r3wp [groups: 83 posts: 189283]
  • Home
  • Script library
  • AltME Archive
  • Mailing list
  • Articles Index
  • Site search
 

World: r3wp

[Core] Discuss core issues

Graham
14-Jul-2006
[5059x4]
run 	Runs the system application associated with a file.
 	
I'm sure that is not working ...
http://www.rebol.org/cgi-bin/cgiwrap/rebol/ml-display-thread.r?m=rmlPRTS
on how to set the environment variables in win32
using the win32 api
hmm.  Doesn't work.
BrianH
14-Jul-2006
[5063]
Graham, there are two good reasons for that: Security and portability. 
Some platforms have one environment, some have per-process, some 
have global and per-user (like Windows) - which environment do you 
want to set? As for security, if you set any variable other than 
per-process it can affect the behavior of other programs, an ability 
that should be restricted in a sandboxed environment.


You should check out command line apps that you can call to set the 
various environments on your platform. If the REBOL process doesn't 
have call because of security restrictions, it shouldn't be able 
to set environment variables anyways.
Volker
14-Jul-2006
[5064]
run is available in ios. Somewhere i read in windows one can type 
a filename in shell, like text.txt, and windows launches the accosiated 
application. If that works, maybe with 'call too?
Graham
14-Jul-2006
[5065]
I think the lack of 'run in other than IOS is some type of oversight.
Volker
14-Jul-2006
[5066]
I agree
Graham
14-Jul-2006
[5067]
I wonder why using the windows 32 api, I still can't set the user 
environment variables
Volker
14-Jul-2006
[5068]
In linux they are copied for each process, not shared AFAIK. MAybe 
windows does the same?
Gregg
15-Jul-2006
[5069]
; Does this work for you Graham?

REBOL []

; GET-ENV is a standard REBOL function now

; environment variable APIs
; msvcrt.dll
; getenv _putenv _environ
; char *getenv( const char *varname );
; int _putenv( const char *envstring );

lib:  load/library %msvcrt.dll

get-env: make routine! [
	varname [string!]
	return: [string!]
] lib "getenv"

put-env: make routine! [
	env-string [string!]
	return:    [integer!]
] lib "_putenv"

remove-env-var: func [name [string!]] [put-env join name "="]

env-var-exists?: func [name [string!]] [
	either "^@" = get-env name [false][true]
]

tz-set: make routine! [
	return:    [integer!]
] lib "_tzset"

print get-env "path"
print get-env "lib"
print get-env "temp"
print get-env "test"

if 0 <> put-env "test=blah" [
	print "error writing environment variable"
]
print get-env "test"
remove-env-var "test"
print mold get-env "test"

print get-env "TZ"
tz-set

free lib

halt
Graham
15-Jul-2006
[5070]
No, it doesn't.  It appears to .. but as soon as you quit the rebol 
session and see if the env variable is still set, it's not.
BrianH
16-Jul-2006
[5071]
What you are doing there is setting the per-process environment. 
To set any of the other environments you need some other, somewhat 
more complex APIs. Or some much more simple command line apps.
Graham
16-Jul-2006
[5072]
I guess that's why we have get-env and not set-env
BrianH
18-Jul-2006
[5073x2]
I have a few command line apps that do the job, but the best one 
I've found I got from a web site that isn't there any more. If you 
want it PM me and I will email it to you. Otherwise, try setx.exe 
from the Windows Resource Kit - it can do all sorts of stuff.
There are 4 environments on Windows (aside from the per-process ones), 
the system, user, default user and volatile environments. The setenv.exe 
I found can set any of these 4, the setx.exe in the Resource Kit 
can set the system and user environments.
Graham
18-Jul-2006
[5075]
I'm only looking at present to set the current user environment
BrianH
18-Jul-2006
[5076x2]
Do you mean the current user's initial environment, or something 
less lasting?
Or do you mean the REBOL process' environment that is inherited by 
the subprocesses started by CALL (assuming that CALL internally passes 
along the current environment to its subprocesses)? Or do you mean 
the environment of the parent process?


Every started process is passed an environment, usually a copy of 
the parent environment (sometimes with some modifications). On Windows 
(NT kernel, not 9x), the initial environment is a combination of 
variables associated with the system (or machine), the user and volatile 
values, in that order. The initial values of these variables are 
constructed from data in the registry. Once these variables are constructed 
and compiled into an environment, this environment is passed to a 
process. Changes to the environment of that process (with getenv 
and setenv) don't affect the environment of the parent processes, 
and certainly don't affect the global values.


