From lists at d-ra.ath.cx  Sun Feb  1 10:21:19 2009
From: lists at d-ra.ath.cx (d-ra)
Date: Sun, 1 Feb 2009 10:21:19 +0100
Subject: new snapshot
In-Reply-To: <48425.85.24.201.138.1233324280.squirrel@webmail.stargirl.org>
References: <35276.85.24.200.117.1233246148.squirrel@webmail.stargirl.org>
	<48425.85.24.201.138.1233324280.squirrel@webmail.stargirl.org>
Message-ID: <20090201092119.GA9094@workstation.local>

Hi Richard,

On Fri, Jan 30, 2009 at 03:04:40PM +0100, Richard Nyberg wrote:
> And another one, which should get rid of some tracker problems.
> http://www.murmeldjur.se/btpd/btpd-20090130.tar.gz

Works fine on OpenBSD. Thanks!

        -d-ra

From btpd-users at vuux.net  Sun Feb  1 20:53:59 2009
From: btpd-users at vuux.net (Anton Shcherbinin)
Date: Sun, 1 Feb 2009 22:53:59 +0300
Subject: Cross-compilation for mips platform
Message-ID: <E27BAF67690C4B0FB5DDB0008726D8D8@quux>

I'd like to cross-compile btpd on a usual i386 Ubuntu Linux 8.10 for 
platform `mips' (Networked Media Tank, Popcorn Hour A-110 video player).
I'm quite new to compiling (let alone cross-compiling) because I write 
in Perl and never had to compile anything :)
Installed a Sigma Designs `toolchain' for cross-compilation. I can 
compile `Hello, world!' application and it runs ok on the target 
platform.
Then I downloaded btpd snapshot of Jan 30 and tried to compile it.
At first, I run `configure' without parameters. It failed because it 
tried to compile a dummy application and run it right on Ubuntu, and the 
app was compiled for mips, so running it evidently failed.
Then I tried, without actually understanding what I am doing, to run
./configure --host=mips
This time it did not try to run a.out, so this step was successful, but 
it failed saying

...
checking for SHA1_Final in -lcrypto... no
configure: error: btpd needs openssl's libraries and headers
See `config.log' for more details.

So the question is: how do I make these libraries and headers available 
for the toolchain? I do have necessary libraries and headers in the base 
(Ubuntu) system, I made this sure by running `configure' for i386, and 
the Makefile was successfully created.

Thanks for any clues. Sorry if I missed any important information (which 
I am ready to provide). 


From queueram at gmail.com  Sun Feb  1 21:42:11 2009
From: queueram at gmail.com (Marq Schneider)
Date: Sun, 1 Feb 2009 14:42:11 -0600
Subject: Cross-compilation for mips platform
In-Reply-To: <E27BAF67690C4B0FB5DDB0008726D8D8@quux>
References: <E27BAF67690C4B0FB5DDB0008726D8D8@quux>
Message-ID: <d9b431e0902011242n2fccc903k247cc694ee0e0748@mail.gmail.com>

On Sun, Feb 1, 2009 at 13:53, Anton Shcherbinin <btpd-users at vuux.net> wrote:
> I'd like to cross-compile btpd on a usual i386 Ubuntu Linux 8.10 for
> platform `mips' (Networked Media Tank, Popcorn Hour A-110 video player).
> I'm quite new to compiling (let alone cross-compiling) because I write
> in Perl and never had to compile anything :)
> Installed a Sigma Designs `toolchain' for cross-compilation. I can
> compile `Hello, world!' application and it runs ok on the target
> platform.
> Then I downloaded btpd snapshot of Jan 30 and tried to compile it.
> At first, I run `configure' without parameters. It failed because it
> tried to compile a dummy application and run it right on Ubuntu, and the
> app was compiled for mips, so running it evidently failed.
> Then I tried, without actually understanding what I am doing, to run
> ./configure --host=mips
> This time it did not try to run a.out, so this step was successful, but
> it failed saying
>
> ...
> checking for SHA1_Final in -lcrypto... no
> configure: error: btpd needs openssl's libraries and headers
> See `config.log' for more details.
>
> So the question is: how do I make these libraries and headers available
> for the toolchain? I do have necessary libraries and headers in the base
> (Ubuntu) system, I made this sure by running `configure' for i386, and
> the Makefile was successfully created.
>
> Thanks for any clues. Sorry if I missed any important information (which
> I am ready to provide).
>

You need to have libcrypto built for your target platform (in your
case, MIPS).  libcrypto is available through openssl which you will
need to build using your cross-compiler toolchain.

-Marq

From rnyberg at murmeldjur.se  Sun Feb  8 01:07:15 2009
From: rnyberg at murmeldjur.se (Richard Nyberg)
Date: Sun, 8 Feb 2009 01:07:15 +0100 (CET)
Subject: released 0.14
Message-ID: <58509.85.24.200.109.1234051635.squirrel@webmail.stargirl.org>

Hi users!

I just released 0.14. Lots of bug fixes and improvements since
the last version, including banning peers with bad data.

Hopefully not too many new bugs. ;)

Mmm, sleep...
        -Richard



From axelgenus at gmail.com  Sun Feb  8 09:30:52 2009
From: axelgenus at gmail.com (=?UTF-8?Q?Alessandro_Calor=C3=AC?=)
Date: Sun, 8 Feb 2009 09:30:52 +0100
Subject: released 0.14
In-Reply-To: <58509.85.24.200.109.1234051635.squirrel@webmail.stargirl.org>
References: <58509.85.24.200.109.1234051635.squirrel@webmail.stargirl.org>
Message-ID: <eeb7df810902080030u1f53a2e5yc98f7e5c81385f0f@mail.gmail.com>

2009/2/8, Richard Nyberg <rnyberg at murmeldjur.se>:
> Hi users!
>
> I just released 0.14. Lots of bug fixes and improvements since
> the last version, including banning peers with bad data.

Great news! Thanks... I'm gonna try ASAP.

Bye, Alessandro.

From barone_rosso2 at yahoo.it  Sun Feb  8 09:33:16 2009
From: barone_rosso2 at yahoo.it (barone rosso)
Date: Sun, 8 Feb 2009 09:33:16 +0100
Subject: released 0.14
In-Reply-To: <58509.85.24.200.109.1234051635.squirrel@webmail.stargirl.org>
References: <58509.85.24.200.109.1234051635.squirrel@webmail.stargirl.org>
Message-ID: <ADA5B911-6330-49E6-B8D4-9DE7C6EC5658@yahoo.it>


Il giorno 08/feb/09, alle ore 01:07, Richard Nyberg ha scritto:

> Hi users!
>
> I just released 0.14. Lots of bug fixes and improvements since
> the last version, including banning peers with bad data.
>
> Hopefully not too many new bugs. ;)
>
> Mmm, sleep...
>        -Richard
>


Ok compiling  on MacOsx, btpd is connecting to trackers, but   is  
extremly slow in beginning download and btcli has always zero statistic.
I'm downloading a file just now at 580 Kb/sec and this is btcli stat -in

  2 +.   0.0%   0.00M   0.00kB/s   0.00M   0.00kB/s    0.00    0    
0.0%   0




--
Cordialmente, vostro

barone rosso




From goatspice at gmail.com  Sun Feb  8 10:34:53 2009
From: goatspice at gmail.com (Yassen Roussev)
Date: Sun, 8 Feb 2009 09:34:53 +0000
Subject: released 0.14
In-Reply-To: <58509.85.24.200.109.1234051635.squirrel@webmail.stargirl.org>
References: <58509.85.24.200.109.1234051635.squirrel@webmail.stargirl.org>
Message-ID: <34becb040902080134w64668778r7b8451bc9adbeebd@mail.gmail.com>

Thank you!

Y.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.stargirl.org/pipermail/btpd-users/attachments/20090208/69045406/attachment.htm 

From borg at uu3.net  Sun Feb  8 11:43:46 2009
From: borg at uu3.net (Unknown)
Date: Sun, 8 Feb 2009 11:43:46 +0100 (CET)
Subject: git repo and stuff
In-Reply-To: <52842.85.24.201.38.1231938051.squirrel@webmail.stargirl.org>
References: <40577.85.24.200.233.1231891136.squirrel@webmail.stargirl.org>
	<496D95F6.6060506@computer.org>
	<eeb7df810901140020q4674d43dl9998a29826750b4b@mail.gmail.com>
	<496DA996.7020002@computer.org>
	<52687.85.24.201.38.1231933654.squirrel@webmail.stargirl.org>
	<eeb7df810901140401u5f6699f7k93be5e40ee24e096@mail.gmail.com>
	<52842.85.24.201.38.1231938051.squirrel@webmail.stargirl.org>
Message-ID: <Pine.LNX.4.64.0902081138590.17350@cube>



On Wed, 14 Jan 2009, Richard Nyberg wrote:

>> Also, keep in mind that you should add a compile flag to
>> enable/disable IPv6 support. ;)
>
> Why? I'd rather not, unless there's a very good reason to.

