1. enhancing prompt() functions
- Posted by euphoric (admin) Sep 15, 2011
- 1086 views
I would like to enhance the prompt_string() function to allow the user to supply a default or initial string to be used.
For example:
uname = "Steve Jones" -- could be retrieved from a database, prefs file, or whatever uname = prompt_string( "Enter username: ", uname )
Or is there another/a better way to go about this?
2. Re: enhancing prompt() functions
- Posted by jimcbrown (admin) Sep 15, 2011
- 1072 views
I would like to enhance the prompt_string() function to allow the user to supply a default or initial string to be used.
For example:
uname = "Steve Jones" -- could be retrieved from a database, prefs file, or whatever uname = prompt_string( "Enter username: ", uname )
Or is there another/a better way to go about this?
This approach is quite difficult, because we just use gets(0). Even if we fall back to C code, to handle fgets() we'd have to use ungetc() to push back the characters in the initial string - and ungetc() only guarantees one push back (thus one character).
So, we have to drop gets() and use wait_key() and puts() and handle the edit processing ourselves. This is a lot more complicated than the existing prompt_string() function. OTOH, ed.ex already does this.
3. Re: enhancing prompt() functions
- Posted by mattlewis (admin) Sep 15, 2011
- 1056 views
I would like to enhance the prompt_string() function to allow the user to supply a default or initial string to be used.
For example:
uname = "Steve Jones" -- could be retrieved from a database, prefs file, or whatever uname = prompt_string( "Enter username: ", uname )
Or is there another/a better way to go about this?
This approach is quite difficult, because we just use gets(0). Even if we fall back to C code, to handle fgets() we'd have to use ungetc() to push back the characters in the initial string - and ungetc() only guarantees one push back (thus one character).
So, we have to drop gets() and use wait_key() and puts() and handle the edit processing ourselves. This is a lot more complicated than the existing prompt_string() function. OTOH, ed.ex already does this.
I'm not sure why we'd need to do all of that. Using the example above, it might look like:
Enter username: [Steve Jones]Then, if you simply press enter, it's as though you entered Steve Jones. Otherwise, whatever the user enters is returned.
Matt
4. Re: enhancing prompt() functions
- Posted by euphoric (admin) Sep 15, 2011
- 1013 views
I would like to enhance the prompt_string() function to allow the user to supply a default or initial string to be used.
This approach is quite difficult, because we just use gets(0).
Yes, I discovered that after looking at the function. I'm guessing a roll-your-own solution will have to do for now.
Thank you! :)
5. Re: enhancing prompt() functions
- Posted by euphoric (admin) Sep 15, 2011
- 1062 views
I would like to enhance the prompt_string() function to allow the user to supply a default or initial string to be used.
For example:
uname = "Steve Jones" -- could be retrieved from a database, prefs file, or whatever uname = prompt_string( "Enter username: ", uname )
Or is there another/a better way to go about this?
This approach is quite difficult, because we just use gets(0). Even if we fall back to C code, to handle fgets() we'd have to use ungetc() to push back the characters in the initial string - and ungetc() only guarantees one push back (thus one character).
So, we have to drop gets() and use wait_key() and puts() and handle the edit processing ourselves. This is a lot more complicated than the existing prompt_string() function. OTOH, ed.ex already does this.
I'm not sure why we'd need to do all of that. Using the example above, it might look like:
Enter username: [Steve Jones]Then, if you simply press enter, it's as though you entered Steve Jones. Otherwise, whatever the user enters is returned.
I don't see an easy way to get "Steve Jones" into the buffer, which I think is what would have to happen.
I'm going to try a puts(0,default) and see if that works. :)
Edit: It doesn't work. :/
6. Re: enhancing prompt() functions
- Posted by jimcbrown (admin) Sep 15, 2011
- 1051 views
I'm not sure why we'd need to do all of that. Using the example above, it might look like:
Enter username: [Steve Jones]Then, if you simply press enter, it's as though you entered Steve Jones. Otherwise, whatever the user enters is returned.
Matt
Hmm. I guess that's fine.
The user couldn't edit the default string (e.g. change it from "Steve Jones" to "Steve Jone") but would either have to use it as-is or type in the whole thing from scratch, but that might not be a problem.
7. Re: enhancing prompt() functions
- Posted by euphoric (admin) Sep 15, 2011
- 1057 views
I'm not sure why we'd need to do all of that. Using the example above, it might look like:
Enter username: [Steve Jones]Then, if you simply press enter, it's as though you entered Steve Jones. Otherwise, whatever the user enters is returned.
