Issues 2008-08-11
 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
21242DEFECTtriageP3azverkanInstall() race condition
4
consensus
5
6
Bill
7
Brandon
8
David
9
GarySteven: if it's virus scanners, does the bug go away when you turn them off?
10
GregBrandon was supposed to analyze this in more detail and report back for this bug party. Steven has added a possible explanation as issue 2165 below, but I don't see what that gets us. (Steven, I'm curious why you used Install() instead of Command(Copy())?) @Steven: I agree that using Command(Copy()) will probably still fail.
11
Jim
12
Ken
13
Sohail
14
StevenGary: I'll check. Greg: sounds like I should have added 2165 to this instead of opening a separate issue. Greg: used it for the dumb reason of force-of-abit; Install() has historically been used as a Copy() substitute. I should try Command(Copy()) to see if that changes the behavior. My guess is it won't because Copy() is also implemented by Python.
15
16
15000DEFECTtriageP3stevenknightParseConfig gets confused about drive-labels under MSYS
17
consensus
18
19
Bill
20
Brandon
21
David
22
GaryAgree, WONTFIX. Virtually impossible to know that msys is in use (and what about combinations of msys and native tools?)
23
GregSteven was supposed to resolve this issue with an explanation as to why we WONTFIX it.
24
Jim
25
Ken
26
Sohail
27
StevenMy bad, I'll WONTFIX it.
28
29
21611DEFECTtriageP4stevenknightRace condition with #import
30
consensus
31
32
Bill1.0.xAccording to visual studio docs, #import is current, not obsolete for visual studio c++ files. http://msdn.microsoft.com/en-us/library/8etzzkb6.aspx So this is likely a real bug for folks using visual studio. It looks like the generated .tlh file is the #included by other files. This is used for Type Libraries. (like code generated for IDL rpc calls I beleive). It may make sense to have a visual studio specific scanner in msvc.py ?
33
Brandon
34
David
35
GaryAgree w/ Bill that it's in current use in MSVS, but it's only used to import type libs. So I'd say later than 1.0.x. There are 2 issues: (1) it's hard to add a custom scanner for #import, and (2) what do you do even once you have it to prevent multiple commands making the same .tlh? SCons won't be happy no matter what happens here unless users make separate .tlh files and combine them later.
36
GregwontfixThis is a legitimate race, with several aspects: First, #import is obsolete, and not in the current standard (should use #incude with guards). Second, adding SideEffect()s for these files would make compiles unnecessary sequential (all should wait until the first builds the file then proceed in parallel). Third, it sounds like a pretty specialized corner case and not a lot of people are beating down our doors about it. Right now, #import is ignored by cpp.py; the Classic CPP scanner treats it the same as #include. I don't have a feel for how important this case is, or how long it will be with us, but I'm inclined to WONTFIX it. At a minimum, it's not a regression, so 1.0.x is not in scope.
37
Jim
38
Ken
39
Sohail
40
Steven1.xp3Agree w/Gary and Greg re: not urgent enough for 1.0.x. I'd re-evaluate for 1.x time frame, with non-zero chance of pushing it to 2.0.
41
42
21590DEFECTtriageP4stevenknightMan page errors for Packaging() and Tag()
43
consensus
44
45
Bill1.0.1gregnoel
46
Brandon
47
David
48
Garyagree w/ all
49
Greg1,0,1??meMy issue; my comment is in the issue.
50
Jim
51
Ken
52
Sohail
53
Steven1.0.1gregnoel
54
55
21600DEFECTtriageP4stevenknightGetBuildPath, one arg or two?
56
consensus
57
58
Bill1.xA thought, should we send email to users list requesting people who have access to uncommon platforms? So we can put a matrix up in the wiki for use when we need to reproduce bugs/test fixes/etc?
59
Brandon
60
David
61
Gary1.x?Bill, there's supposed to be such a matrix but it doesn't get a lot of attention. See http://scons.org/wiki/SconsDevelopers.
62
Greg1.xp4?AIX guruMy issue. It could never have worked, so it's not a regression that needs to be fixed. Someone who uses AIX needs to tell us what it's supposed to do.
63
Jim
64
Ken
65
Sohail
66
Steven1.xp4agree w/Greg
67
68
21620PATCHtriageP4gregnoelbison in OS X now creates file.hpp like everybody else
69
consensus
70
71
Bill1.0.1gregnoel
72
Brandon
73
David
74
GaryIdeally scons should figure out the actual behavior at runtime. Otherwise I think there has to be an option or an OS version detection switch. But if this patch passes on 10.4 and 10.5 I think that's sufficient (can ignore 10.2 definitely, and 10.3 probably).
75
Greg1.0.1p3meMy issue. I wonder if it is 'yacc' that generates file.cpp.h, in which case the actual problem (and fix) is different.
76
Jim
77
Ken
78
Sohail
79
Steven1.0.1p3gregnoelReinforcing consensus.
80
81
21630DEFECTtriageP4stevenknightcreateSharedLibBuilder not called for ilink32 tool
82
consensus
83
84
Bill1.0.xAgree with Greg.
85
Brandon
86
David
87
Gary1.0.xNot me; Borland-related. Don't think many people use borland tools anymore. If it satisfies the original poster, I say take it.
88
Gregsee commentgaryo?If it's really that easy, 1.0.x and aim it for 1.0.2. Anything more complex (or having other side effects) should be 1.x p3.
89
Jim
90
Ken
91
Sohail
92
Steven1.0.xagree with greg
93
94
21640DEFECTtriageP4stevenknightbcc32/ilink32 wrong settings for building DLLs
95
consensus
96
97
Bill1.0.xAgree with Greg.
98
Brandon
99
David
100
Garysee commentgaryo?Same comment as previous.
Loading...
 
 
 
Sheet 1