Hi, William.
Thaks for your response. I appreciated it.
I'm triing to reproduce the problem with a more basic program, and some
constants autogenerated, but we obtain another diferent error or
limitation on the tests.
It seems that there is a limit when you define more than 64K constants
with .h files:
Up to 64K constants a pike will compile correctly (32 files .h with
2001 constants each one) .
After include the 33rth file, seems that exceeds the current capacity:
'too many identifiers'.
datos_33.h:1549:Too many identifier_references.
datos_33.h:1550:Too many identifiers.
Violación de segmento (`core' generado)
I've done the test in two systems:
Rocky with 4GB RAM Pike v8.0 release 1116
Mint19 with 15GB RAM Pike v8.0 release 498
If I compare this situation whith the other one, that results in the
malloc error (corrupted size vs. prev_size) is not really comprensible,
because the program that fails
does not exceed these limits, but fails when I add a new .h.
** malloc error (corrupted size vs. prev_size)
number of constants in .h files: 1545
size of .h files .........................: 680KB
** max constants error (Too many identifiers.)
number of constants in .h files....: Fine up to 64.000 (up to 33.h)
total size for the 40 .h files..........: 298MB
As you can see I'm getting the malloc error with only 1500 constants
defined (<64K) and no more thant 1MB of data payload.
Below you have the content of 3 files to autogenerate the .h payload and
test these situations:
First run with includes commented to generate a payload.
Then run with the include.h enabled to see the result.
$ cat include_0.h
constant SALTO="\n\n";
$ cat includes.h
#ifndef INCLUDES_H
#define INCLUDES_H
#include "./datos_1.h"
#include "./datos_2.h"
#include "./datos_3.h"
#include "./datos_4.h"
#include "./datos_5.h"
#include "./datos_6.h"
#include "./datos_7.h"
#include "./datos_8.h"
#include "./datos_9.h"
#include "./datos_10.h"
#include "./datos_11.h"
#include "./datos_12.h"
#include "./datos_13.h"
#include "./datos_14.h"
#include "./datos_15.h"
#include "./datos_16.h"
#include "./datos_17.h"
#include "./datos_18.h"
#include "./datos_19.h"
#include "./datos_20.h"
#include "./datos_21.h"
#include "./datos_22.h"
#include "./datos_23.h"
#include "./datos_24.h"
#include "./datos_25.h"
#include "./datos_26.h"
#include "./datos_27.h"
#include "./datos_28.h"
#include "./datos_29.h"
#include "./datos_30.h"
#include "./datos_31.h"
#include "./datos_32.h"
#include "./datos_33.h"
#include "./datos_34.h"
#include "./datos_35.h"
#include "./datos_36.h"
#include "./datos_37.h"
#include "./datos_38.h"
#include "./datos_39.h"
#include "./datos_40.h"
#endif
$ cat includes.pike
#!/usr/bin/pike
#include "./include_0.h"
//Once autogenerated the payload, uncomment next line to include the of
h files: datos 1 a 40
// #include "./includes.h"
int main()
{
string
PAYLOAD="ABBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB";
PAYLOAD=PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+
PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+
PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD+PAYLOAD;
for (int i=1;i<=40; i++)
{
string f="./datos_"+i+".h";
if (Stdio.is_file(f))
continue;
else
Stdio.stdout.write("#include \"./datos_"+i+".h\"\n");
string AUX=""+
"#ifdef DATOS_"+i+"_H\n"+
"#warning Ya se ha hecho include previamente del fichero de constantes
tsa_mail_ES.h\n"+
"\n"+
"#else\n"+
"#define DATOS_"+i+"_H\n"+
"\n";
Stdio.append_file(f, AUX);
for (int j=0; j<2000; j++)
{
Stdio.append_file(f, "constant
CT_"+i+"_"+j+"=\""+PAYLOAD+"\"\+SALTO;\n");
}
Stdio.append_file(f, "\n#endif\n");
}
Stdio.stdout.write("ALL DONE!\n");
thread_create(nothing);
return -1;
}
void nothing()
{
while(1)
{
sleep(1);
}
}
On the test callbacks are enabled only to reproduce another running
condition.
If anyone has more ideas to try, comments will be appreciated.
Best Regards
David Estopà
El 28/4/22 a las 2:56, H William Welliver escribió:
Just a quick update: I wrote a (basically empty) program that included 20 empty
files and had no problems on 8.0.1116. So there must be something more to it.
If you can find a minimal test case, we might have better luck.
Bill
On Apr 26, 2022, at 11:01 AM, David Estopa <dest...@lleida.net> wrote:
Hello, Friends of Pike!
On Pike 8.0 1116, have anybody experienced similar malloc error trying to
include several .h files into a program?
malloc(): invalid next size (unsorted)
corrupted size vs. prev_size
My program includes several .h to define several languages constants. If I
include 18 works fine, but cannot include one more.
Is there a max size limit to include files into a pike program?
Thanks in advance.
David Estopà