The user couldn't edit the default string (e.g. change it from "Steve Jones" to "Steve Jone") but would either have to use it as-is or type in the whole thing from scratch, but that might not be a problem.
Oh, I see what you're saying. heh.
8. Re: enhancing prompt() functions
- Posted by jimcbrown (admin) Sep 15, 2011
- 1057 views
I'm going to try a puts(0,default) and see if that works. :)
That's hard to do (even in C) for reasons I previously stated.
On the Windows (non-MinGW) side, we can add support in be_runtime.c's key_gets() function. We already appear to be handling editing in there, so we'd just need to add a pushback buffer.
We don't use key_gets() on MinGW or POSIX, however. On MinGW we use the W32API console api, so it might be possible to push back characters that way.
9. Re: enhancing prompt() functions
- Posted by euphoric (admin) Sep 15, 2011
- 1039 views
This seems to work OK:
public function prompt_string(sequence prompt, sequence default = "") object answer puts(1, prompt) if length(default) > 0 then puts(1," [" & default & "] ") end if answer = gets(0) puts(1, '\n') if sequence(answer) and length(answer) > 0 then answer = answer[1..$-1] -- trim the \n end if if length(answer) > 0 then return answer else return default end if end function
To test:
sequence s = prompt_string( "Enter username:", "Billy Ray Jones") puts(1,"\n'" & s & "'\n\n")
10. Re: enhancing prompt() functions
- Posted by mattlewis (admin) Sep 15, 2011
- 1061 views
I'm not sure why we'd need to do all of that. Using the example above, it might look like:
Enter username: [Steve Jones]Then, if you simply press enter, it's as though you entered Steve Jones. Otherwise, whatever the user enters is returned.
I don't see an easy way to get "Steve Jones" into the buffer, which I think is what would have to happen.
I'm going to try a puts(0,default) and see if that works. :)
Edit: It doesn't work. :/
Well, in my console (konsole), I can copy and paste, and edit whatever's there. And then I get the end result:
include std/console.e printf(1, "you entered: [%s]\n", { prompt_string("enter something:") })
Matt
11. Re: enhancing prompt() functions
- Posted by euphoric (admin) Sep 15, 2011
- 1089 views
I guess I could brush off my Textmode Dialogs lib. I think it has all sorts of functionality in there, even text prompts with default values... IIRC. :D
12. Re: enhancing prompt() functions
- Posted by jeremy (admin) Sep 15, 2011
- 1079 views
I guess I could brush off my Textmode Dialogs lib. I think it has all sorts of functionality in there, even text prompts with default values... IIRC. :D
If it works across platforms, it would be a nice addition to the std lib.
Jeremy
13. Re: enhancing prompt() functions
- Posted by euphoric (admin) Sep 15, 2011
- 1061 views
I guess I could brush off my Textmode Dialogs lib. I think it has all sorts of functionality in there, even text prompts with default values... IIRC. :D
If it works across platforms, it would be a nice addition to the std lib.
Jeremy
I have an updated version on my PC. I've squashed some bugs and will make it available to others for testing.
It's based on Irv Mullins' original text-mode dialogs library, just with some added usability enhancements.
14. Re: enhancing prompt() functions
- Posted by DerekParnell (admin) Sep 15, 2011
- 1055 views
I'm going to try a puts(0,default) and see if that works. :)
That's hard to do (even in C) for reasons I previously stated.
On the Windows (non-MinGW) side, we can add support in be_runtime.c's key_gets() function. We already appear to be handling editing in there, so we'd just need to add a pushback buffer.
We don't use key_gets() on MinGW or POSIX, however. On MinGW we use the W32API console api, so it might be possible to push back characters that way.
I'm currently rewritting the console input functionality so that in Windows (regardless of compiler) it just uses the Windows API and thus putting characters into the console buffer is quite easy to do.
15. Re: enhancing prompt() functions
- Posted by jimcbrown (admin) Sep 15, 2011
- 1065 views
I'm going to try a puts(0,default) and see if that works. :)
That's hard to do (even in C) for reasons I previously stated.
On the Windows (non-MinGW) side, we can add support in be_runtime.c's key_gets() function. We already appear to be handling editing in there, so we'd just need to add a pushback buffer.
We don't use key_gets() on MinGW or POSIX, however. On MinGW we use the W32API console api, so it might be possible to push back characters that way.
I'm currently rewritting the console input functionality so that in Windows (regardless of compiler) it just uses the Windows API and thus putting characters into the console buffer is quite easy to do.
Right now it's in three places. That welcome rewrite would bring it down to two.