DO IT! You are using C.. use the goodnes of #ifdef .... #endif stuff.
Dont do a mistake and run into troubles when doing your btpd portable.
Not all platforms have IPv6... Same about kqueue, epoll(), select(), 
/dev/poll...

Also.. I dont like idea of IPv4 in IPv6 mapping.. im not sure if someone
is using it. I feel it was like IPv6 design mistake. 90% of APPs I know
are coded dual stack.. so they open 2 listenres.. IPv4 only and IPv6 only
aware (via IPV6_V6ONLY if platform support it).

I know.. more code.. but not so much really if you design it properly.
Wrapped function for most socket calls so you can handle different
event methods and protocols.

Regards,
Borg


From borg at uu3.net  Sun Feb  8 11:47:58 2009
From: borg at uu3.net (Unknown)
Date: Sun, 8 Feb 2009 11:47:58 +0100 (CET)
Subject: git repo and stuff
In-Reply-To: <20090124024924.DABD9478088@ws1-5.us4.outblaze.com>
References: <20090124024924.DABD9478088@ws1-5.us4.outblaze.com>
Message-ID: <Pine.LNX.4.64.0902081145390.17350@cube>



On Fri, 23 Jan 2009, BSD Fanatic wrote:

> As for the git thing, for months I tracked your svn repo..
> I really have no interest in installing the GPL mess that is git,
> would you be willing to mirror git changes to the svn server perodically?
> or perhaps.. release snapshot tarballs?
Hehe.. Same here.. I tried to compile GIT on my accient platform I use.
But failed.. I started to porting but stopped.. too much changes was 
needed to core functions.. (uh oh.. Linus is so L33n0xich bastards).
I stick to svn-1.4.6 for the same reason.. 1.5.x is broken :/
(btw, whats going on w/ coders these days?).
I would be glat to see git to svn replication aswell..
But if not.. at least night tarballs :)

Regards,
Borg


From ruslan.savchenko at gmail.com  Sun Feb  8 11:50:48 2009
From: ruslan.savchenko at gmail.com (Ruslan Savchenko)
Date: Sun, 8 Feb 2009 13:50:48 +0300
Subject: Add multiple torrents at once
Message-ID: <a20a4ca60902080250g4a7b8149yf6425b7ddc8cf710@mail.gmail.com>

Hi all!

Attached trivial patch allows to add multiple torrents at once. Just
like 'btcli add -d somewhere *.torrent'
I'm not sure if I hacked btcli in a right way but I'd like to see this
functionality in mainstream.

Also why topdir is 0 by default? Wouldn't 1 be usual behavior for a
typical torrent client?

-- 
Regards,
savrus
-------------- next part --------------
diff --git a/cli/add.c b/cli/add.c
index 7af08af..32192fd 100644
--- a/cli/add.c
+++ b/cli/add.c
@@ -42,6 +42,7 @@ cmd_add(int argc, char **argv)
     int ch, topdir = 0, start = 1;
     size_t dirlen = 0;
     char *dir = NULL, *name = NULL;
+    int nfile;
 
     while ((ch = getopt_long(argc, argv, "Nd:n:", add_opts, NULL)) != -1) {
         switch (ch) {
@@ -66,7 +67,7 @@ cmd_add(int argc, char **argv)
     argc -= optind;
     argv += optind;
 
-    if (argc != 1 || dir == NULL)
+    if (argc < 1 || dir == NULL)
         usage_add();
 
     btpd_connect();
@@ -76,9 +77,11 @@ cmd_add(int argc, char **argv)
     char dpath[PATH_MAX];
     struct iobuf iob;
 
-    if ((mi = mi_load(argv[0], &mi_size)) == NULL)
-        diemsg("error loading '%s' (%s).\n", argv[0], strerror(errno));
-
+    for (nfile = 0; nfile < argc; nfile++) {
+    if ((mi = mi_load(argv[nfile], &mi_size)) == NULL) {
+        printf("error loading '%s' (%s).\n", argv[nfile], strerror(errno));
+        continue;
+    }
     iob = iobuf_init(PATH_MAX);
     iobuf_write(&iob, dir, dirlen);
     if (topdir && !mi_simple(mi)) {
@@ -89,8 +92,11 @@ cmd_add(int argc, char **argv)
         iobuf_write(&iob, td, tdlen);
     }
     iobuf_swrite(&iob, "\0");
-    if ((errno = make_abs_path(iob.buf, dpath)) != 0)
-        diemsg("make_abs_path '%s' failed (%s).\n", dpath, strerror(errno));
+    if ((errno = make_abs_path(iob.buf, dpath)) != 0) {
+        printf("make_abs_path '%s' failed (%s).\n", dpath, strerror(errno));
+        iobuf_free(&iob);
+        continue;
+    }
     code = btpd_add(ipc, mi, mi_size, dpath, name);
     if (code == 0 && start) {
         struct ipc_torrent tspec;
@@ -99,6 +105,9 @@ cmd_add(int argc, char **argv)
         code = btpd_start(ipc, &tspec);
     }
     if (code != IPC_OK)
-        diemsg("command failed (%s).\n", ipc_strerror(code));
+        printf("command failed for '%s' (%s).\n", argv[nfile], ipc_strerror(code));
+
+    iobuf_free(&iob);
+    }
     return;
 }
-- 
1.4.4.4
-------------- next part --------------
diff --git a/cli/add.c b/cli/add.c
index 32192fd..19b0e88 100644
--- a/cli/add.c
+++ b/cli/add.c
@@ -78,36 +78,36 @@ cmd_add(int argc, char **argv)
     struct iobuf iob;
 
     for (nfile = 0; nfile < argc; nfile++) {
-    if ((mi = mi_load(argv[nfile], &mi_size)) == NULL) {
-        printf("error loading '%s' (%s).\n", argv[nfile], strerror(errno));
-        continue;
-    }
-    iob = iobuf_init(PATH_MAX);
-    iobuf_write(&iob, dir, dirlen);
-    if (topdir && !mi_simple(mi)) {
-        size_t tdlen;
-        const char *td =
-            benc_dget_mem(benc_dget_dct(mi, "info"), "name", &tdlen);
-        iobuf_swrite(&iob, "/");
-        iobuf_write(&iob, td, tdlen);
-    }
-    iobuf_swrite(&iob, "\0");
-    if ((errno = make_abs_path(iob.buf, dpath)) != 0) {
-        printf("make_abs_path '%s' failed (%s).\n", dpath, strerror(errno));
-        iobuf_free(&iob);
-        continue;
-    }
-    code = btpd_add(ipc, mi, mi_size, dpath, name);
-    if (code == 0 && start) {
-        struct ipc_torrent tspec;
-        tspec.by_hash = 1;
-        mi_info_hash(mi, tspec.u.hash);
-        code = btpd_start(ipc, &tspec);
-    }
-    if (code != IPC_OK)
-        printf("command failed for '%s' (%s).\n", argv[nfile], ipc_strerror(code));
+        if ((mi = mi_load(argv[nfile], &mi_size)) == NULL) {
+            printf("error loading '%s' (%s).\n", argv[nfile], strerror(errno));
+            continue;
+        }
+        iob = iobuf_init(PATH_MAX);
+        iobuf_write(&iob, dir, dirlen);
+        if (topdir && !mi_simple(mi)) {
+            size_t tdlen;
+            const char *td =
+                benc_dget_mem(benc_dget_dct(mi, "info"), "name", &tdlen);
+            iobuf_swrite(&iob, "/");
+            iobuf_write(&iob, td, tdlen);
+        }
+        iobuf_swrite(&iob, "\0");
+        if ((errno = make_abs_path(iob.buf, dpath)) != 0) {
+            printf("make_abs_path '%s' failed (%s).\n", dpath, strerror(errno));
+            iobuf_free(&iob);
+            continue;
+        }
+        code = btpd_add(ipc, mi, mi_size, dpath, name);
+        if (code == 0 && start) {
+            struct ipc_torrent tspec;
+            tspec.by_hash = 1;
+            mi_info_hash(mi, tspec.u.hash);
+            code = btpd_start(ipc, &tspec);
+        }
+        if (code != IPC_OK)
+            printf("command failed for '%s' (%s).\n", argv[nfile], ipc_strerror(code));
 
-    iobuf_free(&iob);
+        iobuf_free(&iob);
     }
     return;
 }
-- 
1.4.4.4

From borg at uu3.net  Sun Feb  8 11:53:32 2009
From: borg at uu3.net (Unknown)
Date: Sun, 8 Feb 2009 11:53:32 +0100 (CET)
Subject: rough roadmap
In-Reply-To: <42889.85.24.202.250.1233066929.squirrel@webmail.stargirl.org>
References: <54965.85.24.201.198.1233007007.squirrel@webmail.stargirl.org>
	<F43BD16C-24C7-4CC8-B401-B38D0B8603A3@yahoo.it>
	<d9b431e0901261546p2e1bbf6dh32e9e312b3b223fb@mail.gmail.com>
	<DD6A554B-1717-4794-A6BE-7F49E84BA1F2@yahoo.it>
	<42889.85.24.202.250.1233066929.squirrel@webmail.stargirl.org>
Message-ID: <Pine.LNX.4.64.0902081151110.17350@cube>



On Tue, 27 Jan 2009, Richard Nyberg wrote:

> Ok. I want monotonic time, so stuff using gettimeofday won't do,
> but this looks quite simple to do anyway. However, I'll need ssh
> access to a MacOS machine where I can build btpd.
What for? I saw similar problems in Quagga code for that MONOTIC time.
And they have big problems with timers currently, doing a lot of hacks
for platforms. Whats wrong w/ good old time() ? It works fine from 
decades. If admin cannot keep his clock right. its his problem..
gettimeofday() is good too. but DO you need us precision? I doubt anything
serious can happen in less that 1s for torrent protocol.

