- Thanks to Roman Golovin.
- Calls like 'context.columns[name]' autovivify the key 'name' with a default
ctor value, which ends up polluting the context.columns map with every unique
attribute name, which is a lot of 'annotation_nnnnnnnnnn' attributes.
Improves "load" time for all performance tests that load data by ~20%.
- "next" down 15% total
- "list" down 7% total
- "all" down 3% total
- "add" down 15% total
- "export" down 8% total
- "import" down 6% total
fseek() in File::append() was invalidating the file buffer for each call
to append(). Better handling improves "import" performance test by 9%,
45% in "commit", on a system with a spinning disk.
This performance problem affects all operations where
{pending,completed}.data are rewritten. During normal operation a
garbage collection can be enough to trigger it. On storage with high
latency, e.g. networked, this previously took 20 seconds and more.
Reduce the amount of copies necessary for TDB2::gc() by moving the GC
processing to the Task object source - TF2::load_tasks().
This entangles TDB2 and TF2 more than previously, but leads to huge
performance benefits:
- "next" performance test down 21%
- "list" performance test down 11%
- "all" performance test down 4%
- "export" performance test down 9%
The "gc" measurement is down 96% for all performance tests. This is a
result of moving the actual processing into TF2::load_gc() and not
measuring the time taken by that function as "gc" time.
- Thanks to Michele Santullo.
- Although TW-311 references impossible results, there are two problems.
First there is the algorithm that determines estimateѕ completion, and
second there is the reference to wait dates. The algorithm has been replaced
by something better, but wait dates have nothing to do with estimation.
- The 'burndown.bias' configuration setting is removed.
- The estimated completion is based on the net completion rate since the
high water mark of total pending tasks, measured on a daily basis, regardless
of the type of chart produced.
- Vim syntax updated.
- Docs updated.
- Make the Task object's interface more explicit by removing the
std::map inheritance.
- Using this more explicit interface, remove unneeded ctors in order to
allow the compiler to "Do The Right Thing"(tm).
This leads to a performance improvement of 12% in the "add"
performance test, and 7% for "import".
- Leads to 1-3% improvement across all performance tests.
- _I2U could very likely be further optimized by using a sparse vector,
or handling the edge cases of this approach ourselves.
