2008-07-14 |
Sjoerd Mullender |
Fixed error clean up.
|
2008-07-13 |
Niels Nes |
fixed leak
May2008
|
2008-07-13 |
Niels Nes |
fix bug in handling references (old/new values) in triggers
May2008
|
2008-07-13 |
Martin Kersten |
More protection against buffer overflows.
May2008
|
2008-07-13 |
Niels Nes |
approve output
May2008
|
2008-07-13 |
Niels Nes |
backported fix (test independend of other tests)
May2008
|
2008-07-13 |
Martin Kersten |
Approved output
|
2008-07-13 |
Martin Kersten |
Make the test independent of others that might leave tables around
|
2008-07-13 |
Niels Nes |
output errors (allthough unexpected in this part of the code)
|
2008-07-13 |
Niels Nes |
incompatible backport of a fix. SQL functions returning tables return bat[:str,:BAT]. We now correctly return this full type vs only bat before.
May2008
|
2008-07-13 |
Martin Kersten |
Make the return type for table returning functions more precise
|
2008-07-13 |
Martin Kersten |
void expensive clearing of the memory used as a heap
|
2008-07-13 |
Martin Kersten |
garbageCollector - remove always true issue
|
2008-07-13 |
Niels Nes |
dataflow can now include the thread save sql operations
|
2008-07-13 |
Niels Nes |
add locks around trace output
|
2008-07-13 |
Niels Nes |
many fixes for the parallel dataflow
|
2008-07-13 |
Niels Nes |
we like to know about garbage control statements
|
2008-07-13 |
Niels Nes |
4 more signed long longs vs %llu in the scanf format string
|
2008-07-13 |
Niels Nes |
fix compilation (icc even checks for signedness compatible types in scanf)
|
2008-07-13 |
Niels Nes |
initialize stream *s solves compilation issue
|
2008-07-12 |
Martin Kersten |
align documentation with real situation.
|
2008-07-12 |
Martin Kersten |
Make the target variable of an accumulator identical to the first argument.
|
2008-07-11 |
Ying Zhang |
- load module 'lock' before loading xrpc_{client,server} modules
|
2008-07-11 |
Ying Zhang |
the BATs passed to print_result are all void-headed BATs
|
2008-07-11 |
Ying Zhang |
more fixes in param/result parsing
|
2008-07-11 |
Sjoerd Mullender |
Save some more space: this requires a full rebuild.
|
2008-07-11 |
Ying Zhang |
minor change: use index starting from 0 for iter/param/value in error
|
2008-07-11 |
Ying Zhang |
remove tabs
|
2008-07-11 |
Ying Zhang |
- the element type in pre_kind is "ELEMENT", not "ELEM". this fixes the
|
2008-07-11 |
Ying Zhang |
fix of the fix (thank Sjoerd)
|
2008-07-11 |
Sjoerd Mullender |
Got rid of descriptor files. This has a number of consequences which
|
2008-07-11 |
Sjoerd Mullender |
Got rid of descriptor files. This has a number of consequences which
|
2008-07-11 |
Ying Zhang |
check 'seqnr' before write into it.
|
2008-07-11 |
Sjoerd Mullender |
Got rid of descriptor files. This has a number of consequences which
|
2008-07-11 |
Sjoerd Mullender |
showErrors() calling sequence has changed.
|
2008-07-11 |
Niels Nes |
readded mkey.mx
|
2008-07-11 |
Niels Nes |
readded mkey.mx
|
2008-07-11 |
Ying Zhang |
remove unused/unnecessry xrpc env. var
|
2008-07-11 |
Martin Kersten |
New signatures
|
2008-07-11 |
Martin Kersten |
The calling convention for patterns has been changed. It now accepts the
|
2008-07-11 |
Niels Nes |
use mktemp as windws doesn't have mkstemp
|
2008-07-11 |
Niels Nes |
windows compiler seems to tell us that the recycler calls memset with
|
2008-07-11 |
Niels Nes |
add new variance functions
|
2008-07-10 |
Niels Nes |
lots of windows compilation fixes
|
2008-07-10 |
Niels Nes |
fixed support for IS NULL in selection
May2008
|
2008-07-10 |
Niels Nes |
approved tests after hash join changes
May2008
|
2008-07-10 |
Niels Nes |
approving m4 output works for m5 as wel (other way around doesn't this time)
May2008
|
2008-07-10 |
Niels Nes |
test for a bug in the rangejoin 2 select rewriting.
May2008
|
2008-07-10 |
Niels Nes |
approve output after hash join changes
May2008
|
2008-07-10 |
Martin Kersten |
fix all changes before the big bang on patterns
|
2008-07-10 |
Niels Nes |
approve after hash join changes
May2008
|
2008-07-10 |
Niels Nes |
approve output after hash join fix
May2008
|
2008-07-10 |
Niels Nes |
we always build the hash on the smallest bat. All other cases should be handled on a higher level (the frontend optimizers)
May2008
|
2008-07-10 |
Ying Zhang |
Try to fix
May2008
|
2008-07-10 |
Ying Zhang |
Added test for bug
May2008
|
2008-07-10 |
Ying Zhang |
add the new demo files to the RPMs
|
2008-07-10 |
Sjoerd Mullender |
propagated changes of Tuesday Jul 08 2008 - Thursday Jul 10 2008
|
2008-07-10 |
Sjoerd Mullender |
propagated changes of Thursday Jul 10 2008
|
2008-07-10 |
Niels Nes |
it seems we approved incorrect output, so lets approve the correct output now.
May2008
|
2008-07-10 |
Martin Kersten |
Silence compiler
|