Keep it as simple as you can.

Regards,
Borg

From borg at uu3.net  Sun Feb  8 11:56:50 2009
From: borg at uu3.net (Unknown)
Date: Sun, 8 Feb 2009 11:56:50 +0100 (CET)
Subject: Cross-compilation for mips platform
In-Reply-To: <d9b431e0902011242n2fccc903k247cc694ee0e0748@mail.gmail.com>
References: <E27BAF67690C4B0FB5DDB0008726D8D8@quux>
	<d9b431e0902011242n2fccc903k247cc694ee0e0748@mail.gmail.com>
Message-ID: <Pine.LNX.4.64.0902081155010.17350@cube>



On Sun, 1 Feb 2009, Marq Schneider wrote:

> You need to have libcrypto built for your target platform (in your
> case, MIPS).  libcrypto is available through openssl which you will
> need to build using your cross-compiler toolchain.
I belive you want SHA1() function..
On new platforms it can be in libcryto but on old ones.. libssl.
Fix configure script.. or add a hack to check it.

Regards,
Borg


From borg at uu3.net  Sun Feb  8 11:58:30 2009
From: borg at uu3.net (Unknown)
Date: Sun, 8 Feb 2009 11:58:30 +0100 (CET)
Subject: Add multiple torrents at once
In-Reply-To: <a20a4ca60902080250g4a7b8149yf6425b7ddc8cf710@mail.gmail.com>
References: <a20a4ca60902080250g4a7b8149yf6425b7ddc8cf710@mail.gmail.com>
Message-ID: <Pine.LNX.4.64.0902081157290.17350@cube>



On Sun, 8 Feb 2009, Ruslan Savchenko wrote:
> Hi all!
>
> Attached trivial patch allows to add multiple torrents at once. Just
> like 'btcli add -d somewhere *.torrent'
> I'm not sure if I hacked btcli in a right way but I'd like to see this
> functionality in mainstream.
>
> Also why topdir is 0 by default? Wouldn't 1 be usual behavior for a
> typical torrent client?
Hehe.. not bad :) but was it necesaary?

for f in *.torrent do btcli add -d somewhere $f; done
use the power of SHELL.

Regards,
Borg

From ruslan.savchenko at gmail.com  Sun Feb  8 12:02:10 2009
From: ruslan.savchenko at gmail.com (Ruslan Savchenko)
Date: Sun, 8 Feb 2009 14:02:10 +0300
Subject: Add multiple torrents at once
In-Reply-To: <Pine.LNX.4.64.0902081157290.17350@cube>
References: <a20a4ca60902080250g4a7b8149yf6425b7ddc8cf710@mail.gmail.com>
	<Pine.LNX.4.64.0902081157290.17350@cube>
Message-ID: <a20a4ca60902080302g526cf4a6i8df3ebd20b726d5e@mail.gmail.com>

> Hehe.. not bad :) but was it necesaary?
>
> for f in *.torrent do btcli add -d somewhere $f; done
> use the power of SHELL.

Thanks, i'm fed up with shell cycle. Try to add 2-3 torrents several
times a day and you'll understand what i'm saying :)

-- 
Regards,
savrus

From hex29a at openbsd.se  Sun Feb  8 12:08:37 2009
From: hex29a at openbsd.se (Martin)
Date: Sun, 8 Feb 2009 12:08:37 +0100
Subject: Add multiple torrents at once
In-Reply-To: <a20a4ca60902080302g526cf4a6i8df3ebd20b726d5e@mail.gmail.com>
References: <a20a4ca60902080250g4a7b8149yf6425b7ddc8cf710@mail.gmail.com>
	<Pine.LNX.4.64.0902081157290.17350@cube>
	<a20a4ca60902080302g526cf4a6i8df3ebd20b726d5e@mail.gmail.com>
Message-ID: <d3e9906d0902080308h3cac5899n8b48e5e68da9bed2@mail.gmail.com>

Maybe it would be possible to have a "watch" dir like in rtorrent which
checks for newly added torrents? That would definitely make my day :)
 ...sadly enough I'm not that good in programming :(


On Sun, Feb 8, 2009 at 12:02 PM, Ruslan Savchenko <
ruslan.savchenko at gmail.com> wrote:

> > Hehe.. not bad :) but was it necesaary?
> >
> > for f in *.torrent do btcli add -d somewhere $f; done
> > use the power of SHELL.
>
> Thanks, i'm fed up with shell cycle. Try to add 2-3 torrents several
> times a day and you'll understand what i'm saying :)
>
> --
> Regards,
> savrus
> _______________________________________________
> btpd-users mailing list
> btpd-users at murmeldjur.se
> http://lists.stargirl.org/listinfo/btpd-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.stargirl.org/pipermail/btpd-users/attachments/20090208/041ed1b1/attachment.htm 

From lists at d-ra.ath.cx  Sun Feb  8 13:04:23 2009
From: lists at d-ra.ath.cx (d-ra)
Date: Sun, 8 Feb 2009 13:04:23 +0100
Subject: Small tool for buildung .torrent files
Message-ID: <20090208120423.GB31842@workstation.local>

Hi,

a small tool for building .torrent files is:
http://www.rootshell.be/~mitya/ports/mktorrent.tar.gz
For buildung just extract the source and type:

$ make

on the terminal. Then run the executable. The Makefile did not include
a install target. You have to copy it on your own to a appropriate
location. The license is BSD like.

For me, this small program is the perfect buddy to btpd.

        -Dieter

From lists at d-ra.ath.cx  Sun Feb  8 12:55:40 2009
From: lists at d-ra.ath.cx (d-ra)
Date: Sun, 8 Feb 2009 12:55:40 +0100
Subject: released 0.14
In-Reply-To: <58509.85.24.200.109.1234051635.squirrel@webmail.stargirl.org>
References: <58509.85.24.200.109.1234051635.squirrel@webmail.stargirl.org>
Message-ID: <20090208115540.GA31842@workstation.local>

Hi,

On Sun, Feb 08, 2009 at 01:07:15AM +0100, Richard Nyberg wrote:
> I just released 0.14.

Thank you!

Works fine on OpenBSD (but just tested seeding).

Regards
        -Dieter

From rnyberg at murmeldjur.se  Sun Feb  8 15:00:03 2009
From: rnyberg at murmeldjur.se (Richard Nyberg)
Date: Sun, 8 Feb 2009 15:00:03 +0100 (CET)
Subject: released 0.14
In-Reply-To: <ADA5B911-6330-49E6-B8D4-9DE7C6EC5658@yahoo.it>
References: <58509.85.24.200.109.1234051635.squirrel@webmail.stargirl.org>
	<ADA5B911-6330-49E6-B8D4-9DE7C6EC5658@yahoo.it>
Message-ID: <51150.85.24.201.204.1234101603.squirrel@webmail.stargirl.org>

On Sun, February 8, 2009 09:33, barone rosso wrote:
>
> Ok compiling  on MacOsx, btpd is connecting to trackers, but   is
> extremly slow in beginning download and btcli has always zero statistic.
> I'm downloading a file just now at 580 Kb/sec and this is btcli stat -in
>
0.14 is totally broken on MacOs. I'm working on fixing it.

Regards,
        -Richard



From rnyberg at murmeldjur.se  Sun Feb  8 19:43:04 2009
From: rnyberg at murmeldjur.se (Richard Nyberg)
Date: Sun, 8 Feb 2009 19:43:04 +0100 (CET)
Subject: released 0.14
In-Reply-To: <51150.85.24.201.204.1234101603.squirrel@webmail.stargirl.org>
References: <58509.85.24.200.109.1234051635.squirrel@webmail.stargirl.org>
	<ADA5B911-6330-49E6-B8D4-9DE7C6EC5658@yahoo.it>
	<51150.85.24.201.204.1234101603.squirrel@webmail.stargirl.org>
Message-ID: <45183.85.24.201.204.1234118584.squirrel@webmail.stargirl.org>

On Sun, February 8, 2009 15:00, Richard Nyberg wrote:
> On Sun, February 8, 2009 09:33, barone rosso wrote:
>>
>> Ok compiling  on MacOsx, btpd is connecting to trackers, but   is
>> extremly slow in beginning download and btcli has always zero statistic.
>> I'm downloading a file just now at 580 Kb/sec and this is btcli stat -in
>>
> 0.14 is totally broken on MacOs. I'm working on fixing it.
>
Please try the latest snapshot.

