Changes between Version 26 and Version 27 of CottoTasklist
- Timestamp:
- 08/03/10 22:18:14 (11 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
CottoTasklist
v26 v27 2 2 3 3 == todo (important, in decreasing order): == 4 - some way to measure how much time is being spent in each nqp-rx subroutine or method? I know we had stuff using kcachegrind at one point, but I could never get it to work much with pge or nqp-rx because it seemed to have difficulty with tailcalls. 4 5 - make --stagestats for PCT::HLLCompiler less noisy 5 6 - Take a shot at implementing one-pass lazy PMC class initialization. 6 7 - http://irclog.perlgeek.de/parrot/2010-06-20#i_2460692 7 - Look into a process for GPG signing releases8 - as a bare minimum, add SHA1 (or better) checksums of the tarballs to the release announcement9 - get feedback from #ps folks, update release_manager_guide.pod et al10 8 - Make the [wiki:Deprecation Deprecation] page not suck 11 - add sub-pages for migration between subsequent supported releases12 9 - troll through svn logs and fill in what needs to be done 13 - make such updates a *required* *non-optional* *you-will-get-yelled-at-if-you-forget-this* part of the deprecation cycle14 - see if Coke, kthakore and the Rakudo hackers get less crabby15 - Look at khairul's code:16 - look through meeting log, add tasks to this todo list17 10 - add stuff from http://irclog.perlgeek.de/parrot/2010-05-28#i_2378589 to LoritoRoadmap 18 11 - Add a test for #1652 to t/profiling/profiling.t … … 24 17 - docs/parrotbyte.pod duplicates some of pdd13. Merge parrotbyte.pod into the pdd. It's hard enough to keep one document up-to-date. 25 18 - It'd also be nice to speed up Parrot_Sub_get_line_from_pc since it currently eats about 15% of the processing time *after* caching its results. Knowing PackFiles might help. 26 - ~~Figure out how to properly fix pbc_merge (TT #1419)~~27 - Once this is working, pir files can be compiled to individual pbc files and be merged similar to how C code is compiled.28 - This will make dependency tracking for pir code much more reliable.29 - Add --hash-seed=xxx to parrot so that hash order-related failures can be diagnosed more quickly. [http://irclog.perlgeek.de/parrot/2009-12-18#i_1853127 Coke] recently ran a case where this would have made life easier. (submitted in tt #1383)30 19 - test profiling output (pprof format) 31 20 - test pprof to callgrind conversion and callgrind-style output … … 34 23 - Look into spitting out NYTProf-compatible output too. 35 24 - Optimize the profiling runloop code. Major refactors should wait until some tests are in place. 36 - ~~Do a code review of r43196.~~ 37 - ~~Abstract output in the profiling runcore to minimize the amount of code that cares about the output format.~~ 38 - ~~Figure out a nice way to integrate annotations into the profile.~~ 25 39 26 40 27 == todo (would be nice): == … … 45 32 46 33 == profiling testing todo: == 47 - Make docs/dev/profiling.pod more skimmable.34 - Make docs/dev/profiling.pod helpful to a (hypothetical) annoyed and scared newbie. 48 35 - Figure out what to test. (see [wiki:TestingProfiling]) 49 36 - Write some fake tests to figure out what the profiling testing interface should look like. … … 57 44 - proper namespace support 58 45 - threads 46 - make profiler useful to HLLs (nqp especially, pge would be helpful too) 47 - time spent per sub/method would be very helpful to pmichaud 48 - deal gracefully with tailcalls 49 - optimize pprof2cg, see [http://irclog.perlgeek.de/parrot/2010-08-03#i_2659752 this] and the ensuing discussion 50 59 51 60 52 == Questions: == 61 * What are "basic blocks" in Callgrind's format? Callgrind's docs aren't at all helpful here. RTFSapplies.53 * What are "basic blocks" in Callgrind's format? Callgrind's docs aren't at all helpful here. UTSL applies. 62 54 * A basic block is a straight line piece of code without any branches or branch targets. It's the smallest individual unit of code to which you can apply a compiler optimization.