 2017, at 3:46 PM, 'John Clements' via Racket Users 
>  wrote:
> 
> I’m surprised by this interaction with raco planet:
> 
> % raco planet open ../margrave3-dev-0203.plt .
> cannot install; version (400) of collection ("mzscheme") is required, but 
> version  is installed
>  context...:
>   /Users/clements/racket/racket/collects/setup/unpack.rkt:408:31: loop
>   
> /Users/clements/racket/racket/collects/setup/../racket/private/map.rkt:57:19: 
> loop
>   /Users/clements/racket/racket/collects/setup/unpack.rkt:274:7
>   /Users/clements/racket/racket/share/pkgs/planet-lib/planet/raco.rkt: 
> [running body]
>   /Users/clements/racket/racket/collects/raco/raco.rkt: [running body]
>   /Users/clements/racket/racket/collects/raco/main.rkt: [running body]
> %
> 
> Specifically, it appears to me that “open” shouldn’t perform checks 
> associated with installation. Before diving into the code for PLaneT…is this 
> a bug? Or am I misunderstanding how PLaneT is (was) supposed to work?
> 
> Thanks!
> 
> John
> 
> 
> 
> -- 
> You received this message because you are subscribed to the Google Groups 
> "Racket Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to racket-users+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.



-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[racket-users] raco planet open tries to install?

['John Clements' via Racket Users]

I’m surprised by this interaction with raco planet:

% raco planet open ../margrave3-dev-0203.plt .
cannot install; version (400) of collection ("mzscheme") is required, but 
version  is installed
  context...:
   /Users/clements/racket/racket/collects/setup/unpack.rkt:408:31: loop
   
/Users/clements/racket/racket/collects/setup/../racket/private/map.rkt:57:19: 
loop
   /Users/clements/racket/racket/collects/setup/unpack.rkt:274:7
   /Users/clements/racket/racket/share/pkgs/planet-lib/planet/raco.rkt: 
[running body]
   /Users/clements/racket/racket/collects/raco/raco.rkt: [running body]
   /Users/clements/racket/racket/collects/raco/main.rkt: [running body]
%

Specifically, it appears to me that “open” shouldn’t perform checks associated 
with installation. Before diving into the code for PLaneT…is this a bug? Or am 
I misunderstanding how PLaneT is (was) supposed to work?

Thanks!

John



-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] Re: preferences file

[William G Hatch]

Recent discussion thread: 
https://groups.google.com/forum/#!topic/racket-users/6hn9J-r0Nek


Since we're discussing dotfiles and I missed the previous thread, I'm
going to shamelessly promote my basedir package[1], which provides
functions for conveniently reading/writing config files according to
the XDG base directory specification (eg. in ~/.config/, or similar
directory on Windows) respecting the standard environment variables.
Using the standard also makes it more convenient for a user to have
parts of their configuration in multiple files, eg. having both public
and private components, as well as local and global components, and
have the parts composed together.

Let's not clutter up home directories anymore with one-off config
files or directories.  Let's let them live in a sane and configurable
space. 

[1] http://docs.racket-lang.org/basedir/index.html

--
You received this message because you are subscribed to the Google Groups "Racket 
Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[racket-users] Bluetooth LE - Typed Racket

[Ray Racine]

While deep lurking I noticed a mention concerning some older Racket code of
mine.  Specifically porting sha-256 from typed to untyped racket.

Well that led to a wiki page that listed a number of suggested starter
tasks for new Racket inductees wanting to get involved.  A number of those
were in and around IOT and Bluetooth (Wii controller is one I recall).  So
while hesitant to dump yet more unfinished Racket code onto the
unsuspecting, I think there maybe enough in a side project in hiatus that
someone may find a modicum of utility in the dross.

Thora-HCI is a (just started, unfinished) native Typed Racket
implementation of Bluetooth HCI stack.  This is the lowest layer in the
Bluetooth stack which underlies all the other Bluetooth layers and
protocols.  Where it was going was to have just enough of a slice
implemented to support hobbiest Home IOT stuff targeting Bluetooth LE,
specifically GATT -> ATT -> HCI.  By having such a slice one would be able
to do some Raspberry PI, small sensors, light bulb control, robotics etc.

Bluetooth is a big umbrella term for a number of disparate protocols and
Bluetooth vs Bluetooth LE (low energy) vs profile type stuff (audio
streaming or keyboards) are all varying protocols / layers that are build
over HCI.

https://en.wikipedia.org/wiki/List_of_Bluetooth_profiles

Thora-HCI

https://github.com/RayRacine/thora-hci-racket/tree/dev/src

What can Thora-HCI do now to varying degrees:
 - Create a standard Racket I/O Port over an underlying RAW socket
connection on Linux.
 - Open a Bluetooth Adapter.
 - Send a subset of HCI command packets to the Adapter
 - Receive a subset of HCI events from the Adapter.
 - Bluetooth and Bluetooth LE device scans (e.g. scan for Rokus, lightb
ulbs, WII controllers).
 - Create a connection to a device. e.g. connect to a light bulb
 - Send raw command packets to a connected device. e.g Query a BLE sensor,
flip light bulbs on/off.

