From astan.chee at al.com.au  Fri Jun  5 09:16:17 2009
From: astan.chee at al.com.au (Astan Chee)
Date: Fri, 5 Jun 2009 17:16:17 +1000
Subject: listing torrent files in btpd
Message-ID: <4A28C641.2000905@al.com.au>

Hi,
I have btpd running with some torrents. Unfortunately they are writing 
to the same folder but each have different files. Is there a way to list 
what files are associated with which torrent in btpd?
Does this make sense?
Cheers
Astan

From axelgenus at gmail.com  Fri Jun  5 09:42:45 2009
From: axelgenus at gmail.com (=?UTF-8?Q?Alessandro_Calor=C3=AC?=)
Date: Fri, 5 Jun 2009 09:42:45 +0200
Subject: listing torrent files in btpd
In-Reply-To: <4A28C641.2000905@al.com.au>
References: <4A28C641.2000905@al.com.au>
Message-ID: <eeb7df810906050042h71b9094eu62d02d43672c9509@mail.gmail.com>

2009/6/5 Astan Chee <astan.chee at al.com.au>:
> Hi,
> I have btpd running with some torrents. Unfortunately they are writing
> to the same folder but each have different files.
You should specify a different directory for each torrent when you add
them with btcli:

Example: btcli add filename.torrent -d directory/to/use

> Is there a way to list what files are associated with which torrent in btpd?

Sorry, there's not any (with btpd).

> Cheers
> Astan

Bye,
Alessandro.

From nemosoft at smcc.demon.nl  Mon Jun 15 22:45:25 2009
From: nemosoft at smcc.demon.nl (Nemosoft Unv.)
Date: Mon, 15 Jun 2009 22:45:25 +0200
Subject: Potiential resource leak in save_info()
Message-ID: <200906152245.25833.nemosoft@smcc.demon.nl>

Hello,

While browsing through the btpd code, I noticed the following code in tlib.c, 
function save_info():

    if ((fflush(fp) == EOF || fsync(fileno(fp)) != 0
            || ferror(fp) || fclose(fp) != 0))
        btpd_err("failed to write '%s'.\n", wpath);

The nature of the C '||' operator is thus, that is evaluated left-to-right, 
and any statements after the first one that evaluates to 'true' are not 
executed. So, if fflush(), fsync() or ferror() return an error, the file 'fp' 
is never closed properly. Suppose there is a temporary condition (like disk 
full), btpd would slowly run out of file descriptors... I thinks it's better 
to call each function, store the result and then compare the return values:

 int e1 = fflush (fp);
 int e2 = fsync (fp);
 int e3 = ferror (fp);
 int e4 = fclose (fp);
 if (e1 == EOF || e2 != 0 ...)

Something to consider for the next release...

 - Nemosoft

From graue at oceanbase.org  Mon Jun 15 22:51:18 2009
From: graue at oceanbase.org (Scott Feeney)
Date: Mon, 15 Jun 2009 16:51:18 -0400
Subject: Potiential resource leak in save_info()
In-Reply-To: <200906152245.25833.nemosoft@smcc.demon.nl>
Message-ID: <o0GeyZkG.1245099078.8910380.graue@pendrell.textdrive.com>

Nemosoft,

Your concerns are moot because btpd_err() quits the program.

Scott



On 6/15/2009, "Nemosoft Unv." <nemosoft at smcc.demon.nl> wrote:

>Hello,
>
>While browsing through the btpd code, I noticed the following code in tlib.c,
>function save_info():
>
>    if ((fflush(fp) == EOF || fsync(fileno(fp)) != 0
>            || ferror(fp) || fclose(fp) != 0))
>        btpd_err("failed to write '%s'.\n", wpath);
>
>The nature of the C '||' operator is thus, that is evaluated left-to-right,
>and any statements after the first one that evaluates to 'true' are not
>executed. So, if fflush(), fsync() or ferror() return an error, the file 'fp'
>is never closed properly. Suppose there is a temporary condition (like disk
>full), btpd would slowly run out of file descriptors... I thinks it's better
>to call each function, store the result and then compare the return values:
>
> int e1 = fflush (fp);
> int e2 = fsync (fp);
> int e3 = ferror (fp);
> int e4 = fclose (fp);
> if (e1 == EOF || e2 != 0 ...)
>
>Something to consider for the next release...
>
> - Nemosoft
>_______________________________________________
>btpd-users mailing list
>btpd-users at murmeldjur.se
>http://lists.stargirl.org/listinfo/btpd-users

From nemosoft at smcc.demon.nl  Mon Jun 15 23:33:09 2009
From: nemosoft at smcc.demon.nl (Nemosoft Unv.)
Date: Mon, 15 Jun 2009 23:33:09 +0200
Subject: Potiential resource leak in save_info()
In-Reply-To: <o0GeyZkG.1245099078.8910380.graue@pendrell.textdrive.com>
References: <o0GeyZkG.1245099078.8910380.graue@pendrell.textdrive.com>
Message-ID: <200906152333.10047.nemosoft@smcc.demon.nl>

On Monday 15 June 2009 22:51:18 Scott Feeney wrote:
> Nemosoft,
>
> Your concerns are moot because btpd_err() quits the program.

Erhm, right. Methinks that function should be renamed to btpd_Fatal() then, 
since btpd_err() suggest it just logs an error, not quit. And maybe I code a 
bit too much on the defensive side, but I don't think it's really nice to 
depend on exit() to close your file descriptors. Oh well...

 - Nemosoft