Temporary files in Ada95.

Temporary files in Ada95.

[jboulais]

I am trying to create a temporary output file, that the system would cleanup
after I close it.  In the Ada 83 LRM this is specifically called out as
supplying a blank name.  No mention of this is made in the Ada95 LRM, and
when I try I get a Name_Error.	Any help?

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Temporary files in Ada95.

[Matthew Heaney]

jboulais@sikorsky.com writes:

> I am trying to create a temporary output file, that the system would cleanup
> after I close it.  In the Ada 83 LRM this is specifically called out as
> supplying a blank name.  No mention of this is made in the Ada95 LRM, and
> when I try I get a Name_Error.	Any help?

Are you sure you're calling Create?  The Name parameter is defaulted to
null string, which means temporary file.  See RM95 A.8.2 (4).

Post the smallest fragment of code that exhibits the behavior you
describe, and say what OS and compiler you're using (including version
numbers).

Re: Temporary files in Ada95.

[robert_dewar]

In article <7de1t6$usl$1@nnrp1.dejanews.com>,
  jboulais@sikorsky.com wrote:
> I am trying to create a temporary output file, that the
> system would cleanup
> after I close it.  In the Ada 83 LRM this is specifically
> called out as
> supplying a blank name.  No mention of this is made in
> the Ada95 LRM, and
> when I try I get a Name_Error.	Any help?


The Ada 95 RM is identical to the Ada 83 RM in this
respect, so it is not clear what claim you are making.
Please quote the two reference manuals to support your
claim of an incompatibility.


-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

Re: Temporary files in Ada95.

[Boulais, Jeffrey M]

In response to Mathew, Yes I am sure I am using Create.  I looked at RM95
A.8.2(4).  That is a Sequental_IO
section, and I should have specified the first time that I am trying to do
this under Text_IO.  However, I still
did not see anything under A.8.2(4) that discussed temporary files.  The
reference I made is from
"Ada Programmer's Handbook and LRM" (83), section 14.2.1(3)

Here's what I am trying to do:

Text_IO.Create (
   File => Temp_File ,
   Mode => Text_IO.Out_File ,
   Name => "" ) ;

Now, this technique was kindof working compiling under Gnat3.04, using the
-gnatp switch (suppress runtime checks).
It would create files like "ADA_TEMP_a00068", that were not begin destroyed
upon program completion.
I recently upgraded to Gnat3.11, and this no longer works for me.  I know
this sounds like a Gnat problem at
first, but I have found that Gnat3.04 let a lot of things work that
shouldn't.



   -**** Posted from remarQ, Discussions Start Here(tm) ****-
http://www.remarq.com/ - Host to the the World's Discussions & Usenet

Re: Temporary files in Ada95.

[Boulais, Jeffrey M]

In reponse to Robert,

   I am refernecing a book entitled "Ada Programmer's Handbook and Language
Reference Manual",
by Dean W. Gonzalez, section 14.2.1(3) for the Ada83 LRM.

   For Ada95, I am referencing a book entitled "Ada95 - The Language
Reference Manual & Standard Libraries",
International Standard ANSI/ISO/IEC-8652:1995, section A.10 Text Input-
Output (and on), where I can not find
any detailed infomation on Text_IO.Create, let alone temporary files.  In
A.10.2(2) there is mention of Create
in reference to file modes, but that's it.



   -**** Posted from remarQ, Discussions Start Here(tm) ****-
http://www.remarq.com/ - Host to the the World's Discussions & Usenet

Re: Temporary files in Ada95.

[Jeffrey D. Cherry]

Jeffrey,

Look in the Ada95 manual that you reference in section A.8.2, paragraphs
2, 3, and 4.  You will find the following:

----------
procedure Create(File : in out File_Type;
                 Mode : in File_Mode := default_mode;
                 Name : in String := "";
                 Form : in String := "");

Establishes a new external file, with the given name and form, and
associates this external file with the given file. The given file is
left open. The current mode of the given file is set to the given access
mode. The default access mode is the mode Out_File for sequential and
text input-output; it is the mode Inout_File for direct input-output.
For direct access, the size of the created file is implementation
defined.

A null string for Name specifies an external file that is not accessible
after the completion of the main program (a temporary file). A null
string for Form specifies the use of the default options of the
implementation for the external file.
----------

Except for case changes in the non-reserved words, and breaking up one
paragraph into two, this quote from the Ada95 LRM is identical to
section 14.2.1, paragraphs 2 and 3 of the Ada83 LRM. 