What Thora-HCI has some pieces of low level machinery for, but is missing
major pieces to do includes things like Bluetooth audio etc.

I can't emphasis enough its unfinished state as well as its scattered
nature from being both in mid-flight and hiatus.  Long story.  It's more of
a code dump that someone could cherry-pick from or start from.  It is alpha
code at best.

FWIW if someone is interested in doing some IOT stuff via Bluetooth (on
Linux) one could also leverage DBus from Racket as well.  One could also do
Racket FFI -> Bluez.  Complicated story but I found things I didn't like
when I went down those paths and I ended up doing the HCI stack layer
implementation. If/when I pickup this effort I would stick with this
approach (native stack impl).

If anyone does attempt to mess around with what is there, I do have time to
assist with questions, maybe even help a bit with something simple they are
trying to do to get things jump started.  Maybe even jump back into it
again.

If poking around in the code.  See scratch.rkt showing code opening an
adapter, scanning, sending commands etc.  In addition focus on adapter.rkt
and what modules / code it leverages.

Caveat Emptor

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] Re: preferences file

[Geoffrey Knauth]

Awesome, Neil, great suggestions!

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] Re: preferences file

[Neil Van Dyke]

For a "home directory dotfile" kind of preferences file that's arbitrary 
Racket code, I'd do `dynamic-require` instead of `require`.


If you want to use `require` instead, maybe the application is a 
framework, and the user-specific `.rkt` file is the program that is run 
and is what `require`s the framework.


Recent discussion thread: 
https://groups.google.com/forum/#!topic/racket-users/6hn9J-r0Nek


If you don't want arbitrary code in your dotfile, then maybe use the 
Racket reader with some of the features disabled.  (When I've done this, 
I went through the current documentation for the reader, and wrote a 
procedure that does the read while explicitly disabling each reader 
feature that was problematic.)  But I prefer the programmer-centric 
arbitrary-code way for many purposes now, like I say in that earlier 
discussion thread.


--
You received this message because you are subscribed to the Google Groups "Racket 
Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] preferences file

[Geoffrey Knauth]

Thanks John!  Good ideas, and no, nothing is too obvious for me!  --Geoff

On Thursday, November 2, 2017 at 12:56:14 PM UTC-4, johnbclements wrote:
>
>
> One solution is to use a .gitignore. I have a convention that files ending 
> with ‘private.rkt’ don’t go in the repo. 
>
> Alternatively, you could have a single directory—a private repo, or not a 
> repo at all—that’s linked as a package, containing a collection called 
> ‘private’ or whatever you like, and then require them by using paths like 
> (require private/dbaccess) 
>
> Apologies if these were all too obvious for you! :) 
>
> John 
>

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[racket-users] Re: preferences file

[Geoffrey Knauth]

The following seems to work for me:

  (require (file "~/.dbaccess.rkt"))

Is that normal / ok ?  Just wondering what other people do.

Geoff

On Thursday, November 2, 2017 at 12:35:25 PM UTC-4, Geoffrey Knauth wrote:
>
> Let's suppose I want to have a preferences file, e.g., "~/.dbaccess.rkt".  
> I used to have a module I would import via 
> "../some-relative-path/dbaccess.rkt", but I thought I would try 
> "~/.dbaccess.rkt" as a module path.  It turned out a module path cannot 
> begin with "/", which happens with the expansion of "~" (it seems).
>
> This leads me to the more general question, what do other people do, what 
> is the convention when, say, they will have a GitHub repo with generally 
> public code, but some private files with, say, database-access codes that 
> are specific to a person's use of a package?
>
> Geoff
>

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] preferences file

['John Clements' via Racket Users]

> On Nov 2, 2017, at 9:35 AM, Geoffrey Knauth  wrote:
> 
> Let's suppose I want to have a preferences file, e.g., "~/.dbaccess.rkt".  I 
> used to have a module I would import via 
> "../some-relative-path/dbaccess.rkt", but I thought I would try 
> "~/.dbaccess.rkt" as a module path.  It turned out a module path cannot begin 
> with "/", which happens with the expansion of "~" (it seems).
> 
> This leads me to the more general question, what do other people do, what is 
> the convention when, say, they will have a GitHub repo with generally public 
> code, but some private files with, say, database-access codes that are 
> specific to a person's use of a package?

One solution is to use a .gitignore. I have a convention that files ending with 
‘private.rkt’ don’t go in the repo.

Alternatively, you could have a single directory—a private repo, or not a repo 
at all—that’s linked as a package, containing a collection called ‘private’ or 
whatever you like, and then require them by using paths like (require 
private/dbaccess)

Apologies if these were all too obvious for you! :)

John



-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[racket-users] preferences file

[Geoffrey Knauth]

Let's suppose I want to have a preferences file, e.g., "~/.dbaccess.rkt".  
I used to have a module I would import via 
"../some-relative-path/dbaccess.rkt", but I thought I would try 
"~/.dbaccess.rkt" as a module path.  It turned out a module path cannot 
begin with "/", which happens with the expansion of "~" (it seems).

This leads me to the more general question, what do other people do, what 
is the convention when, say, they will have a GitHub repo with generally 
public code, but some private files with, say, database-access codes that 
are specific to a person's use of a package?

Geoff

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.