Regards,
        -Richard



From barone_rosso2 at yahoo.it  Sun Feb  8 20:33:11 2009
From: barone_rosso2 at yahoo.it (barone rosso)
Date: Sun, 8 Feb 2009 20:33:11 +0100
Subject: released 0.14
In-Reply-To: <45183.85.24.201.204.1234118584.squirrel@webmail.stargirl.org>
References: <58509.85.24.200.109.1234051635.squirrel@webmail.stargirl.org>
	<ADA5B911-6330-49E6-B8D4-9DE7C6EC5658@yahoo.it>
	<51150.85.24.201.204.1234101603.squirrel@webmail.stargirl.org>
	<45183.85.24.201.204.1234118584.squirrel@webmail.stargirl.org>
Message-ID: <4F69C6AB-6BF0-4902-AC95-8CD2CD4E82D1@yahoo.it>


Il giorno 08/feb/09, alle ore 19:43, Richard Nyberg ha scritto:

> On Sun, February 8, 2009 15:00, Richard Nyberg wrote:
>> On Sun, February 8, 2009 09:33, barone rosso wrote:
>>>
>>> Ok compiling  on MacOsx, btpd is connecting to trackers, but   is
>>> extremly slow in beginning download and btcli has always zero  
>>> statistic.
>>> I'm downloading a file just now at 580 Kb/sec and this is btcli  
>>> stat -in
>>>
>> 0.14 is totally broken on MacOs. I'm working on fixing it.
>>
> Please try the latest snapshot.
>
>

it seems ok  now. I try it

--
Cordialmente, vostro

barone rosso




From bsdfan666 at cyberdude.com  Sun Feb  8 22:14:04 2009
From: bsdfan666 at cyberdude.com (BSD Fanatic)
Date: Sun, 8 Feb 2009 16:14:04 -0500
Subject: released 0.14
Message-ID: <20090208211404.A0F30326774@ws1-8.us4.outblaze.com>

Richard wrote:
> Hi users!
> 
> I just released 0.14. Lots of bug fixes and improvements since
> the last version, including banning peers with bad data.
> 
> Hopefully not too many new bugs. ;)
> 
> Mmm, sleep...
>          -Richard

Thanks man!

I compiled the snapshot from today instead of the release, the slow peer 
find bug appears to be gone now! yay! it instantly jumped above 20 peers, 
1MB/s download speeds.:-)

Just some random feature suggestions:

* Selective file downloading, like the README before an ISO image.
* Detailed peer information, IP's.. country.. seed vs peer.. selective 
blocking of peers.. etc.

Bugs:
* Torrent counting bug, if you finish a torrent.. del it.. add a new one.. 
it still becomes #1 instead of #0. (You need to kill btpd.)

You still rock! love the new website.. your main index page is still ugly 
though.

TC!

-- 
Be Yourself @ mail.com!
Choose From 200+ Email Addresses
Get a Free Account at www.mail.com


From bsdfan666 at cyberdude.com  Sun Feb  8 22:21:10 2009
From: bsdfan666 at cyberdude.com (BSD Fanatic)
Date: Sun, 8 Feb 2009 16:21:10 -0500
Subject: Small tool for buildung .torrent files
Message-ID: <20090208212110.D68A41CE8FB@ws1-6.us4.outblaze.com>

Dieter wrote:
> Hi,
> 
> a small tool for building .torrent files is:
> http://www.rootshell.be/~mitya/ports/mktorrent.tar.gz
> For buildung just extract the source and type:
> 
> $ make
> 
> on the terminal. Then run the executable. The Makefile did not include
> a install target. You have to copy it on your own to a appropriate
> location. The license is BSD like.
> 
> For me, this small program is the perfect buddy to btpd.
> 
>          -Dieter

Hi,

I didn't try this yet, but thanks.. definitely need more
simple torrent creators.

A member of this list also created a similar tool called
mktorrent.

It's on this page: http://borg.uu3.net/~borg/

Perhaps that can be integrated? presuming the author is in
agreement.

TC!

-- 
Be Yourself @ mail.com!
Choose From 200+ Email Addresses
Get a Free Account at www.mail.com


From graue at oceanbase.org  Sun Feb  8 22:30:26 2009
From: graue at oceanbase.org (Scott Feeney)
Date: Sun, 08 Feb 2009 16:30:26 -0500
Subject: Small tool for buildung .torrent files
In-Reply-To: <20090208212110.D68A41CE8FB@ws1-6.us4.outblaze.com>
Message-ID: <fTGUYRlp.1234128626.7017890.graue@pendrell.textdrive.com>

On 2/8/2009, "BSD Fanatic" <bsdfan666 at cyberdude.com> wrote:

>Hi,
>
>I didn't try this yet, but thanks.. definitely need more
>simple torrent creators.

We need MORE? Really? :)

Obligatory "me too": I also wrote my own torrent creator,
"Torrentize": http://oceanbase.org/graue/wip.html

It is a "work in progress" because it has certain limitations, which
apparently don't affect me because I've been using it for years
without being motivated to fix them. :)

Scott

From lists at d-ra.ath.cx  Sun Feb  8 22:44:24 2009
From: lists at d-ra.ath.cx (d-ra)
Date: Sun, 8 Feb 2009 22:44:24 +0100
Subject: Small tool for buildung .torrent files
In-Reply-To: <20090208212110.D68A41CE8FB@ws1-6.us4.outblaze.com>
References: <20090208212110.D68A41CE8FB@ws1-6.us4.outblaze.com>
Message-ID: <20090208214424.GA9621@workstation.local>

Hi,

On Sun, Feb 08, 2009 at 04:21:10PM -0500, BSD Fanatic wrote:
> A member of this list also created a similar tool called
> mktorrent.
> 
> It's on this page: http://borg.uu3.net/~borg/

You are right. This is the site. I accidently posted the OpenBSD port.
My fault.

Thanks
        -d-ra

From btpd-users at vuux.net  Mon Feb  9 04:21:37 2009
From: btpd-users at vuux.net (Anton Shcherbinin)
Date: Mon, 9 Feb 2009 06:21:37 +0300
Subject: Failed to initialize evloop (Function not implemented).
Message-ID: <C5AEFD012ED94658AE5FD29FC7344171@quux>

First of all, Richard, thank you very much for you work in general and 
release 0.14 in particular :)

Finally I was able to compile BTPD for mipsel-linux-uclibc platform on a 
usual i386 Linux Ubuntu machine. I copied libcrypto.so from the target 
platform to the Ubuntu box. After that BTPD 0.14 was built successfully.

But when I copied the executables to the target machine and run btpd, it 
exited with code 1 saying:

Failed to initialize evloop (Function not implemented).

What is it that I am missing? Thanks for any help.

btcli and btinfo run successfully (though of course I could not check 
`btcli add` and other commands involving btpd daemon).

-- 
Anton Shcherbinin 


From bsdfan666 at cyberdude.com  Mon Feb  9 04:43:40 2009
From: bsdfan666 at cyberdude.com (BSD Fanatic)
Date: Sun, 8 Feb 2009 22:43:40 -0500
Subject: Failed to initialize evloop (Function not implemented).
Message-ID: <20090209034340.2FE8C11581F@ws1-7.us4.outblaze.com>

Anton Shcherbinin wrote:
> 
> First of all, Richard, thank you very much for you work in general and
> release 0.14 in particular :)
> 
> Finally I was able to compile BTPD for mipsel-linux-uclibc platform on a
> usual i386 Linux Ubuntu machine. I copied libcrypto.so from the target
> platform to the Ubuntu box. After that BTPD 0.14 was built successfully.
> 
> But when I copied the executables to the target machine and run btpd, it
> exited with code 1 saying:
> 
> Failed to initialize evloop (Function not implemented).
> 
> What is it that I am missing? Thanks for any help.
> 
> btcli and btinfo run successfully (though of course I could not check
> `btcli add` and other commands involving btpd daemon).
> 
> --
> Anton Shcherbinin

Hi,

That string isn't in the executable, it's in your C library.. clearly
btpd is using a function that's only stubbed by uClibc.

You'll have to trace the execution and figure out where it bails out.

If you determine the missing function, you can either see if the latest
version provides it, or write it yourself.

TC!

-- 
Be Yourself @ mail.com!
Choose From 200+ Email Addresses
Get a Free Account at www.mail.com


From rnyberg at murmeldjur.se  Mon Feb  9 08:52:00 2009
From: rnyberg at murmeldjur.se (Richard Nyberg)
Date: Mon, 9 Feb 2009 08:52:00 +0100 (CET)
Subject: Failed to initialize evloop (Function not implemented).
In-Reply-To: <C5AEFD012ED94658AE5FD29FC7344171@quux>
References: <C5AEFD012ED94658AE5FD29FC7344171@quux>
Message-ID: <41118.85.24.200.146.1234165920.squirrel@webmail.stargirl.org>

On Mon, February 9, 2009 04:21, Anton Shcherbinin wrote:
> First of all, Richard, thank you very much for you work in general and
> release 0.14 in particular :)

You're welcome. :)

> Finally I was able to compile BTPD for mipsel-linux-uclibc platform on a
> usual i386 Linux Ubuntu machine. I copied libcrypto.so from the target
> platform to the Ubuntu box. After that BTPD 0.14 was built successfully.

