Issues 2008-08-18
 Share
The version of the browser you are using is no longer supported. Please upgrade to a supported browser.Dismiss

 
$
%
123
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
|
 
Still loading...
ABCDEFG
1
IDVotesType/MemberMilestonePriOwnerSummary/Comments
2
3
15000DEFECTtriageP3stevenknightParseConfig gets confused about drive-labels under MSYS
4
consensus
5
6
Bill
7
Brandon
8
David
9
Gary2.0p3Steven's right; good for you for sticking with it until the fstab thing came up. It's not about slash dir, it's about this fs mapping layer. If this were done, it could be done as a function attribute on a node (.msys_path?) and those used rather than windows or regular paths. Shouldn't be hi pri though; msys works OK w/o parseconfig now.
10
Greg@Steven, OK, that's comprehensible. However, it doesn't convince me that the problem is specific to ParseConfig(). I'd think it would need to be remapped in a number of different areas, so the hook would have to be to do the mapping whenever _any_ full pathname was decoded. Yes?
11
Jim
12
Ken
13
Sohail
14
Steven1.xp3I finally figured out what he wants, and it's not unreasonable. MSYS maintains an fstab / mount table of mappings from (e.g.) C:\subdir\mingw to /mingw. The MSYS version of ParseConfg() will spit out the MSYS path, e.g. /mingw/usr/include. ParseConfig() needs a hook so that it will recognize that /mingw prefix realy means C:\subdir\mingw and change the lookup string accordingly so it gets the right underlying Windows path. Hokay.
15
16
21650DEFECTtriageP4stevenknightpossible race condition explanations
17
consensus
18
19
Bill
20
Brandon
21
David
22
Gary
23
GregSteven was supposed to merge this information into 2124 and close this one.
24
Jim
25
Ken
26
Sohail
27
StevenDUPJust updated it.
28
29
21760PATCHtriageP4stevenknightBuilderWrapper __getattr__ should not raise KeyError
30
consensus
31
32
Bill
33
Brandon
34
David
35
Gary1.0.xp2Ludwigagree w/ Greg
36
Greg1.0.xp2LudwigHe's absolutely right, and the fix is simple. After he fixes the patch, let him apply it. Not a regression, not 1.0.1, but soon thereafter.
37
Jim
38
Ken
39
Sohail
40
Steven1.0.xp2Ludwigagree w/Greg
41
42
21770PATCHtriageP4stevenknightDir get_contents kills memory
43
consensus
44
45
Bill
46
Brandon
47
David
48
Gary1.0.xp2anyone"foolish consistency is the hobgoblin of little minds." IMHO, fix asap. Rebuilds only in certain cases, and bug is severe, could be showstopper for adoption for some users.
49
Greg2,0 {;-}p1anyoneCauses a rebuild in some cases, so our policy says it cannot be applied until a major release. Yeah, that sucks, but I'm not the one who espoused that policy.
50
Jim
51
Ken
52
Sohail
53
Steven1.0.xp1anyoneI'll go with the crowd on whether we stick strictly to stated policy, or whether this is severe enough that we bend the rules. If so, we should have a concrete reason why that rule is being bent and use it to build up a more nuanced policy. For example, is the fact that the rebuilds would only happen in a corner case enough?
54
55
21780PATCHtriageP4pankratLight-weight Attrs class using __slots__
56
consensus
57
58
Bill
59
Brandon
60
David
61
Gary2.0p3LudwigDon't some builds get really large numbers of Nodes? Could be significant for some. Also could lead to better cache behavior (perhaps).
62
Greg2.0p2LudwigIf we set the floor to Python 2.3, this is a slam dunk.
63
Jim
64
Ken
65
Sohail
66
Steven2.0p4LudwigPretty small amount of memory saved, espcially since the Attrs class isn't heavily used. I'd need a more compelling reason than this to have no qualms about moving to 2.2 as the base.
67
68
21790PATCHtriageP4pankratNull Executor memory reductions
69
consensus
70
71
Bill
72
Brandon
73
David
74
Gary1.xp2Ludwigsounds good.
75
Greg1,xp2LudwiigI don't see any dependencies on any fancy features, so apply it as soon as we're out of the bug fixes.
76
Jim
77
Ken
78
Sohail
79
Steven1.xp2LudwigAgree w/Greg.
80
81
21800PATCHtriageP4stevenknightLight-weight Build/NodeInfo classes using __slots__
82
consensus
83
84
Bill
85
Brandon
86
David
87
Gary2.xp2Ludwigagree w/ you guys. Should go into 2.0 (not later) since it changes the .sconsign format.
88
Greg2.xp2LudwigIf we set the floor to Python 2.3, this is a slam dunk.
89
Jim
90
Ken
91
Sohail
92
Steven2.xp2LudwigI think this is a slam dunk even with Python 2.2. The only difference is that 2.3 supports the __weakref__ assignment causing weak references to be used. If I'm reading the doc right, 2.2 will still let you set the __weakref__ variable, it just won't have the effect of using weak references.
93
94
21810PATCHtriageP4pankratintern() redundant strings to reduce memory footprint
95
consensus
96
97
Bill
98
Brandon
99
David
100
Gary1.xp2Ludwigok
Loading...
 
 
 
Sheet 1