To change the initial environment variables, you need to change them 
in their original registry entries. You can either do that directly 
or through using external applications. Keep in mind that changes 
to these initial values won't affect your current environment, or 
those of any running processes, as those environments are already 
set and can only be changed internally.
Graham
18-Jul-2006
[5078x2]
I want to change the user environment strings that othewise I have 
to do in the control panel -> system -> advanced
so that they remain changed when I next login
BrianH
18-Jul-2006
[5080]
Do you need to support Win9x, or just WinNT derivatives like 2000, 
XP and 2003?
Graham
18-Jul-2006
[5081x2]
Just Win32
and Win64
BrianH
18-Jul-2006
[5083]
Windows 95 is Win32 as well, but its environment handling is completely 
different. My question still stands :)
Graham
18-Jul-2006
[5084]
Ok, 2000, XP, 2003
BrianH
18-Jul-2006
[5085x4]
Then it's registry values you need to change. Give me a moment and 
I'll figure out which ones.
Environment variables are based on values of the following registry 
keys, for each type:

- Local machine: HKLM\SYSTEM\CurrentControlSet\Control\Session Manager\Environment
- Current user: \Environment
- Default user: HKEY_USERS\.DEFAULT\Environment

- Volatile (probably should be treated as readonly): HKCU\Volatile 
Environment
Sorry, mistype...
- Current user: HKCU\Environment
You can check in regedit for your current values. Remember to use 
REG_EXPAND_SZ values if you want references to other environment 
variables to be expanded, but keep in mind that these are evaluated 
in one pass for each category, and that local machine is evaluated 
before current user. A value can't make references to other variables 
in its own category, just references to values in other categories 
that are evaluated earlier.
Graham
18-Jul-2006
[5089x3]
found them!
My Computer\HKEY_CURRENT\Environment
and else where :(
BrianH
18-Jul-2006
[5092]
The environments you are likely interested in are in the locations 
I specified. If you want to make changes to all users' local environments, 
look at the other keys under HKEY_USERS. The .DEFAULT key is what 
gets copied to a new user's local registry when that user is created.
Graham
18-Jul-2006
[5093]
Ok.
BrianH
18-Jul-2006
[5094x2]
Obviously you will need admin priviledges for any keys not under 
HKCU.
HKEY_CLASSES_ROOT is one of those fake, mirror keys.
Sunanda
18-Jul-2006
[5096]
. (apologies for the dot, but the web shows this group has failed 
to sync on my machine)
Louis
26-Jul-2006
[5097x2]
Hi everyone, I am presently in Indonesia. I can send email to myself, 
but when I try to send email to other I get the following error message:


** User Error: Server error: tcp 550-(LATURK-WS-2) [61.94.196.24] 
is currently not permitted
 to relay through

Is this a problem I can correct in the script? If yes, how?
I am able to email using Eudora just fine.
Graham
26-Jul-2006
[5099x5]
nothing to do with your script
hmm.
maybe you need to authenticate first before sending mail.
So, use rebol to read your mail first and then try sending.
since you're not getting a smtp authentication error.
Louis
26-Jul-2006
[5104x2]
Ok, I'll try that and get back with the results.
That fixed it. Thank you so much, Graham!
Henrik
26-Jul-2006
[5106x2]
how does one avoid that variables listed in a block in an object 
become global in this instance:

>> unset [a b c]
>> y: make object! [a: [b c d] set a [1 2 3]]
>> a
** Script Error: a has no value
** Near: a
>> b
== 1
>> c
== 2
>> d
== 3

These give the same result:

>> y: make object! [a: [b c d] set bind a self [1 2 3]]
>> y: make object! [a: [b c d] set bind/copy a self [1 2 3]]
>> y: make object! [a: copy [b c d] set a [1 2 3]]
>> y: make object! [a: bind [b c d] self set a [1 2 3]]
hm... seem to be the old "how to add new variables to an existing 
object" problem. looks like I'll have to find a different approach...
Ladislav
27-Jul-2006
[5108]
do you know this one?

d: [a b c]
d: use d reduce [d]
set d [1 2 3]