Using GNAT v3.11p, I tried the following:

----------
with Ada.Text_IO;             use Ada.Text_IO;
with Ada.Strings.Unbounded;   use Ada.Strings.Unbounded;

procedure For_Jeff is

   Temp_File  : File_Type;
   Buffer     : string(1..512);
   Last       : natural;
   Saved_Name : Unbounded_String;

begin -- For_Jeff
   Create(File => Temp_File, Mode => Out_File);
   Saved_Name := To_Unbounded_String(Name(Temp_File));
   Put_Line("Temp file name is " & To_String(Saved_Name));
   Put_Line(Temp_File, "This is a temporary file.");
   Reset(Temp_File, In_File);
   Get_Line(Temp_File, Buffer, Last);
   if Last in Buffer'range then
      Put_Line(Buffer(Buffer'first .. Last));
   end if;
   Close(Temp_File);

   Put_Line("Attempting to open " & To_String(Saved_Name) & " ... ");
   begin
      Open(Temp_File, In_File, To_String(Saved_Name));
      Put_Line("Opened file " & To_String(Saved_Name) & "
successfully.");
      Close(Temp_File);
   exception
      when others =>
         Put_Line("File was not opened successfully.");
   end;
end For_Jeff;
----------

Running the above, I got the following output:

----------
Temp file name is D:\s3vvl40h
This is a temporary file.
Attempting to open D:\s3vvl40h ... 
Opened file D:\s3vvl40h successfully.
----------

The file does not exist after the program terminates as you would
expect.  The only odd thing I found is that the temporary file is
created at the root directory of the current drive and not in the
current directory of the current drive.  I would have expected the file
to be created in the current directory or in the system's TEMP
directory.

Regards,

Jeffrey D. Cherry
Logicon Geodynamics

Re: Temporary files in Ada95.

[Matthew Heaney]

jboulais@sikorsky.com (Boulais, Jeffrey M) writes:

> In response to Mathew, Yes I am sure I am using Create.  I looked at
> RM95 A.8.2(4).  That is a Sequental_IO section,

No.  That section applies to _all_ the predefined file packages.


> and I should have specified the first time that I am trying to do this
> under Text_IO.  However, I still did not see anything under A.8.2(4)
> that discussed temporary files.

The number is parens (here, 4) denotes the paragraph number, which
reads:

4   A null string for Name specifies an external file that is not
    accessible after the completion of the main program (a temporary 
    file).

So that paragraph does indeed speak about temporary files.


> The reference I made is from "Ada Programmer's Handbook and LRM" (83),
> section 14.2.1(3)
> 
> Here's what I am trying to do:
> 
> Text_IO.Create (
>    File => Temp_File ,
>    Mode => Text_IO.Out_File ,
>    Name => "" ) ;


The following programs runs without error using GNAT 3.11p, gcc-2.8.1,
linuxppc R4:


with Text_IO;
procedure Test_Create is

   Temp_File : Text_IO.File_Type;

begin

   Text_IO.Create (File => Temp_File,
                   Mode => Text_IO.Out_File,
                   Name => "" );

end;


 
> Now, this technique was kindof working compiling under Gnat3.04, using
> the -gnatp switch (suppress runtime checks).  It would create files
> like "ADA_TEMP_a00068", that were not begin destroyed upon program
> completion.  I recently upgraded to Gnat3.11, and this no longer works
> for me.  I know this sounds like a Gnat problem at first, but I have
> found that Gnat3.04 let a lot of things work that shouldn't.

If you're using GNAT, you can always post a note to the gnat chat list.
ACT employees read that list and may be able to give you more specific
advice as to why you're having a problem.

<mailto:chat@gnat.com>

Re: Temporary files in Ada95.

[Matthew Heaney]

Read the first sentence of RM95 A.8.2 (4).


jboulais@sikorsky.com (Boulais, Jeffrey M) writes:

> In reponse to Robert,
> 
>    I am refernecing a book entitled "Ada Programmer's Handbook and Language
> Reference Manual",
> by Dean W. Gonzalez, section 14.2.1(3) for the Ada83 LRM.
> 
>    For Ada95, I am referencing a book entitled "Ada95 - The Language
> Reference Manual & Standard Libraries",
> International Standard ANSI/ISO/IEC-8652:1995, section A.10 Text Input-
> Output (and on), where I can not find
> any detailed infomation on Text_IO.Create, let alone temporary files.  In
> A.10.2(2) there is mention of Create
> in reference to file modes, but that's it.
> 
> 
> 
>    -**** Posted from remarQ, Discussions Start Here(tm) ****-
> http://www.remarq.com/ - Host to the the World's Discussions & Usenet

Re: Temporary files in Ada95.

[Boulais, Jeffrey M]

I'd like to say "Thanks to all", you were quite helpful.  After closer
inspection, it appears that I was not
reading section A.8.2 closely enough because I thought it delt with
Sequential IO.