Very nice.

> But when I copied the executables to the target machine and run btpd, it
> exited with code 1 saying:
>
> Failed to initialize evloop (Function not implemented).
>
> What is it that I am missing? Thanks for any help.
>
I guess your'e missing epoll. Why though? Are you using a very old kernel?
Try configuring with "--with-evloop-method=poll".

Regards,
        -Richard


From rnyberg at murmeldjur.se  Mon Feb  9 09:00:00 2009
From: rnyberg at murmeldjur.se (Richard Nyberg)
Date: Mon, 9 Feb 2009 09:00:00 +0100 (CET)
Subject: Failed to initialize evloop (Function not implemented).
In-Reply-To: <41118.85.24.200.146.1234165920.squirrel@webmail.stargirl.org>
References: <C5AEFD012ED94658AE5FD29FC7344171@quux>
	<41118.85.24.200.146.1234165920.squirrel@webmail.stargirl.org>
Message-ID: <49344.85.24.200.146.1234166400.squirrel@webmail.stargirl.org>

On Mon, February 9, 2009 08:52, Richard Nyberg wrote:
> Try configuring with "--with-evloop-method=poll".
>
Forgot to mention that changing the evloop method means
you have to recompile everything, so "make clean" first
if the directory is dirty.



From barone_rosso2 at yahoo.it  Mon Feb  9 10:45:46 2009
From: barone_rosso2 at yahoo.it (barone rosso)
Date: Mon, 9 Feb 2009 10:45:46 +0100
Subject: 2 request
Message-ID: <BC234E97-31D8-49DC-8B72-CEE5F2C030A1@yahoo.it>

1: btcli start -a     as the stop -a command

2. https support


--
Cordialmente, vostro

barone rosso




From btpd-users at vuux.net  Mon Feb  9 11:56:08 2009
From: btpd-users at vuux.net (Anton Shcherbinin)
Date: Mon, 9 Feb 2009 13:56:08 +0300
Subject: Failed to initialize evloop (Function not implemented).
In-Reply-To: <41118.85.24.200.146.1234165920.squirrel@webmail.stargirl.org>
References: <C5AEFD012ED94658AE5FD29FC7344171@quux>
	<41118.85.24.200.146.1234165920.squirrel@webmail.stargirl.org>
Message-ID: <89C2A51D44E443A589F5AB1549DC75AC@quux>

Mon, 9 Feb 2009 08:52:00 +0100, Richard Nyberg wrote in 
<41118.85.24.200.146.1234165920.squirrel at webmail.stargirl.org> :
> On Mon, February 9, 2009 04:21, Anton Shcherbinin wrote:
>> Finally I was able to compile BTPD for mipsel-linux-uclibc platform 
>> on a
>> usual i386 Linux Ubuntu machine. I copied libcrypto.so from the 
>> target
>> platform to the Ubuntu box. After that BTPD 0.14 was built 
>> successfully.
>
> Very nice.
>
>> But when I copied the executables to the target machine and run btpd, 
>> it
>> exited with code 1 saying:
>>
>> Failed to initialize evloop (Function not implemented).
>>
>> What is it that I am missing? Thanks for any help.
>>
> I guess your'e missing epoll. Why though? Are you using a very old 
> kernel?

I wouldn't say it's very old:
/ # uname -a
Linux nmt 2.6.15-sigma #2265 PREEMPT Wed Oct 29 16:56:47 MYT 2008 mips 
unknown

> Try configuring with "--with-evloop-method=poll".

It worked! Thank you. After configuring `--with-evloop-method=poll' btpd 
runs on my media player. It's seeding my old (from btpd 0.13) torrents. 
Seems ok so far (just a few minutes passed though).

Now I'm wondering:
1) With version 0.13, I used to see only 1 btpd process. Now with 
v.0.14, I've run it in the same way (/mnt/syb8634/bin/btpd -d 
/share/.btpd -p 2222 --bw-in 1111 --bw-out 3333 --ipcprot 0660) but I 
see 3 identical btpd processes:
  PID  Uid        VSZ Stat Command
 5213 root       7136 S   /mnt/syb8634/bin/btpd -d /share/.btpd -p 
2222 --bw-in 1111 --bw-out 3333 --ipcprot 0660
 5214 root       7136 S   /mnt/syb8634/bin/btpd -d /share/.btpd -p 
2222 --bw-in 1111 --bw-out 3333 --ipcprot 0660
 5215 root       7136 S   /mnt/syb8634/bin/btpd -d /share/.btpd -p 
2222 --bw-in 1111 --bw-out 3333 --ipcprot 0660

Why 3 processes?

2) The amount of memory (VSZ column in the output of `ps` above) btpd 
uses has increased considerably (used to be something like 4500 KB, now 
7136 KB). On a media player with just 90 MB of RAM and a whole bunch of 
services (2 different http servers, Apache and proprietary, ftp server, 
nfs server, Samba, telnet server, web browser, video player etc. etc.) 
memory is at a premium.

3) Sizes of binaries. Old btpd 0.13 binary (which I did not compile, got 
it installed with the media player) is even bigger (591934 bytes) than 
the new 0.14 one (495553 bytes). At the same time, new btcli is nearly 2 
times bigger (~140 KB vs. ~70 KB) than the old one. New btinfo is ~ 80 
KB compared to old ~ 50 KB. It's not a concern at all, I have half a 
terabyte free at the player's HDD. Just wondering why btpd is smaller 
and btcli/btinfo are bigger.

-- 
Anton Shcherbinin 


From rnyberg at murmeldjur.se  Mon Feb  9 13:41:10 2009
From: rnyberg at murmeldjur.se (Richard Nyberg)
Date: Mon, 9 Feb 2009 13:41:10 +0100 (CET)
Subject: Failed to initialize evloop (Function not implemented).
In-Reply-To: <89C2A51D44E443A589F5AB1549DC75AC@quux>
References: <C5AEFD012ED94658AE5FD29FC7344171@quux>
	<41118.85.24.200.146.1234165920.squirrel@webmail.stargirl.org>
	<89C2A51D44E443A589F5AB1549DC75AC@quux>
Message-ID: <60549.85.24.200.146.1234183270.squirrel@webmail.stargirl.org>

On Mon, February 9, 2009 11:56, Anton Shcherbinin wrote:
> Mon, 9 Feb 2009 08:52:00 +0100, Richard Nyberg wrote in
> <41118.85.24.200.146.1234165920.squirrel at webmail.stargirl.org> :
>> On Mon, February 9, 2009 04:21, Anton Shcherbinin wrote:

>>> What is it that I am missing? Thanks for any help.
>>>
>> I guess your'e missing epoll. Why though? Are you using a very old
>> kernel?
>
> I wouldn't say it's very old:
So why are you missing epoll? Is there some test I could do?

> Now I'm wondering:
> 1) With version 0.13, I used to see only 1 btpd process. Now with
> v.0.14, I've run it in the same way (/mnt/syb8634/bin/btpd -d
> /share/.btpd -p 2222 --bw-in 1111 --bw-out 3333 --ipcprot 0660) but I
> see 3 identical btpd processes:
>
> Why 3 processes?

Why indeed? Could you run "strace -f btpd [possible opts] --empty-start"
do "btcli kill" after a second or so and send me the strace output?

Did btpd even return control of the terminal to you?

> 2) The amount of memory (VSZ column in the output of `ps` above) btpd
> uses has increased considerably (used to be something like 4500 KB, now
> 7136 KB). On a media player with just 90 MB of RAM and a whole bunch of
> services (2 different http servers, Apache and proprietary, ftp server,
> nfs server, Samba, telnet server, web browser, video player etc. etc.)
> memory is at a premium.

0.14 uses two threads and possibly different shared libraries from 0.13.
It also does more alloc's and free's due to bad peer detection. All of
this may influence the virtual size of a process. I don't think you'll
need to worry.

> 3) Sizes of binaries. Old btpd 0.13 binary (which I did not compile, got
> it installed with the media player) is even bigger (591934 bytes) than
> the new 0.14 one (495553 bytes). At the same time, new btcli is nearly 2
> times bigger (~140 KB vs. ~70 KB) than the old one. New btinfo is ~ 80
> KB compared to old ~ 50 KB. It's not a concern at all, I have half a
> terabyte free at the player's HDD. Just wondering why btpd is smaller
> and btcli/btinfo are bigger.

I hope you still have room for you torrents. ;)
Anyway,
Larger: different compiler flags? "-g" for instance? maybe your 0.13
        binaries are stripped.
Smaller: probably removal of libevent.

