Re: dbmail 2.0 RC2 pipe.c

Hi,

Thanks for this fix!

I've made a small additional change. Because the while-loop now also 
continues if linemem != 0, I've added a statement to set linemem = 0 if 
the end of the message (.\r\n as one line) is reached.

Ilja

all changes to pipe.c:
ilja@test01:~/dbmail2_0$ cvs diff pipe.c
Index: pipe.c
===================================================================
RCS file: /cvsroot-dbmail/dbmail/pipe.c,v
retrieving revision 1.99.2.4
diff -r1.99.2.4 pipe.c
582c582
<   while (!feof(instream) && !myeof)
---
>   while ((!feof(instream) && (!myeof)) || (linemem != 0))
587a588
>         usedmem += linemem;
596c597
<       while (!feof(instream) && usedmem < READ_BLOCK_SIZE)
---
>       while (!feof(instream) && (usedmem + linemem < READ_BLOCK_SIZE))
614a616,618
>             /* make sure that the function stops by setting linemem 
to zero
>              */
>             linemem = 0;
628a633
>


Chingson Chen wrote:

> I found some problem at pipe.c that cannot read "stdin" larger than 512k ..
> fixed is the diff below ... 
> **************************************************************************************************8
> 582c582
> <   while (!feof(instream) && !myeof)
> ---
>
> >  while ((!feof(instream) && (!myeof)) || (linemem!=0))
>
> 587c587
> <           strncpy(strblock, tmpline, linemem);
> ---
>
> >          strncpy(strblock, tmpline, linemem); usedmem+=linemem;
>
> 596c596
> <       while (!feof(instream) && usedmem < READ_BLOCK_SIZE)
> ---
>
> >      while (!feof(instream) && (usedmem+linemem < READ_BLOCK_SIZE))
>
> 628a629,631
>
> >               else
> >                       trace(TRACE_ERROR,"test large block.");
>
> ¹šŠYšŠX§‚X¬´6æj)]nf¢–Šà†Ûi³ÿæj)fjwÚ²Ülž_æj)fjåŠËb?ú?u¹šil==