However, I did compile the test program supplied by Jeffrey D. Cherry, and I
still get a name error.  This leads
me to believe that this is either a Gnat problem, or a problem with my
install of Gnat.  Thanks again...



   -**** Posted from remarQ, Discussions Start Here(tm) ****-
http://www.remarq.com/ - Host to the the World's Discussions & Usenet

Re: Temporary files in Ada95.

[Larry Kilgallen]

In article <36FBB978.D44CF370@utech.net>, "Jeffrey D. Cherry" <jdcherry@utech.net> writes:

> The file does not exist after the program terminates as you would
> expect.  The only odd thing I found is that the temporary file is
> created at the root directory of the current drive and not in the
> current directory of the current drive.  I would have expected the file
> to be created in the current directory or in the system's TEMP
> directory.

So what happens if the user running the program does not have write
access to the root directory of the current drive ?

Larry Kilgallen

Re: Temporary files in Ada95.

[Jeffrey D. Cherry]

Larry Kilgallen wrote:

> So what happens if the user running the program does not have write
> access to the root directory of the current drive ?

Great question!!  I ran the executable on our NT machine (no one except
the Administrator has write permission for the root directory on any
partition) and got the following output:

raised ADA.IO_EXCEPTIONS.NAME_ERROR

Yuk!  Perhaps there is a way to get the GNAT compiler to use the system
TEMP directory for temporary files?  

Thanks for catching that problem in my code.  I guess it's always better
to be safe than sorry, so if I ever use a temporary file, I won't assume
it will be created safely.  Pardon me while I go slap my forehead and
say "duh".

Regards,
Jeff Cherry

Re: Temporary files in Ada95.

[Boulais, Jeffrey M]

That's exactly what's happening to me.  I am developing on an NT machine
where I am not a system
administrator, perhaps I should have mentioned that.  Is there a way to make
ada not write temporary files
to the root directory?



   -**** Posted from remarQ, Discussions Start Here(tm) ****-
http://www.remarq.com/ - Host to the the World's Discussions & Usenet

Re: Temporary files in Ada95.

[Jeffrey D. Cherry]

Jeffrey,

(Great name BTW.)

I don't think you want to ask "Is there a way to make ada not write
temporary files to the root directory?" since it's not really a language
issue.  I view it as an implementation issue.  I would have thought that
GNAT would create the temporary file in a directory where the user has
write access.  My test showed I was mistaken and your results are
apparently the same.  From this limited experiment I conclude that the
default behavior for creating temporary files, as defined in the Ada
LRM, must be used with caution, i.e., trap exceptions for the case where
write access is denied by the operating system. In terms of GNAT, I'm
guessing their implementation decision was due to either portability
issues or simplification based on standard C I/O functions.  To their
credit, I haven't found any violation of the rules defined in the Ada
LRM.

For your particular development effort on Windows NT, may I suggest you
look in the Win32 library of services.  Specifically, look in the
package Win32.Winbase and you will find two functions that are designed
for creating temporary files.  These functions are named GetTempPath and
GetTempFileName.  The GetTempPath function provides the path defined by
the environment variable TMP, or TEMP is TMP does not exist, or the
current directory if neither TMP or TEMP exist.  This path can then be
passed to GetTempFileName which will provide you with a unique file name
within the given directory.  Refer to the documentation for
GetTempFileName for additional options.  This should work for all users
under NT unless the Administrator has denied them access to the system
TEMP directory; an unusual condition in which case your program may have
to abort or ask the user to provide a location where a temporary file
may be created.  Please note that the downside of this method is that
the file is NOT deleted once the program terminates.  You must remember
to delete it yourself.

Hope this helps!

Regards,

Jeffrey D. Cherry

Re: Temporary files in Ada95.

[Matthew Heaney]

"Jeffrey D. Cherry" <jdcherry@utech.net> writes:

> I don't think you want to ask "Is there a way to make ada not write
> temporary files to the root directory?" since it's not really a language
> issue.  I view it as an implementation issue.

The Form parameter is usually used for this kind of thing.


> I would have thought that GNAT would create the temporary file in a
> directory where the user has write access.

Is there a Form parameter for GNAT, to specify where the temporary file
gets declared?

Re: Temporary files in Ada95.

[Jeffrey D. Cherry]

>> I don't think you want to ask "Is there a way to make ada not write
>> temporary files to the root directory?" since it's not really a language
>> issue.  I view it as an implementation issue.

> The Form parameter is usually used for this kind of thing.

