Mailing List Archive: 49091 messages
  • Home
  • Script library
  • AltME Archive
  • Mailing list
  • Articles Index
  • Site search
 

Morpheus (was: set-net question. No, TWOquestions)

 [1/10] from: petr:krenzelok:trz:cz at: 30-Oct-2001 10:03


Hallvard Ystad wrote:
> Petr Krenzelok skrev (Tuesday 30.10.2001, kl. 08.57): > > > >Inside the 'open func in system/schemes/http/handler I find the
<<quoted lines omitted: 9>>
> >http-accept-header?? > I mean the Accept: "*/*" header, as seen above. Certain browsers cannot
huh, what is the problem of re-setting the word upon your need then? Although 'create-request seems to be pretty nested, you can: write %my-http-handler.txt mold system/schemes/http/handler ...then change its source, and rebind it back in your user.r file: system/schemes/http/handler: load %my-http-handler.txt But I am not sure if it is secure way to do so, as RT can change handler code in future releases ... One other note - do you know the difference in none vs false in protocol schemes settings? If you use e.g. system/schemes/ftp/user: none, it will look at parent scheme - system/schemes/default for 'user field setting, and if present, it will override your none above. If you want to prevent it, just use system/schemes/ftp/user: false e.g. I hope I do remember it correctly :-) Cheers, -pekr-

 [2/10] from: hallvard:ystad:helpinhand at: 30-Oct-2001 11:12


Petr Krenzelok skrev (Tuesday 30.10.2001, kl. 10.03):
>huh, what is the problem of re-setting the word upon your need then?
I wanted a simple, neat way to do it. I'm inside a CGI script that fetches a url, extracts some information and writes it out. I could rebind system/schemes/http/handler in user.r, but the accpet-setting isn't static! I find it very... bothersome! to have to rebind system/schemes/http/handler several times in the script.
>But I am not sure if it is secure way to do so, as RT can change handler >code >in future releases ...
..and that's another point, yes.
>One other note - do you know the difference in none vs false in protocol >schemes settings? If you use e.g. system/schemes/ftp/user: none, it will
<<quoted lines omitted: 3>>
>use >system/schemes/ftp/user: false e.g. I hope I do remember it correctly :-)
I'll check it out. Thnaks- ~H

 [3/10] from: cyphre::seznam::cz at: 30-Oct-2001 12:35


Hi Halvard, here you are two(either for /core 2.5 or /View1.2.1) little ugly hard-patches for your needs: --------try this on start of your View script-------- poke first at first at second get in system/schemes/http/handler 'open 17 4 2 "your-new-value" --------end of hack for REBOL/View 1.2.1.3.1 21-Jun-2001-------- --------try this on start of your Core script-------- poke first at second get in system/schemes/http/handler 'open 15 2 your-new-value --------end of hack REBOL/Core 2.5.0.3.1-------- Althourgh I didn't test them I hope this will change HTTP-Get-Header/Accept value as you want. Let me know if it works. I'm just curious ;) regards Cyphre

 [4/10] from: brett:codeconscious at: 30-Oct-2001 23:10


I just worked out how to test it - run a little local rebol webserver that shows the connection requests. The following successfully changed the user-agent field of the header. Very rebol ;-) read/custom http://localhost:8000/ [header [User-Agent: "bretts agent"]] Brett.

 [5/10] from: hallvard:ystad:helpinhand at: 30-Oct-2001 13:06


Cyphre skrev (Tuesday 30.10.2001, kl. 12.35):
>poke first at second get in system/schemes/http/handler 'open 15 2 >"your-new-value"
Worked fine, thanks. Setting headers like this is a bit dumb. I even feel that 'send should be simpler. When I want to send an email with, say "X-diddely-do: hello!" in the header, I am forced to make a complete header-object, thus: send/header [hallvard--ystad--helpinhand--com] "test" make object! [X-diddely-do: "hello!" from: none to: none date: none] The three nones are there for rebol to use my default settings. A simpler and more reasonable way to do it, in my opinion, would be like this: send/header [hallvard--ystad--helpinhand--com] "test" [X-diddely-do: "hello!"] Rebol could then simply add [X-diddely-do: "hello!"] to the other headers. In the same way, read could have a refinement like this: read/header http://www.rebol.com/ [Accept: "text/html" user-agent: "007"] I hope something like this will be implemented in future releases... ~H PS. Is there any good documentation for the /custom refinement block dialect in 'read anywhere?

 [6/10] from: deadzaphod:flyingparty at: 30-Oct-2001 4:20


> Setting headers like this is a bit dumb. I even feel that 'send should be > simpler. When I want to send an email with, say "X-diddely-do: hello!" in
<<quoted lines omitted: 4>>
> and more reasonable way to do it, in my opinion, would be like this: > send/header [hallvard--ystad--helpinhand--com] "test" [X-diddely-do: "hello!"]
well, we can already do: send/header [hallvard--ystad--helpinhand--com] "test" make system/standard/email [X-diddely-do: "hello!"] it might be better if send/header did the make implicitly, but this isn't too bad (once you find the standard email object). Cal

 [7/10] from: petr:krenzelok:trz:cz at: 30-Oct-2001 13:29


Cal Dixon wrote:
> > Setting headers like this is a bit dumb. I even feel that 'send should be > > simpler. When I want to send an email with, say "X-diddely-do: hello!" in
<<quoted lines omitted: 10>>
> send/header [hallvard--ystad--helpinhand--com] "test" make > system/standard/email [X-diddely-do: "hello!"]
OK, then everything is solvable - put following into your user.r using: func [blk][make system/standard/email blk] send/header [hallvard--ystad--helpinhand--com] "test" using [X-diddely-do: "hello!"] btw: as for 3 nones - you don't need to use none 3 times. Rebol allowes you to use following construct too - context [a: b: c: d: none] to explicitly set all values ... -pekr-

 [8/10] from: hallvard:ystad:helpinhand at: 30-Oct-2001 13:33


Brett Handley skrev (Tuesday 30.10.2001, kl. 13.10):
>The following successfully changed the user-agent field of the header. >Very >rebol ;-) > read/custom http://localhost:8000/ [header [User-Agent: "bretts > agent"]]
That's funny, because it didn't change anything when I tried. What could be the difference? I'm testing from a win2k machine using /core version 2.5.0.3.1. ~H

 [9/10] from: hallvard:ystad:helpinhand at: 30-Oct-2001 13:36


Cal Dixon skrev (Tuesday 30.10.2001, kl. 13.20):
>well, we can already do: > > send/header [hallvard--ystad--helpinhand--com] "test" make >system/standard/email [X-diddely-do: "hello!"]
Sounds real good, but:
>> send/header [hallvard--ystad--helpinhand--com] "test" make
system/standard/email [X-di ddely-do: "hello!"] connecting to: helpinhand.com ** Access Error: Port none not open ** Where: confirm ** Near: result: make string! (20 * length? words) foreach ~H

 [10/10] from: lmecir:mbox:vol:cz at: 30-Oct-2001 13:43


Hi, how about: send/header [hallvard--ystad--helpinhand--com] "test" make system/standard/email [X-diddely-do: "hello!"] or: sh: func [email [email!] text [string!] spec [block!]] [ send/header email text make system/standard/email block ] and then sh [hallvard--ystad--helpinhand--com] "test" [X-diddely-do: "hello!"] Cheers Ladislav

Notes
  • Quoted lines have been omitted from some messages.
    View the message alone to see the lines that have been omitted