[QUOTE=Kevin;169973]So today I thought I was stocking up on 31M first-time tests, but after checking online it appears all the new ones I picked up have already been tested once (and aren't listed as suspect). Any ideas about why the server might be doing this?[/QUOTE]

Got any specific exponents you want me to look at?

[QUOTE=Prime95;170103]Got any specific exponents you want me to look at?[/QUOTE]

I just ended up unreserving them all and lowering my "Days of work" so I didn't get anymore exponents. I'll check tonight to see if I'm still getting already tested exponents as first-time tests. The same thing happened on two computers in different locations.

Already tested exps reassigned for a first test
 

George,

I have two examples of the phenomenon :

30112051 No factors below 2^68
Unverified LL 4757E0F394A965__ by "Gunnar Tufvesson"
Assigned LL testing to "S485122" on 2009-03-28
31945157 No factors below 2^68
Unverified LL C47311E971E543__ by "C. Cooper / S. Boone"
Assigned LL testing to "S485122" on 2009-04-19

are two exponents that have been tested with no errors that have been assigned as first time tests.

Jacob

... both of which had not been P-1 tested. (hint?, hint?)

[QUOTE=Prime95;170103]Got any specific exponents you want me to look at?[/QUOTE]

Not sure it's the same thing but I've had several "first-time" LL tests in the 28-29M range that completed lately along with the message:

[QUOTE]LL test successfully completes double-check of M29460733[/QUOTE]

[QUOTE=petrw1;170192]Not sure it's the same thing but I've had several "first-time" LL tests in the 28-29M range that completed lately along with the message:[/QUOTE]

That is a normal message when testing an exponent that had a suspect first-time test.

[QUOTE=S485122;170168]I have two examples of the phenomenon : 30112051[/QUOTE]

Bug found and fixed. Any exponent that needed P-1 factoring was handed out as a first-time test even if it had a clean LL test reported.

[QUOTE=Prime95;170221]Bug found and fixed. Any exponent that needed P-1 factoring was handed out as a first-time test even if it had a clean LL test reported.[/QUOTE]

I have a bunch of them. Basically I take first-time test assignments and do P-1 on them only.

I assume I will keep these assignments until I return them or they expire (which I won't let happen). My question is, when I return them, will they remain available for first-time testing, or revert to doublechecks? The reason is that I prefer to P-1 exponents that are likely to be tested immediately I'm done with them, which won't happen if these will now be treated as doublechecks.

Another question: would it be possible for the server to identify those clients not doing P-1 (whether because of misconfiguration or because they're very old) and ensure that they are only given assignments which already have P-1 done?

[QUOTE=Mr. P-1;170254]My question is, when I return them, will they remain available for first-time testing, or revert to doublechecks?[/QUOTE]

They (exponents that had a clean first test) will revert to double-checks. Those that had a suspect first test will revert to first-time tests.

[QUOTE=Mr. P-1;170255]Another question: would it be possible for the server to identify those clients not doing P-1 (whether because of misconfiguration or because they're very old) and ensure that they are only given assignments which already have P-1 done?[/QUOTE]

No. The primary source of did-not-bother-with-P-1-tests are manual testers. I cannot control what they type into worktodo.txt. The other source would be pre-v21(?) clients. All v24 and earlier clients look the same to the new server.