*** Bug 1167 has been marked as a duplicate of this bug. ***
http://sources.redhat.com/ml/systemtap/2005-q2/msg00249.html
good luck, dude!
version in CVS now handles global log_hist(foo,NBUCKETS) and global linear_hist(foo,LO,HI,STEP) as "quasi scalar" variables and as statistic-valued maps. can't be copied, can't be assigned to or from. operator <<< inserts pe_long values into stats, expressions of type <<< are pe_long, expression value is the inserted value. not lockless as advertized; they're currently just as SMP-expensive as any other sort of variable. only rudimentary (build-stage) testing, no print or inspection operators, no changes to docs yet.
Is it an accident that the translation of the "<<<" operator currently causes an exclusive (write) lock to be taken on the stats array for the operation?
Final bit of functionality for this (histograms in pseudo-array contexts) should finally now be committed. Let me know if you find further missing support.
fnord