Text IO for files not using the standard line termination for the OS

Text IO for files not using the standard line termination for the OS

[jdgressett]

I use Ada 2012 to write Windows programs; my usual compiler is 32-bit TDM-gcc which implements the gcc-5.1.0 version of the Ada compiler.

My problem is a simple one: i need to produce a program which reads lines from a text file; if a line meets a requirement, it is written to an output file; if it does not meet the requirement it is not written to the output file.

The annoyance that I must deal with is the fact that the files are not Windows text files, which use the <cr><lf> convention to terminate text lines.

The files use a single <cr> as the line terminator.

Is there a way to get Ada.Text_IO to use a line terminator that is not the default for the operation system on which the program runs?

Re: Text IO for files not using the standard line termination for the OS

[Shark8]

On Saturday, November 17, 2018 at 7:01:45 PM UTC-7, jdgre...@hotmail.com wrote:
> I use Ada 2012 to write Windows programs; my usual compiler is 32-bit TDM-gcc which implements the gcc-5.1.0 version of the Ada compiler.
> 
> My problem is a simple one: i need to produce a program which reads lines from a text file; if a line meets a requirement, it is written to an output file; if it does not meet the requirement it is not written to the output file.
> 
> The annoyance that I must deal with is the fact that the files are not Windows text files, which use the <cr><lf> convention to terminate text lines.
> 
> The files use a single <cr> as the line terminator.
> 
> Is there a way to get Ada.Text_IO to use a line terminator that is not the default for the operation system on which the program runs?

I think you have to do it character-by-character in this case.
There's a library called PragmaARC (IIRC) that has Text_IO packages for non-native text, you might want to take a look there.

Re: Text IO for files not using the standard line termination for the OS

[Jeffrey R. Carter]

On 11/18/18 4:44 AM, Shark8 wrote:
> 
> I think you have to do it character-by-character in this case.
> There's a library called PragmaARC (IIRC) that has Text_IO packages for non-native text, you might want to take a look there.

These are the PragmAda Reusable Components. Package PragmARC.Text_IO handles 
files with 3 kinds of line terminators:

    type EOL_ID is (DOS_Windows_EOL, Mac_EOL, Unix_EOL);
    -- Used to specify what line terminator to use on output
    -- DOS_Windows_EOL = CR-LF
    -- Mac_EOL         = CR
    -- Unix_EOL        = LF

For reading, it automatically handles EOLs of all 3 kinds, which may be mixed in 
the same file (though I've never encountered this). For output, when the files 
is created/opened, the desired EOL kind is specified.

The PragmARCs are at

https://github.com/jrcarter/PragmARC

HTH

-- 
Jeff Carter
"Ditto, you provincial putz?"
Blazing Saddles
86

Re: Text IO for files not using the standard line termination for the OS

[jdgressett]

On Sunday, November 18, 2018 at 3:26:31 AM UTC-6, Jeffrey R. Carter wrote:
> On 11/18/18 4:44 AM, Shark8 wrote:
> > 
> > I think you have to do it character-by-character in this case.
> > There's a library called PragmaARC (IIRC) that has Text_IO packages for non-native text, you might want to take a look there.
> 
> These are the PragmAda Reusable Components. Package PragmARC.Text_IO handles 
> files with 3 kinds of line terminators:
> 
>     type EOL_ID is (DOS_Windows_EOL, Mac_EOL, Unix_EOL);
>     -- Used to specify what line terminator to use on output
>     -- DOS_Windows_EOL = CR-LF
>     -- Mac_EOL         = CR
>     -- Unix_EOL        = LF
> 
> For reading, it automatically handles EOLs of all 3 kinds, which may be mixed in 
> the same file (though I've never encountered this). For output, when the files 
> is created/opened, the desired EOL kind is specified.
> 
> The PragmARCs are at
> 
> https://github.com/jrcarter/PragmARC
> 
> HTH
> 
> -- 
> Jeff Carter
> "Ditto, you provincial putz?"
> Blazing Saddles
> 86

That is what I need - thanks.

Re: Text IO for files not using the standard line termination for the OS

[Keith Thompson]

jdgressett@hotmail.com writes:
> I use Ada 2012 to write Windows programs; my usual compiler is 32-bit
> TDM-gcc which implements the gcc-5.1.0 version of the Ada compiler.
>
> My problem is a simple one: i need to produce a program which reads
> lines from a text file; if a line meets a requirement, it is written
> to an output file; if it does not meet the requirement it is not
> written to the output file.
>
> The annoyance that I must deal with is the fact that the files are not
> Windows text files, which use the <cr><lf> convention to terminate
> text lines.
>
> The files use a single <cr> as the line terminator.

That's unusual.  The most recent system system I know of that uses that
representation is pre-OSX MacOS.

> Is there a way to get Ada.Text_IO to use a line terminator that is not
> the default for the operation system on which the program runs?

There's no standard way to do it.  GNAT does provide some non-standard
Form parameters for Text_IO.Create and Text_IO.Open, but none of then do
just what you want.  ("Text_Translation=No" would let you handle
Unix-style text files on Windows, with line endings marked by <lf>.)

http://docs.adacore.com/live/wave/gnat_rm/html/gnat_rm/gnat_rm/the_implementation_of_standard_i_o.html#text-translation

Your best bet might be to translate the files before you feed them to
your program.  Or you can read the files in binary mode.

-- 
Keith Thompson (The_Other_Keith) kst@mib.org  <http://www.ghoti.net/~kst>
Will write code for food.
"We must do something.  This is something.  Therefore, we must do this."
    -- Antony Jay and Jonathan Lynn, "Yes Minister"

Re: Text IO for files not using the standard line termination for the OS

[Simon Wright]

Keith Thompson <kst-u@mib.org> writes:

> Or you can read the files in binary mode.

Or via a stream.