Message242210
Hi Steve:
I have limited expertise in most of these areas, but I looked at starunpack40.diff and have these comments:
* tests look to have good coverage of the feature (can't speak to coverage of the parser/compiler code)
* parsermodule.c changes comprehension handling, but I thought we pulled this out?
There was some code taken common in various places, but there should be no code for unpacking comprehensions left in. Do you have a question about a specific line?
* why was dictobject.c.h added? I don't understand clinic thoroughly, but it seems a lot of new code for what was changed in dictobject.c
They weren't added. They were moved by someone else. The only changes are exposing a method.
* can the BUILD_(TUPLE|LIST)_UNPACK code in ceval.c share most of the processing? Or is there an unwanted perf impact to that?
Good idea. I'll make that change.
* whoever applies the patch should regenerate importlib.h themselves - just a reminder
Otherwise, I didn't see anything that particularly scared me. Since we apparently don't have anyone willing and with the expertise to thoroughly check the patch, I'd vote for checking it in asap so it has more releases to bake before 3.5 final.
Thanks! |
|
Date |
User |
Action |
Args |
2015-04-29 02:51:38 | NeilGirdhar | set | recipients:
+ NeilGirdhar, gvanrossum, twouters, georg.brandl, terry.reedy, paul.moore, ncoghlan, belopolsky, benjamin.peterson, ezio.melotti, andybuckley, r.david.murray, zbysz, ethan.furman, eric.snow, berker.peksag, Joshua.Landau, steve.dower, pconnell, Jeff.Kaufman, SpaghettiToastBook |
2015-04-29 02:51:38 | NeilGirdhar | set | messageid: <1430275898.11.0.212907027428.issue2292@psf.upfronthosting.co.za> |
2015-04-29 02:51:38 | NeilGirdhar | link | issue2292 messages |
2015-04-29 02:51:37 | NeilGirdhar | create | |
|