On Tue, 2013-04-30 at 07:56 -0700, Jon Taylor wrote:
> Not an exact copy, but:
>
>
> variable buf:line;
> variable i:integer;
>
>
> begin
>
>
> file_open(fstatus, readfile, temp_string, read_mode);
> if (fstatus = OPEN_OK) then
> while not (endfile(readfile)) loop
> readline(readfile, buf);
> read(buf, i);
> [...]
That isn't a testcase. This is a testcase : it should just compile and
run, and report what it's doing.
-------------------------------------------------------------
entity Jon is
end Jon;
use std.textio.all;
architecture Taylor of Jon is
begin
process is
variable buf:line;
variable i:integer;
-- variable s : string(1 to 12);
variable s : string(1 to 1);
variable fstatus : file_open_status;
file readfile : text;
constant temp_string : string := "hello.txt";
begin
file_open(fstatus, readfile, temp_string, read_mode);
report "open " & file_open_status'image(fstatus) severity note;
if fstatus = OPEN_OK then
while not endfile(readfile) loop
readline(readfile, buf);
-- read(buf, i);
-- report "i= "& integer'image(i) severity note;
read(buf, s);
report "s= "& s severity note;
end loop;
end if;
report "done" severity note;
wait;
end process;
end Taylor;
-------------------------------------------------------------
and test data "hello.txt"
-------------------------------------------------------------
1
2
hello
3
-------------------------------------------------------------
And I can't see the problems you describe : reading integers would fail
at line 3 of course.
> ../../../src/std/textio_body.v93:1329:5:@0ms:(assertion failure):
> string read failure
I suspect these string read failures occur where the line does not
contain enough data to fill the string; that's the only way I have found
to reproduce them. But that is expected behaviour, not a (GHDL) error.
> I don't know if the two errors are related, but they occur at ther
> same line of code.
I still haven't seen the first "error" so I don't know what it is. If
you can modify the testcase to produce it - OR the one you reported as
https://gna.org/bugs/?20771, please post that, we may be able to find
answers to these.
But we can be reasonably sure that conflicts with libc aren't involved.
- Brian
>
> ...which compiles, so I assume that the read() of an integer is
> properly overloaded with a function that can handle integers. No EOF
> condition and at least 4 bytes of data left, as far as I can see.
> Also, a new wierd error when reading from files: I switched the file
> contents to textual representation of numbers instead of binary
> (replacing the read(buf, [intvar]) with read(buf, [stringvar]) instead
> (which also compiles fine), and at runtime I see this:
>
>
> ../../../src/std/textio_body.v93:1329:5:@0ms:(assertion failure):
> string read failure
>
>
>
> I don't know if the two errors are related, but they occur at ther
> same line of code.
>
>
> Jon
>
>
>
>
> On Mon, Apr 29, 2013 at 1:45 PM, Brian Drummond
> <[email protected]> wrote:
> On Mon, 29 Apr 2013 10:12:03 -0700, you wrote:
>
> >I have a NULL pointer crash occuring on a read() call:
> >
> >read(linevar, intvar);
> >
> >There should be no (normal) way this could compile and then
> not run
> >correctly. Any ideas?
> >
> >Jon
>
>
> Again, as with the other bug you reported; any chance of a
> small testcase?
>
> Thanks,
> - Brian
>
> _______________________________________________
> Ghdl-discuss mailing list
> [email protected]
> https://mail.gna.org/listinfo/ghdl-discuss
>
>
>
_______________________________________________
Ghdl-discuss mailing list
[email protected]
https://mail.gna.org/listinfo/ghdl-discuss
