[mnet-devel] what Erik Moeller wants in Mnet v0.6.1

Sun Feb 2 18:44:26 GMT 2003

The mention of "10 MB" in this letter is because I estimated that bundling 
wxPython, wxGTK, etc. into the downloadable would make it be 10 MB instead of 
2 MB.

--Z

------- Forwarded Message

Subject: Re: what do you want in Mnet v0.6.1
From: Erik Moeller <moeller at scireview.de>
To: Zooko <zooko at zooko.com>
Date: 02 Feb 2003 16:38:59 +0100

Bryce,

I managed to install wxPython2.4. The Debian unstable package wouldn't
work on testing(=sarge), but I could convert the RPM using alien. This
may be worth pointing out in the docs.

I don't know how I would feel about downloading 10 megs, that should
IMHO not be required. Just tell people in plain language how to satisfy
the dependencies. Linux people are used to that :-). Would it be
possible to check in the cvsinstall.sh script first if the dependencies
are met, and if they aren't, print out some explanations for the
different distros and dependencies?

I just restarted my MNet broker and the GUI works. One thing I noticed
after restarting is that my started downloads were lost. I presume this
is a bug. In the GUI, there were now several 0% downloads like this:

jrT7J-UJj5X6EzswXHieIJIaB8k  
cCCZyv2c2V9cvM50cgbtilYEI40
qUFZrd-yUQXAIJL1pS6cy01Ei2o 
nB5yV-0Kz0blP7OIusPHxS5CO30 

One of them climbed to 85%. I tried downloading the same file from
before the restart again, and that one, in a few seconds, also climbed
to the exact same percentage. So they may be the same file. But IIRC,
the file was previously only at 20% or so. What's going on here? At this
point all my downloads were stalling. A few seconds later, the whole
broker froze, hard (I have attached the last log messages). Killing it
took some effort.

Some comments on the GUI:

It's fairly clean, and that's good. I like the way downloading from the
search window is handled (flagging the file instead of switching the
window).  In the search, I'd like the "Search" button to be "defaulted",
that is, when I hit enter, the search should be started.

When I publish a file, it would be nice to prefill the "file format"
field with the file extension or the file type derived from it (using
magic numbers would be even nicer).

Another feature I'd like to see is search result caching. Since it can
take a while until results come in, I'd like to be able to view recent
search results, even if my broker has been restarted in between. Can you
simply serialize the search screen results upon exit?

Regards,

Erik

Here are the log messages of the last few seconds my broker ran:

2003-02-02_15:28:27 (Conversation) MTM: msgId: <qt747> :: 'request blob'
with <htmww>, completed
2003-02-02_15:28:27 (MojoTransaction) dynamic timing: message
'roundtrip_time[request blob]' to <htmww> took '11.24' seconds (mu:
'23.27', sigma: '16.11')
2003-02-02_15:28:27 (BlockWrangler) <BlockWrangler
1>._handle_block(<cgzeu>); no chunkobj
2003-02-02_15:28:27 (BlockWrangler) <BlockWrangler
1>.start_more_chunks(<FileObj <FileMap 3 with 294 chunkmaps>>) want 21,
have 27, queue 16
2003-02-02_15:28:27 (commstrats) <TCPCommsHandler.TCPCommsHandler
listening :5004 at 0x4130ae44>.use_comm_strategy(<htmww>, <TCP to
212.84.209.24:27741 via <socket object, fd=40, family=2, type=1,
protocol=0>, 893e8c4>): asyncsock: <TCPConnection connected to <htmww>
at 212.84.209.24:27741, 8aa214c>, peername: None
2003-02-02_15:28:27 (Conversation) MTM: msgId: <uzwmy> :: 'request blob'
with <htmww>, completed
2003-02-02_15:28:27 (MojoTransaction) dynamic timing: message
'roundtrip_time[request blob]' to <htmww> took '11.31' seconds (mu:
'23.15', sigma: '16.07')
2003-02-02_15:28:27 (BlockWrangler) <BlockWrangler
1>._handle_block(<c47xm>); no chunkobj
2003-02-02_15:28:27 (Conversation) MTM: msgId: <5zsgt> :: 'do you have
blobs' with <bjsc4>, failed, failure_reason: 'timeout'
2003-02-02_15:28:27 (commstrats) <TCPCommsHandler.TCPCommsHandler
listening :5004 at 0x4130ae44>.use_comm_strategy(<htmww>, <TCP to
212.84.209.24:27741 via <socket object, fd=40, family=2, type=1,
protocol=0>, 893e8c4>): asyncsock: <TCPConnection connected to <htmww>
at 212.84.209.24:27741, 8aa214c>, peername: None
2003-02-02_15:28:27 (Conversation) MTM: msgId: <1o74y> :: 'request blob'
with <htmww>, completed
2003-02-02_15:28:27 (MojoTransaction) dynamic timing: message
'roundtrip_time[request blob]' to <htmww> took '11.39' seconds (mu:
'23.04', sigma: '16.03')
2003-02-02_15:28:27 (BlockWrangler) <BlockWrangler
1>._handle_block(<undxs>); no chunkobj
2003-02-02_15:28:27 (Conversation) MTM: msgId: <tw3xr> :: 'content
tracker lookup' with <518c4>, failed, failure_reason: 'timeout'
2003-02-02_15:28:27 (contentlib) content tracker lookup failed from
<518c4>, reason: 'timeout'
2003-02-02_15:28:28 (commstrats) <TCPCommsHandler.TCPCommsHandler
listening :5004 at 0x4130ae44>.use_comm_strategy(<htmww>, <TCP to
212.84.209.24:27741 via <socket object, fd=40, family=2, type=1,
protocol=0>, 893e8c4>): asyncsock: <TCPConnection connected to <htmww>
at 212.84.209.24:27741, 8aa214c>, peername: None
2003-02-02_15:28:28 (Conversation) MTM: msgId: <i5wzu> :: 'request blob'
with <htmww>, completed
2003-02-02_15:28:28 (MojoTransaction) dynamic timing: message
'roundtrip_time[request blob]' to <htmww> took '11.49' seconds (mu:
'22.92', sigma: '16.00')
2003-02-02_15:28:28 (BlockWrangler) <BlockWrangler
1>._handle_block(<t9psk>); no chunkobj
2003-02-02_15:28:28 (counterparty) using global average dynamic timeout
'170.46' with for 'roundtrip_time[do you have blobs]' to <y54hr>
2003-02-02_15:28:28 (counterparty) accounting with <y54hr>: requesting
free transaction [reason: 'do you have blobs']
2003-02-02_15:28:28 (conversation) ---> <y54hr>: failed to send,
forgetting comm strat and trying to look up new one...
2003-02-02_15:28:31 (BlockWranglingStrategy) <GSR 1>: searching on peer
<36tbf> for blocks [<9tsr4>, <6ha5x>, <6ac7p>, <6iiqd>, <t9psk>,
<t7x11>, ...]
2003-02-02_15:28:31 (commstrats) <TCPCommsHandler.TCPCommsHandler
listening :5004 at 0x4130ae44>.use_comm_strategy(<htmww>, <TCP to
212.84.209.24:27741 via <socket object, fd=40, family=2, type=1,
protocol=0>, 893e8c4>): asyncsock: <TCPConnection connected to <htmww>
at 212.84.209.24:27741, 8aa214c>, peername: None
2003-02-02_15:28:31 (Conversation) MTM: msgId: <w7sdy> :: 'request blob'
with <htmww>, completed
2003-02-02_15:28:31 (MojoTransaction) dynamic timing: message
'roundtrip_time[request blob]' to <htmww> took '14.73' seconds (mu:
'22.81', sigma: '15.96')
2003-02-02_15:28:31 (BlockWrangler) <BlockWrangler
1>._handle_block(<1dzp5>); no chunkobj
