I've put a patch up at sourceforge (#922881) which is somewhat hackish change in warnings.py that gets back about 50% of the startup time we lost between 2.2.2 and 2.3.x, while avoiding the import deadlock problem with delayed import. The general feeling that I get from the developers that I've surveyed is that it sounds/looks good, but I wanted some more people to take a look at it. I'm also working on a plan for solving the import deadlock problem in the long term (per-module locking - more on that later). This patch does pass all tests. The timings are below: PYTHON STARTUP TIME - 100 INVOCATIONS (no site.py) ------------------------------------- Python 2.2.2 release tree, 9:06, 4/24/2004 ------------------------------------- 3.450u 2.200s 0:05.96 94.7% 0+0k 0+0io 0pf+0w 3.410u 2.270s 0:05.96 95.3% 0+0k 0+0io 0pf+0w 3.400u 2.340s 0:05.92 96.9% 0+0k 0+0io 0pf+0w ------------------------------------------------- 3.420u 2.270s 0:05.95 ------------------------------------------------- ------------------------------------- Current 2.4 tree, 21:30, 04/23/2004 ------------------------------------- 5.410u 4.310s 0:10.50 92.5% 0+0k 0+0io 0pf+0w 5.680u 4.270s 0:10.49 94.8% 0+0k 0+0io 0pf+0w 5.720u 4.330s 0:10.52 95.5% 0+0k 0+0io 0pf+0w ------------------------------------------------- 5.603u 4.303s 0:10.50 ------------------------------------------------- ------------------------------------- Current 2.4 with warnings.py patch ------------------------------------- 4.380u 3.450s 0:08.23 95.1% 0+0k 0+0io 0pf+0w 4.420u 3.410s 0:08.31 94.2% 0+0k 0+0io 0pf+0w 4.580u 3.190s 0:08.24 94.2% 0+0k 0+0io 0pf+0w ------------------------------------------------- 4.460u 3.350s 0:08.26 ------------------------------------------------- -- Nick
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4