Regards,
        -Richard


From rnyberg at murmeldjur.se  Mon Feb  9 13:47:24 2009
From: rnyberg at murmeldjur.se (Richard Nyberg)
Date: Mon, 9 Feb 2009 13:47:24 +0100 (CET)
Subject: Failed to initialize evloop (Function not implemented).
In-Reply-To: <60549.85.24.200.146.1234183270.squirrel@webmail.stargirl.org>
References: <C5AEFD012ED94658AE5FD29FC7344171@quux>
	<41118.85.24.200.146.1234165920.squirrel@webmail.stargirl.org>
	<89C2A51D44E443A589F5AB1549DC75AC@quux>
	<60549.85.24.200.146.1234183270.squirrel@webmail.stargirl.org>
Message-ID: <50619.85.24.200.146.1234183644.squirrel@webmail.stargirl.org>

On Mon, February 9, 2009 13:41, Richard Nyberg wrote:
> On Mon, February 9, 2009 11:56, Anton Shcherbinin wrote:

>> Why 3 processes?
>
Or could it be that ps shows an entry per thread?

        -Richard



From rnyberg at murmeldjur.se  Mon Feb  9 22:21:16 2009
From: rnyberg at murmeldjur.se (Richard Nyberg)
Date: Mon, 9 Feb 2009 22:21:16 +0100 (CET)
Subject: btpd 0.15
Message-ID: <46969.85.24.200.146.1234214476.squirrel@webmail.stargirl.org>

Fixed some bugs in 0.14.

btpd 0.15:
----------

Bug fixes:
- The event loop could busy loop for up to a millisecond sometimes.
- On some platforms btpd couldn't initiate connections to other peers.
- The timer code was broken on MacOS X.

Happy torrenting!
        -Richard



From barone_rosso2 at yahoo.it  Tue Feb 10 10:48:17 2009
From: barone_rosso2 at yahoo.it (barone rosso)
Date: Tue, 10 Feb 2009 10:48:17 +0100
Subject: btpd 0.15
In-Reply-To: <46969.85.24.200.146.1234214476.squirrel@webmail.stargirl.org>
References: <46969.85.24.200.146.1234214476.squirrel@webmail.stargirl.org>
Message-ID: <B66171A1-A71E-44EB-82EA-3C88F62FF345@yahoo.it>


Il giorno 09/feb/09, alle ore 22:21, Richard Nyberg ha scritto:

> Fixed some bugs in 0.14.
>
> btpd 0.15:
> ----------
>
> Bug fixes:
> - The event loop could busy loop for up to a millisecond sometimes.
> - On some platforms btpd couldn't initiate connections to other peers.
> - The timer code was broken on MacOS X.
>
>


little bug,maybe only on OsX:

my log file report: Feb 10 09:49:36 error: bad data from tracker ......

but no E* is reported by btcli
--
Cordialmente, vostro

barone rosso




From rnyberg at murmeldjur.se  Tue Feb 10 12:02:00 2009
From: rnyberg at murmeldjur.se (Richard Nyberg)
Date: Tue, 10 Feb 2009 12:02:00 +0100 (CET)
Subject: btpd 0.15
In-Reply-To: <B66171A1-A71E-44EB-82EA-3C88F62FF345@yahoo.it>
References: <46969.85.24.200.146.1234214476.squirrel@webmail.stargirl.org>
	<B66171A1-A71E-44EB-82EA-3C88F62FF345@yahoo.it>
Message-ID: <58536.85.24.200.151.1234263720.squirrel@webmail.stargirl.org>

On Tue, February 10, 2009 10:48, barone rosso wrote:
>
> Il giorno 09/feb/09, alle ore 22:21, Richard Nyberg ha scritto:
>
>
> little bug,maybe only on OsX:
>
> my log file report: Feb 10 09:49:36 error: bad data from tracker ......
>
> but no E* is reported by btcli

Trackers are just handled differently. E* isn't used anymore. At least
for now.

Regards,
        -Richard



From btpd-users at vuux.net  Wed Feb 11 20:55:19 2009
From: btpd-users at vuux.net (Anton Shcherbinin)
Date: Wed, 11 Feb 2009 22:55:19 +0300
Subject: Failed to initialize evloop (Function not implemented).
In-Reply-To: <60549.85.24.200.146.1234183270.squirrel@webmail.stargirl.org>
References: <C5AEFD012ED94658AE5FD29FC7344171@quux>
	<41118.85.24.200.146.1234165920.squirrel@webmail.stargirl.org>
	<89C2A51D44E443A589F5AB1549DC75AC@quux>
	<60549.85.24.200.146.1234183270.squirrel@webmail.stargirl.org>
Message-ID: <117382232.20090211225519@vuux.net>

Mon, 9 Feb 2009 13:41:10 +0100 (CET), Richard Nyberg wrote:
> On Mon, February 9, 2009 11:56, Anton Shcherbinin wrote:
>> Mon, 9 Feb 2009 08:52:00 +0100, Richard Nyberg wrote in
>> <41118.85.24.200.146.1234165920.squirrel at webmail.stargirl.org> :
>>> On Mon, February 9, 2009 04:21, Anton Shcherbinin wrote:

>>>> What is it that I am missing? Thanks for any help.
>>>>
>>> I guess your'e missing epoll. Why though? Are you using a very old
>>> kernel?
>>
>> I wouldn't say it's very old:
> So why are you missing epoll? Is there some test I could do?

First of all, sorry for a delayed reply.
Unfortunately, I'm not that familiar with Linux. I don't even know
what an `epoll' is ;-) let alone why my mips-based Linux 2.6.15-sigma
lacks epoll. But BTPD seems to work ok configured
--with-evloop-method=poll so that's probably not that important.

>> 1) With version 0.13, I used to see only 1 btpd process. Now with
>> v.0.14, I've run it in the same way (/mnt/syb8634/bin/btpd -d
>> /share/.btpd -p 2222 --bw-in 1111 --bw-out 3333 --ipcprot 0660) but I
>> see 3 identical btpd processes:
>>
>> Why 3 processes?

> Why indeed? Could you run "strace -f btpd [possible opts] --empty-start"
> do "btcli kill" after a second or so and send me the strace output?

I don't have strace on the platform.

> Did btpd even return control of the terminal to you?

Yes. The actual answer could be what you wrote in another letter in
this thread: `ps` just shows threads of 1 process as several
processes. Does btpd actually use 3 threads?

I wonder (it's off topic, it's about Linux in general and not BTPD)
how to tell different threads of the same process from actually
different processes (they look like processes cause they have
different PIDs and separate entries in /proc/$PID

>> 2) The amount of memory (VSZ column in the output of `ps` above) btpd
>> uses has increased considerably
[skip]

> 0.14 uses two threads and possibly different shared libraries from 0.13.
> It also does more alloc's and free's due to bad peer detection. All of
> this may influence the virtual size of a process. I don't think you'll
> need to worry.

Agreed. I won't worry :)

>> 3) Sizes of binaries.

> Larger: different compiler flags? "-g" for instance? maybe your 0.13
>         binaries are stripped.

Sure. Stripped them now and they are much smaller.
Thanks for a helpful answer.

-- 
Anton Shcherbinin


From bsdfan666 at cyberdude.com  Thu Feb 12 18:24:00 2009
From: bsdfan666 at cyberdude.com (BSD Fanatic)
Date: Thu, 12 Feb 2009 12:24:00 -0500
Subject: Failed to initialize evloop (Function not implemented).
Message-ID: <20090212172400.10186BE407B@ws1-9.us4.outblaze.com>

Anton Shcherbinin wrote:
> Richard wrote:
> > So why are you missing epoll? Is there some test I could do?
> >
> First of all, sorry for a delayed reply.
> Unfortunately, I'm not that familiar with Linux. I don't even know
> what an `epoll' is ;-) let alone why my mips-based Linux 2.6.15-sigma
> lacks epoll. But BTPD seems to work ok configured
> --with-evloop-method=poll so that's probably not that important.

Hi,

I don't know much about Linux, but.. in my uneducated guess, I think it's 
related to the fact that he's using uClibc, a lighter alternative to GLIBC.

As indicated in his earlier emails, the epoll functions appear to be stubbed.. 
or incomplete.

The last time I checked, years ago.. that library was intended for uClinux 
on MMU-less system, though, I think I read that uClinux merged with the
mainline kernel.

Anton, do you know enough about this architecture? is it possible to use 
GLIBC instead?

http://www.gnu.org/software/libc/

TC!

-- 
Be Yourself @ mail.com!
Choose From 200+ Email Addresses
Get a Free Account at www.mail.com


From lundstrom.jerry at gmail.com  Fri Feb 13 15:06:50 2009
From: lundstrom.jerry at gmail.com (=?ISO-8859-1?Q?Jerry_Lundstr=F6m?=)
Date: Fri, 13 Feb 2009 15:06:50 +0100
Subject: [PATCH] clock_gettime on system that don't support CLOCK_MONOTONIC
Message-ID: <bc772bd10902130606u458f22ddh36eccda48af56d69@mail.gmail.com>

Hi,

Here is a patch that added evtimer_clock_init() that will try and find
some way of getting time on systems that might not support
CLOCK_MONOTONIC (like my MIPS32 OpenWRT 8.09rc2).