Although, if memory serves, this is correct for VMS, in general I would
have to say that I haven't found that to be true.  I've found the Form
parameter to be used for specialized file access, security requests, and
the like.  Not to force the file to be temporary.  If you could, please
provide some examples of compiler/OS pairings that create temporary
files using the Form parameter.  I would appreciate it since that's why
I read CLA ... to learn new things.  Thanks in advance. :)

>> I would have thought that GNAT would create the temporary file in a
>> directory where the user has write access.

> Is there a Form parameter for GNAT, to specify where the temporary file
> gets declared?

I'd have to check the GNAT documentation and the GCC documentation but I
believe the only Form parameter accepted under any of the Windows
systems is "shared=yes" or "shared=no", or, of course, a null string.

Re: Temporary files in Ada95.

[John J Cupak Jr]

[-- Attachment #1: Type: text/plain, Size: 1929 bytes --]

Jeff and group...

Here's the Ada code I used to illustrate printing a file and deleting it after
close on the DEC system:

with Text_IO;
procedure Send_To_Printer is
  Printer : Text_IO.File_Type;
begin
   Text_IO.Create(File => Printer,
                            Mode => Text_IO.Out_File,
                            Name => "Print.dat",
                            Form => "FILE;" &
                                             "PRINT_ON_CLOSE YES;" &
                                             "DELETE_ON_CLOSE YES;");
   Text_IO.Put_Line(Printer, "And away-y-y-y we go-o-o-o-!");
   Text_IO.Close(Printer);
end Send_To_Printer;

Hope this helps!

Yours in Ada!
John J Cupak, Jr., CCP

Jeffrey D. Cherry wrote:

> >> I don't think you want to ask "Is there a way to make ada not write
> >> temporary files to the root directory?" since it's not really a language
> >> issue.  I view it as an implementation issue.
>
> > The Form parameter is usually used for this kind of thing.
>
> Although, if memory serves, this is correct for VMS, in general I would
> have to say that I haven't found that to be true.  I've found the Form
> parameter to be used for specialized file access, security requests, and
> the like.  Not to force the file to be temporary.  If you could, please
> provide some examples of compiler/OS pairings that create temporary
> files using the Form parameter.  I would appreciate it since that's why
> I read CLA ... to learn new things.  Thanks in advance. :)
>
> >> I would have thought that GNAT would create the temporary file in a
> >> directory where the user has write access.
>
> > Is there a Form parameter for GNAT, to specify where the temporary file
> > gets declared?
>
> I'd have to check the GNAT documentation and the GCC documentation but I
> believe the only Form parameter accepted under any of the Windows
> systems is "shared=yes" or "shared=no", or, of course, a null string.



[-- Attachment #2: Card for John J Cupak Jr --]
[-- Type: text/x-vcard, Size: 428 bytes --]

begin:          vcard
fn:             John J Cupak Jr
n:              Cupak Jr;John J
org:            Raytheon Systems Company
adr:            50 Apple Hill Road;;T3MN35;Tewksbury;MA;01876;USA
email;internet: John_J_Cupak@res.raytheon.com
title:          Software Engineering Instructor
tel;work:       978.858.1222
tel;fax:        978.858.4336
x-mozilla-cpt:  ;0
x-mozilla-html: TRUE
version:        2.1
end:            vcard

Re: Temporary files in Ada95.

[Jerry van Dijk]

Boulais, Jeffrey M (jboulais@sikorsky.com) wrote:

: I'd like to say "Thanks to all", you were quite helpful.  After closer
: inspection, it appears that I was not
: reading section A.8.2 closely enough because I thought it delt with
: Sequential IO.
: 
: However, I did compile the test program supplied by Jeffrey D. Cherry, and I
: still get a name error.  This leads
: me to believe that this is either a Gnat problem, or a problem with my
: install of Gnat.  Thanks again...

One of the wonderful advantages of GNAT is that you have the sourcecode...

In this case, if you check you adainclude directory, you can see what happens
if you do not provide a name, by following:

ada.text_io -> system.file_io -> interfaces.c_streams

to find that GNAT will use the MS C runtime library function 'tmpnam' to
create a temporary filename.

Looking up the documentation for 'tmpnam' (it's online if you do not have
it locally), you will find that the directory the file is created in is the
directory pointed to by the TMP environment variable.

So, either make sure you have read/write access to this directory, or have
the TMP variable point to somewhere you do have these rights.

(technical note: there are other C runtime library functions that do not
exhibit this behaviour, but those are less portable (between Win32 versions
and other OS's), so I think this is a reasonable choice).

--
-- Jerry van Dijk | Leiden, Holland
-- Team Ada       | jdijk@acm.org
-- see http://stad.dsl.nl/~jvandyk