2003-02-02_15:28:31 (commstrats) <TCPCommsHandler.TCPCommsHandler
listening :5004 at 0x4130ae44>.use_comm_strategy(<htmww>, <TCP to
212.84.209.24:27741 via <socket object, fd=40, family=2, type=1,
protocol=0>, 893e8c4>): asyncsock: <TCPConnection connected to <htmww>
at 212.84.209.24:27741, 8aa214c>, peername: None
2003-02-02_15:28:31 (Conversation) MTM: msgId: <jreyt> :: 'request blob'
with <htmww>, completed
2003-02-02_15:28:31 (MojoTransaction) dynamic timing: message
'roundtrip_time[request blob]' to <htmww> took '14.87' seconds (mu:
'22.73', sigma: '15.90')
2003-02-02_15:28:31 (BlockWrangler) <BlockWrangler
1>._handle_block(<trbcm>); no chunkobj
2003-02-02_15:28:31 (Conversation) MTM: msgId: <ym7tf> :: 'do you have
blobs' with <rjkaz>, failed, failure_reason: 'timeout'
2003-02-02_15:28:31 (commstrats) <TCPCommsHandler.TCPCommsHandler
listening :5004 at 0x4130ae44>.use_comm_strategy(<htmww>, <TCP to
212.84.209.24:27741 via <socket object, fd=40, family=2, type=1,
protocol=0>, 893e8c4>): asyncsock: <TCPConnection connected to <htmww>
at 212.84.209.24:27741, 8aa214c>, peername: None
2003-02-02_15:28:31 (Conversation) MTM: msgId: <qnhtp> :: 'request blob'
with <htmww>, completed
2003-02-02_15:28:31 (MojoTransaction) dynamic timing: message
'roundtrip_time[request blob]' to <htmww> took '15.02' seconds (mu:
'22.65', sigma: '15.84')
2003-02-02_15:28:31 (BlockWrangler) <BlockWrangler
1>._handle_block(<c8nic>); no chunkobj
2003-02-02_15:28:31 (commstrats) <TCPCommsHandler.TCPCommsHandler
listening :5004 at 0x4130ae44>.use_comm_strategy(<htmww>, <TCP to
212.84.209.24:27741 via <socket object, fd=40, family=2, type=1,
protocol=0>, 893e8c4>): asyncsock: <TCPConnection connected to <htmww>
at 212.84.209.24:27741, 8aa214c>, peername: None
2003-02-02_15:28:31 (Conversation) MTM: msgId: <858zh> :: 'request blob'
with <htmww>, completed
2003-02-02_15:28:31 (MojoTransaction) dynamic timing: message
'roundtrip_time[request blob]' to <htmww> took '15.11' seconds (mu:
'22.57', sigma: '15.78')
2003-02-02_15:28:31 (BlockWrangler) <BlockWrangler
1>._handle_block(<cuqhy>); no chunkobj
2003-02-02_15:28:31 (commstrats) <TCPCommsHandler.TCPCommsHandler
listening :5004 at 0x4130ae44>.use_comm_strategy(<htmww>, <TCP to
212.84.209.24:27741 via <socket object, fd=40, family=2, type=1,
protocol=0>, 893e8c4>): asyncsock: <TCPConnection connected to <htmww>
at 212.84.209.24:27741, 8aa214c>, peername: None
2003-02-02_15:28:31 (Conversation) MTM: msgId: <iwtgw> :: 'request blob'
with <htmww>, completed
2003-02-02_15:28:31 (MojoTransaction) dynamic timing: message
'roundtrip_time[request blob]' to <htmww> took '15.22' seconds (mu:
'22.50', sigma: '15.71')
2003-02-02_15:28:31 (BlockWrangler) <BlockWranglerData 1>.swb(<FileObj
<FileMap 3 with 294 chunkmaps>>); blocks(782 wanted, 265 needed, 3060
completed, 192 queued)
2003-02-02_15:28:31 (BlockWrangler) <BlockWranglerData
1>.stop_wanting_blocks(1: (<c4q56>,)); I now want 737 blocks. I think I
still need 321
2003-02-02_15:28:32 (FileObj) <FileObj <FileMap 3 with 294
chunkmaps>>.bump_progress(), self.processedblocks: 3258
2003-02-02_15:28:32 (counterparty) using global average dynamic timeout
'170.46' with for 'roundtrip_time[do you have blobs]' to <36tbf>
2003-02-02_15:28:32 (counterparty) accounting with <36tbf>: requesting
free transaction [reason: 'do
you have blobs']
2003-02-02_15:28:32 (conversation) ---> <36tbf>: failed to send,
forgetting comm strat and trying to
look up new one...
- -- 
Scientific Reviewer, Freelancer, Humanist -- Berlin / Germany
Phone: +49 (0)30 45491008 -- Web: http://www.humanist.de/erik
Editor of: http://www.violence.de, http://www.infoanarchy.org

	Save the Public Domain - Renaissance Now:
http://lists.infoanarchy.org/mailman/listinfo.cgi/action

------- End of Forwarded Message

-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com
_______________________________________________
mnet-devel mailing list
mnet-devel at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mnet-devel