You can change the order it tries them if you find some of them
better. I just did it in the order of the time.h file.

/Jerry
-------------- next part --------------
A non-text attachment was scrubbed...
Name: btpd-0.15-clock_gettime.diff
Type: text/x-patch
Size: 1959 bytes
Desc: not available
Url : http://lists.stargirl.org/pipermail/btpd-users/attachments/20090213/ce77c810/attachment.bin 

From lundstrom.jerry at gmail.com  Fri Feb 13 15:16:30 2009
From: lundstrom.jerry at gmail.com (=?ISO-8859-1?Q?Jerry_Lundstr=F6m?=)
Date: Fri, 13 Feb 2009 15:16:30 +0100
Subject: [PATCH] tracker &numwant= option
Message-ID: <bc772bd10902130616g48854638w127a731ed00590e6@mail.gmail.com>

Hi,

Here is a suggested patch to send the number of wanted peers to the
tracker. This can speed up the torrent greatly.

The protocol specifications says this is default 50 but if you don't
supply it its default to what the tracker likes and most trackers I've
seen only supply 20-25 peers at start. btpd is set to have a tracker
request interval of 25 to 30 minutes so it takes a while for it to
build up the peer list.

/Jerry
-------------- next part --------------
A non-text attachment was scrubbed...
Name: btpd-0.15-numwant.diff
Type: text/x-patch
Size: 2814 bytes
Desc: not available
Url : http://lists.stargirl.org/pipermail/btpd-users/attachments/20090213/624ef8d4/attachment.bin 

From rnyberg at murmeldjur.se  Fri Feb 13 16:03:32 2009
From: rnyberg at murmeldjur.se (Richard Nyberg)
Date: Fri, 13 Feb 2009 16:03:32 +0100 (CET)
Subject: [PATCH] clock_gettime on system that don't support 
	CLOCK_MONOTONIC
In-Reply-To: <bc772bd10902130606u458f22ddh36eccda48af56d69@mail.gmail.com>
References: <bc772bd10902130606u458f22ddh36eccda48af56d69@mail.gmail.com>
Message-ID: <2441.85.24.200.83.1234537412.squirrel@webmail.stargirl.org>

On Fri, February 13, 2009 15:06, Jerry Lundstr?m wrote:
> Hi,
>
> Here is a patch that added evtimer_clock_init() that will try and find
> some way of getting time on systems that might not support
> CLOCK_MONOTONIC (like my MIPS32 OpenWRT 8.09rc2).
>
> You can change the order it tries them if you find some of them
> better. I just did it in the order of the time.h file.
>
Hi!

CLOCK_REALTIME will work in most cases, but will need a little extra
care in case the time runs backwards. The CPUTIME clocks measures the
wrong time and will not work for btpd.

However, I'm going to add a fallback to gettimeofday. It's equivalent
to CLOCK_REALTIME, but exists on more systems.

Regards,
        -Richard



From rnyberg at murmeldjur.se  Fri Feb 13 16:10:39 2009
From: rnyberg at murmeldjur.se (Richard Nyberg)
Date: Fri, 13 Feb 2009 16:10:39 +0100 (CET)
Subject: [PATCH] tracker &numwant= option
In-Reply-To: <bc772bd10902130616g48854638w127a731ed00590e6@mail.gmail.com>
References: <bc772bd10902130616g48854638w127a731ed00590e6@mail.gmail.com>
Message-ID: <2447.85.24.200.83.1234537839.squirrel@webmail.stargirl.org>

On Fri, February 13, 2009 15:16, Jerry Lundstr?m wrote:
> Hi,
>
> Here is a suggested patch to send the number of wanted peers to the
> tracker. This can speed up the torrent greatly.

Sounds like a good idea. :)

> The protocol specifications says this is default 50 but if you don't
> supply it its default to what the tracker likes and most trackers I've
> seen only supply 20-25 peers at start. btpd is set to have a tracker
> request interval of 25 to 30 minutes so it takes a while for it to
> build up the peer list.
>
Actually, btpd will use the interval specified by the tracker. I've
seen anything from 5 to 120 minutes. The 25 to 30 minutes are just
for buggy trackers.

Thanks!
        -Richard



From lundstrom.jerry at gmail.com  Fri Feb 13 16:12:40 2009
From: lundstrom.jerry at gmail.com (=?ISO-8859-1?Q?Jerry_Lundstr=F6m?=)
Date: Fri, 13 Feb 2009 16:12:40 +0100
Subject: [PATCH] clock_gettime on system that don't support 
	CLOCK_MONOTONIC
In-Reply-To: <2441.85.24.200.83.1234537412.squirrel@webmail.stargirl.org>
References: <bc772bd10902130606u458f22ddh36eccda48af56d69@mail.gmail.com>
	<2441.85.24.200.83.1234537412.squirrel@webmail.stargirl.org>
Message-ID: <bc772bd10902130712xf2943b0wab14b28085538c6d@mail.gmail.com>

On Fri, Feb 13, 2009 at 4:03 PM, Richard Nyberg <rnyberg at murmeldjur.se> wrote:
> However, I'm going to add a fallback to gettimeofday. It's equivalent
> to CLOCK_REALTIME, but exists on more systems.

gettimeofday() will only provide microseconds to clock_gettime()'s
nanoseconds, will that precision be enough?

/Jerry

From rnyberg at murmeldjur.se  Fri Feb 13 16:18:16 2009
From: rnyberg at murmeldjur.se (Richard Nyberg)
Date: Fri, 13 Feb 2009 16:18:16 +0100 (CET)
Subject: [PATCH] clock_gettime on system that don't support  
	CLOCK_MONOTONIC
In-Reply-To: <bc772bd10902130712xf2943b0wab14b28085538c6d@mail.gmail.com>
References: <bc772bd10902130606u458f22ddh36eccda48af56d69@mail.gmail.com>
	<2441.85.24.200.83.1234537412.squirrel@webmail.stargirl.org>
	<bc772bd10902130712xf2943b0wab14b28085538c6d@mail.gmail.com>
Message-ID: <2494.85.24.200.83.1234538296.squirrel@webmail.stargirl.org>

On Fri, February 13, 2009 16:12, Jerry Lundstr?m wrote:
> On Fri, Feb 13, 2009 at 4:03 PM, Richard Nyberg <rnyberg at murmeldjur.se>
> wrote:
>> However, I'm going to add a fallback to gettimeofday. It's equivalent
>> to CLOCK_REALTIME, but exists on more systems.
>
> gettimeofday() will only provide microseconds to clock_gettime()'s
> nanoseconds, will that precision be enough?
>
Hehe! I'd say a tenth of a second would be good enough. :)
Btpd really doesn't require any high precision timers.

        -Richard



From lundstrom.jerry at gmail.com  Fri Feb 13 16:23:54 2009
From: lundstrom.jerry at gmail.com (=?ISO-8859-1?Q?Jerry_Lundstr=F6m?=)
Date: Fri, 13 Feb 2009 16:23:54 +0100
Subject: [PATCH] clock_gettime on system that don't support 
	CLOCK_MONOTONIC
In-Reply-To: <2494.85.24.200.83.1234538296.squirrel@webmail.stargirl.org>
References: <bc772bd10902130606u458f22ddh36eccda48af56d69@mail.gmail.com>
	<2441.85.24.200.83.1234537412.squirrel@webmail.stargirl.org>
	<bc772bd10902130712xf2943b0wab14b28085538c6d@mail.gmail.com>
	<2494.85.24.200.83.1234538296.squirrel@webmail.stargirl.org>
Message-ID: <bc772bd10902130723k2bae9dfcpddfbd6dc22eeea35@mail.gmail.com>

On Fri, Feb 13, 2009 at 4:18 PM, Richard Nyberg <rnyberg at murmeldjur.se> wrote:
> Hehe! I'd say a tenth of a second would be good enough. :)
> Btpd really doesn't require any high precision timers.

Nice, I'll change the patch. Have evtimer_clock_init() to check if
CLOCK_MONOTONIC gives EINVAL and then set it to use gettimeofday
instead and try it out some on my MIPS box.

/Jerry

From nemosoft at smcc.demon.nl  Fri Feb 13 23:36:21 2009
From: nemosoft at smcc.demon.nl (Nemosoft Unv.)
Date: Fri, 13 Feb 2009 23:36:21 +0100
Subject: [PATCH] Renaming torrent
Message-ID: <200902132336.21471.nemosoft@smcc.demon.nl>

Hello,

If people start patching, I can't stay behind :-)

This patch adds something to btpd that has I have missed for a while: a way to 
rename the (display) name of a torrent. You can set a name with 'btcli add', 
but you can't change it afterwards, other than by deleting the torrent and 
adding it again. Of course, this resets your stats/ratio. So I've added the 
following command:

	btcli rename torrent# new_name

The patch is against the 0.15 tarball.

