i think i finally have the way to do it it you'll be up for a bit longer. need to finish the code and run some prelim tests first though.

wow. i was thinking this would be really hard and need some more views and such, but luckily an early design decision makes it lots easier by accident. heh.

Tomcat: it is in testing at the moment, but when you do get it it will include external attachments which will really reduce the cpu load on the db server and when we go to 0.11 it will really reduce the db size.

nice :) so far size is not a issue

db is (looking)

21 GB big

when you do get a crash, it will insert a signature document for each os, cpu, os version and major version not matching the crashing worker's signature.

so if you have 1.9.1, 1.9.2, 1.9.3 and mac os 10.5 x86, mac os 10.6 x86, winxp x65, winvista x86, win7 x86 and say winxp crashes on 1.9.1, it will generate signatures with 0 priority for all combinations except winxp 1.9.1

so, shortly (~hour or so) you should have results showing if the crash is reproducible on the other branches and operating systems.

still need to add a view with by location_id (url) so we can look at results and such by url.

ok

so i need to rebuild the views ?

no, not yet.

i haven't added them yet.

but they will be useful when we want to look at all the results for a particular url..

still testing though. st00pid bc errors. shouldn't be too much longer.

:)

oh well.

harth: are you removing the querystring from crashing urls? we need to keep that.