GPU72.com TF worm fetching tool now available

[swl551]

Originally Posted by Aramis Wyler

ohnoes - we all have all lowercase letters in our password! We could be hacked almost twice as fast! We're DOOMED! They're going to unreserve my exponents!

Edit: Good catch, Mog.

I don't have all lower case..... What I have detected is that only the first 8 chars of the password are checked by GPU72. My 9th is upper case and manual testing via IE and FIREFOX shows I can login with or without my 9th character.

That is why I never noticed it during my unit testing.

Let me fix the code anyway.

[chalsall]

Originally Posted by swl551

What I have detected is that only the first 8 chars of the password are checked by GPU72.

Interesting. Excellent detective work.

I'm actually leveraging on Apache's BASIC AUTH, which has other issues I don't like. I didn't know about this one though...

Yet another reason I'd like to implement Digest authentication, but when I tried to do so many users complained that they had had their browser "Remember this password", and the browsers didn't know how to deal with such a change.

As an aside Scott, I'm assuming your fetching code could handle Digest auth as well? (With modification, of course.) I probably won't do this until I finally release a proper fetching API for you to use instead. And, of course, implement a "I've forgotten my password." function on the site for the humans.

[swl551]

Originally Posted by chalsall

Interesting. Excellent detective work.

I'm actually leveraging on Apache's BASIC AUTH, which has other issues I don't like. I didn't know about this one though...

Yet another reason I'd like to implement Digest authentication, but when I tried to do so many users complained that they had had their browser "Remember this password", and the browsers didn't know how to deal with such a change.

As an aside Scott, I'm assuming your fetching code could handle Digest auth as well? (With modification, of course.) I probably won't do this until I finally release a proper fetching API for you to use instead. And, of course, implement a "I've forgotten my password." function on the site for the humans.

I have never implemented an auth request using digest authentication, but many examples exist clearly demonstrating the c# web libraries support it.
I think we'll be OK on that.

[swl551]

Originally Posted by mognuts

I can't seem to get this tool to work, which is very frustrating because it would be very useful. I keep getting the following error:

 
The remote server returned an error: (401) Unauthorized.

This implies (I think) that I'm using the wrong username and password but I've checked and double checked and still no joy. One thing I've noticed is that looking at the console output it converts the first character of my password from upper case to lower case. For example, lets say that my password is 'Bloggs', it it submitting 'bloxxxxxx'. Could this be the reason and if so, how do I solve it? Any information greatfully received.

mog

Fixed.
version 1.0.4
1. Fixed bug that forced password into all lower case.
2. If program startup contains NO command line parameter it tries to use GPU72config.txt as the default. Exception is thrown if no config file can be resolved.

Get from http://mersenneforum.org/misfit/

[Aramis Wyler]

Default config option - nice.

[mognuts]

Just tried the latest version and it works perfectly. Thank you very much for fixing this so quickly. I can't thank you enough!

mog

[mognuts]

Originally Posted by swl551

Currently the work fetcher determines when to fetch work based on the number of "Factor" lines in the target file. An alternative is to fetch if the calculated GhzDays payload of the file drops below a threshold. This provides a much better method for calibrating the balance between work consumption and work fetching.

Is there any interest in this?

The floor is open.

I would be very interested in this, together with a modification to handle the "Let GPU72 decide" option (just setting the option number to 9 doesn't appear to work) and a 'silent mode' where no console window is displayed.

However, since the work fetcher has now been built into MISFIT it seems unlikely there will be any more developments.

[swl551]

Originally Posted by mognuts

I would be very interested in this, together with a modification to handle the "Let GPU72 decide" option (just setting the option number to 9 doesn't appear to work) and a 'silent mode' where no console window is displayed.

However, since the work fetcher has now been built into MISFIT it seems unlikely there will be any more developments.

The fetcher does not enforce any constraint on the #9 so it should work.

command shell script has START /B option to create with no window.

And you are right: No additional development will be performed since MISFIT provides a full featured system. (I will perform bug fixes only)

[chalsall]

Originally Posted by mognuts

...(just setting the option number to 9 doesn't appear to work)...

The system did see a single request from you (2013-02-21 18:53:07 UTC) using MISFIT (or the fetcher -- I can't tell the difference since they both use the same UA) asking for one candidate with the option set to 9. The request was satisified.

[swl551]

Originally Posted by chalsall

The system did see a single request from you (2013-02-21 18:53:07 UTC) using MISFIT (or the fetcher -- I can't tell the difference since they both use the same UA) asking for one candidate with the option set to 9. The request was satisified.

Wasn't MISFIT as I have not released #9 code support.

[mognuts]

Originally Posted by chalsall

The system did see a single request from you (2013-02-21 18:53:07 UTC) using MISFIT (or the fetcher -- I can't tell the difference since they both use the same UA) asking for one candidate with the option set to 9. The request was satisified.

Yes, and no. The website offered me a different range of workunit than the fetcher gave me. I tried it again a few minutes ago and this happened:

The GPU72 website using 'Let GPU72 Decide' offered me:

Factor 61,998,xxx from 72 to 73

and correctly gave me:

Factor=N/A,61998xxx,72,73

but the fetcher (using option 9) gave me:

Factor=2013-2-22,63115xxx,71,73

The requests were only a few seconds apart.