 - Nemosoft
-------------- next part --------------
A non-text attachment was scrubbed...
Name: btpd-rename.diff
Type: text/x-diff
Size: 8955 bytes
Desc: not available
Url : http://lists.stargirl.org/pipermail/btpd-users/attachments/20090213/58b0de13/attachment-0001.diff 

From borg at uu3.net  Sat Feb 14 09:28:43 2009
From: borg at uu3.net (Unknown)
Date: Sat, 14 Feb 2009 09:28:43 +0100 (CET)
Subject: [PATCH] clock_gettime on system that don't support CLOCK_MONOTONIC
In-Reply-To: <bc772bd10902130712xf2943b0wab14b28085538c6d@mail.gmail.com>
References: <bc772bd10902130606u458f22ddh36eccda48af56d69@mail.gmail.com>
	<2441.85.24.200.83.1234537412.squirrel@webmail.stargirl.org>
	<bc772bd10902130712xf2943b0wab14b28085538c6d@mail.gmail.com>
Message-ID: <Pine.LNX.4.64.0902140922220.21217@cube>

> On Fri, Feb 13, 2009 at 4:03 PM, Richard Nyberg <rnyberg at murmeldjur.se> wrote:
>> However, I'm going to add a fallback to gettimeofday. It's equivalent
>> to CLOCK_REALTIME, but exists on more systems.
>
> gettimeofday() will only provide microseconds to clock_gettime()'s
> nanoseconds, will that precision be enough?
LOOL.. sorry.. but are you sure what are you talking about?
only us ?? Do you understand that in your system not much happens
on interfaces in that 1 us ? most upstreams have 10ms at best to gw.
Also, remeber that scheduler clock is around 100 to 1000Hz...
And you babling about ns.. heh.

For Layer 4 stuff.. its enough to play w/ seconds.. belive me.
On long haul backbones pings start above 100ms.. so TCP handshaking
sometimes takes a whole second..

I still dont belive ppl posting such nonsense.. You will waste more
time getting data from more precision counters (I hould say "precision" 
because they arent that accurate actualy) and keep them in sync w/ btpd.

Regards,
Borg


From barone_rosso2 at yahoo.it  Sat Feb 14 17:09:53 2009
From: barone_rosso2 at yahoo.it (barone rosso)
Date: Sat, 14 Feb 2009 17:09:53 +0100
Subject: banned peers  btpd 0.15
Message-ID: <AC49C01C-013E-413C-9112-425875BEAFF2@yahoo.it>


Feb 14 16:55:13 error: Bad hash for piece 3616 of 'xxxxmyfile'.

It seems btpd bans a peer at the first bad data transmitted.

Maybe is better ban a IP/piece after 5 or 10 corrupted data and  
completly an IP after 20 corrupted data.

On MacOsX 10.5.6

--
Cordialmente, vostro

barone rosso




From rnyberg at murmeldjur.se  Tue Feb 17 17:58:54 2009
From: rnyberg at murmeldjur.se (Richard Nyberg)
Date: Tue, 17 Feb 2009 17:58:54 +0100 (CET)
Subject: [PATCH] Renaming torrent
In-Reply-To: <200902132336.21471.nemosoft@smcc.demon.nl>
References: <200902132336.21471.nemosoft@smcc.demon.nl>
Message-ID: <4652.85.24.201.34.1234889934.squirrel@webmail.stargirl.org>

On Fri, February 13, 2009 23:36, Nemosoft Unv. wrote:
> Hello,
>
> If people start patching, I can't stay behind :-)
>
> This patch adds something to btpd that has I have missed for a while:
> a way to rename the (display) name of a torrent. You can set a name
> with 'btcli add', but you can't change it afterwards, other than by
> deleting the torrent and adding it again. Of course, this resets your
> stats/ratio. So I've added the following command:
>
> 	btcli rename torrent# new_name
>
> The patch is against the 0.15 tarball.
>
Thanks! I might add this. Later on there'll be a more general way of
setting (and getting) btpd and torrent properties.

Regards,
       -Richard



From rnyberg at murmeldjur.se  Tue Feb 17 18:20:18 2009
From: rnyberg at murmeldjur.se (Richard Nyberg)
Date: Tue, 17 Feb 2009 18:20:18 +0100 (CET)
Subject: banned peers  btpd 0.15
In-Reply-To: <AC49C01C-013E-413C-9112-425875BEAFF2@yahoo.it>
References: <AC49C01C-013E-413C-9112-425875BEAFF2@yahoo.it>
Message-ID: <4752.85.24.201.34.1234891218.squirrel@webmail.stargirl.org>

On Sat, February 14, 2009 17:09, barone rosso wrote:
>
> Feb 14 16:55:13 error: Bad hash for piece 3616 of 'xxxxmyfile'.

So a piece failed the hash check. Then,
1. No peer that contributed to that piece will be used when
   downloading that piece again.
2. If a peer happen to be suspect in 3 pieces w/o delivering good
   pieces as well, it'll be blocked until it can be cleared of
   suspicion or banned.
3. When a bad piece is redownloaded successfully the bad peer(s)
   can be identified and banned.

Note: 1. can be a problem and maybe I can relax this. I rate the
problem as less than continuously redownloading bad data though.

> It seems btpd bans a peer at the first bad data transmitted.
>
> Maybe is better ban a IP/piece after 5 or 10 corrupted data and
> completly an IP after 20 corrupted data.

Well, possibly one shouldn't ban a peer because of just one failed
piece, since that may be the only damaged one it has. But how do you
reason? And do you mean bad blocks or pieces by bad data?

Regards,
       -Richard



From barone_rosso2 at yahoo.it  Tue Feb 17 22:25:38 2009
From: barone_rosso2 at yahoo.it (barone rosso)
Date: Tue, 17 Feb 2009 22:25:38 +0100
Subject: banned peers  btpd 0.15
In-Reply-To: <4752.85.24.201.34.1234891218.squirrel@webmail.stargirl.org>
References: <AC49C01C-013E-413C-9112-425875BEAFF2@yahoo.it>
	<4752.85.24.201.34.1234891218.squirrel@webmail.stargirl.org>
Message-ID: <8E4EB77A-778C-473F-BA7D-2734E6DFB1D9@yahoo.it>


Il giorno 17/feb/09, alle ore 18:20, Richard Nyberg ha scritto:

> On Sat, February 14, 2009 17:09, barone rosso wrote:
>>
>> Feb 14 16:55:13 error: Bad hash for piece 3616 of 'xxxxmyfile'.
>
> So a piece failed the hash check. Then,
> 1. No peer that contributed to that piece will be used when
>   downloading that piece again.
> 2. If a peer happen to be suspect in 3 pieces w/o delivering good
>   pieces as well, it'll be blocked until it can be cleared of
>   suspicion or banned.
> 3. When a bad piece is redownloaded successfully the bad peer(s)
>   can be identified and banned.
>
> Note: 1. can be a problem and maybe I can relax this. I rate the
> problem as less than continuously redownloading bad data though.
>
>> It seems btpd bans a peer at the first bad data transmitted.
>>
>> Maybe is better ban a IP/piece after 5 or 10 corrupted data and
>> completly an IP after 20 corrupted data.
>
> Well, possibly one shouldn't ban a peer because of just one failed
> piece, since that may be the only damaged one it has. But how do you
> reason? And do you mean bad blocks or pieces by bad data?
>
> Regards,
>       -Richard
>
>

I try write better.

Downloading a torrent whit only one seeder, I see zero traffic and log  
Feb 14 16:55:13 error: Bad hash for piece 3616 of 'xxxxmyfile'.
I stopped  and arestartin the torrent and after 2 minutes I was in the  
same situation, but with a different piece in log.


--
Cordialmente, vostro

barone rosso




From nemosoft at smcc.demon.nl  Mon Feb 23 19:18:09 2009
From: nemosoft at smcc.demon.nl (Nemosoft Unv.)
Date: Mon, 23 Feb 2009 19:18:09 +0100
Subject: [PATCH] Renaming torrent
In-Reply-To: <4652.85.24.201.34.1234889934.squirrel@webmail.stargirl.org>
References: <200902132336.21471.nemosoft@smcc.demon.nl>
	<4652.85.24.201.34.1234889934.squirrel@webmail.stargirl.org>
Message-ID: <200902231918.09221.nemosoft@smcc.demon.nl>

Hello,

On Tuesday 17 February 2009 17:58:54 Richard Nyberg wrote:
> On Fri, February 13, 2009 23:36, Nemosoft Unv. wrote:
> > Hello,
> >
> > If people start patching, I can't stay behind :-)
> >
> > This patch adds something to btpd that has I have missed for a while:
> > a way to rename the (display) name of a torrent. You can set a name
> > with 'btcli add', but you can't change it afterwards, other than by
> > deleting the torrent and adding it again. Of course, this resets your
> > stats/ratio. So I've added the following command:
> >
> > 	btcli rename torrent# new_name
> >
> > The patch is against the 0.15 tarball.
>
> Thanks! I might add this. Later on there'll be a more general way of
> setting (and getting) btpd and torrent properties.

Okay, cool. While you're at it, then... It might be handy to add a 'group' 
field to a torrent. Something that only btpd uses, but it could be used to 
sort torrents, turn on/off whole groups, etc. 

Regards,

 - Nemosoft