haven't been using public transport recently anyway. cycling every day, which is awesome
ruaok
+1
that's the first thing I did when I got to delft. rented a bike.
now I fell much like a nederlander. I can terrorize the streets and cut off cars with impunity.
alastairp
I went home last weekend and stole dad's bike
ruaok
(more impunity that usual, that is)
warp
:)
ruaok
hi warp!
alastairp
put it on the plane and put it together and rode home from the wirport
warp
hello
alastairp
airport
ruaok
nice.
warp: I'm really looking forward to my ride to rotterdam
warp
ruaok: good :)
ruaok: what time do you think you'll arrive here?
ruaok
good question.
there is a small possibility of me heading to Bologna tomorrow evening.
that would be an 18:35 flight out of schiphol. so I need to work the schedule backwards if that happens. I'd let you know as soon as I know.
if not, I can come whenever.
I'll set aside an hour for riding.
warp nods.
say noon if I don't travel. a bit earlier if I do?
I should know soon.
warp
sure
I'm up early as you know :)
ruaok nods
ruaok
I'll be in rotterdam today... changing trains. I'm headed to Utrect.
+h
warp
ah, right
warp goes back to adding stuff to the webservice.
ruaok: digest authentication sort of works. :)
ruaok
nice.
what is sort of about it?
warp
it needs to store something (the nonce) in $c->cache, which makes mecached mandatory right now.
ruaok
mandatory if you want to submit something to the WS. we can live with that
warp
which is Good Enough. It can probably just use Cache::Memory for dev machines, but I need to poke aCiD2 about how to make that interact with DBDefs.
ruaok
yea, using Cache::Memory is a good idea.
djce
(as long as you have only one fcgi server process, right?) /me not familiar with dev environments these days, but just checking
ruaok
for what case?
if you have no memcached running, NO WS auth request will ever work.
since the nonce gets stored in memached, and fails.
warp
Cache::Memory is fine for development, but I think it could fail in production with multiple processes/threads/whatnots.
djce
just saying, Cache::Memory is a cache within the perl interpreter's memory, right? So if you have >1 perl interpreter at the back end, you get false cache misses.
indeed.
ruaok
yes, right.
djce
And I don't know whether or not it's normal to have >1 back end process in devel.
warp
djce: yes, exactly. but that's quite ok for a developer running script/musicbrainz_server.pl -r -d
djce
ah. ok.
told you I was out of touch with modern mb_server devel :-)
warp
(I'm using Cache::Memory now, and it hasn't failed here yet)
ruaok
and for production we have to ensure that the churn on memcached is low enough that it survives the call duration
djce
yup. separate memcached instance if needs be.
warp
djce: anyway, so the tech stuff is all solved. I just need to know from aCiD2 which old cache stuff can be pruned from DBDefs and how to get the choice between Cache::Memory and memcached in there.
ruaok: ah, right.
ruaok
warp: thats not something we need to worry about yet tho.