Issues 2010-10-26
 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
26890DEFECTtriageP4issues@sconsDir().abspath breaks a UNC windows share path
4
consensus 2.1 p2 decision sgk
5
6
Bill
7
David
8
Gary2.1p2Clearly a bug (oversight). Should be fixed soon. 2.1 would be ideal. Shouldn't be too hard, depending on python's support for UNC paths. I can take it if nobody else volunteers (but I'm trying to get intel11 support in...)
9
Greg
10
JasonThe standard file code in python handles unc path correctly.
11
Jim
12
Ken
13
Sergey
14
Steven2.1p2agree w/gary, we're clearly mapping the root-ness of UNC paths to \, the root for the current drive letter
15
16
26900DEFECTtriageP4issues@sconsClone() fails when some __semi_deepcopy__ attrs are present but are None
17
consensus asap (assumed to be 2.1 p2) decision Gary
18
19
Bill
20
David
21
Garyasapp2Patch is fine, but I'd recommend using callable() instead of is not None -- it's a little safer. Either way though, as Greg says, patch is harmless.
22
Gregtrivial, asapHas patch. Could it have been caused by upgrade (i.e., removing .sconsign would have fixed it)? Worst case, patch is harmless.
23
JasonI don't see any issues with this patch. '+1'
24
Jim
25
Ken
26
Sergey
27
Stevenasapp2patch is fine, agree w/Gary re: callable()
28
29
26910FEATUREtriageP4garyoProject file generation for Visual Studio 2010
30
DONE by garyo
31
32
Bill
33
David
34
Gary2.1p4garyoI integrated 2008 (9.0) and I think someone posted a 2010 version, just needs some tests. (Can't find it right now though)
35
Greg
36
Jason
37
Jim
38
Ken
39
Sergey
40
Steven2.1p4garyothanks...
41
42
26920ENHANCEMENTtriageP4issues@sconsSuggest a InstallDir() action to install the dir and all it's files recursively
43
decision dup 2536
44
45
Bill
46
David
47
Gary2.x?I think I agree w/ Greg, but it does get complicated.
48
Greg???Update is failing because the target directory is already there, so this issue is technically invalid, but I think something like this should be detected dynamically and (at build time) automatically applied. Also note that it overlaps with the SEP on containers (bug #2536).
49
JasonI think the coy logic needs to be improved here to not fail as it currently does. Part of the issue for me is that we really want to be able to expand The Dir node into a set of files that have structure. I feel this is related to the ZipFile issue. This is a core function change that we have to define in SCons in some way.
50
Jim
51
Ken
52
Sergey
53
Steven2.xp3Greg's analysis is correct. We should be more graceful about handling this case, but I don't think a separate InstallDir is the way to go.
54
55
26930PATCHtriageP4issues@sconsRemoving existing read-only files on Windows
56
consensus 2.x p3 decision Jason; catch unlink failure instead of access(); generalize for DOS and *ix
57
58
Bill
59
David
60
Gary2.1 or 2.xp3I like it. Fixes 2 issues, patch is simple, contains test script even. Note that this applies beyond just Windows (and that's appropriate I think). Of course there is some danger in making files writable, but I think it's correct here -- especially on win7. Could be an option if people feel strongly.
61
Greg@Gary: Yes, would not work correctly if files are linked; need to check for DOS as well. Rather than using access(), should catch exception if delete() fails.
62
JasonWe do something simular in Part. However we need to do this on linux as well, as readonly files have the same effect. The code we have is a little different to protect permission that this patch will mess up. ( we found it the hard way :-) )
st = os.stat(fullpath)
os.chmod(fullpath, stat.S_IMODE(st[stat.ST_MODE]) | stat.S_IWRITE)
63
Jim
64
Ken
65
Sergey
66
Steven2.xp3no reason not to take it, Jason's amendment re: preserving permissions is more correct. os.access() isn't perfect here, because it could fail if a directory above the final path become unsearchable, for example, but it's probably adequate for our purposes
67
68
26940DEFECTtriageP4issues@sconsMSVSProject ignores MSVC_VERSION parameter
69
decision research p3 Bill; keep Gary in the loop; pass it to him if needed
70
71
Billresearchbill
72
David
73
GaryClosed as invalid. He was passing MSVC_VERSION to MSVSProject(), not using it in the Environment constructor.
74
Greg
75
Jason
76
Jim
77
Ken
78
Sergey
79
Steven2.x?p3?issue has been reopened, setting it when the environment was created didn't fix it. maybe -research-?
80
81
26950DEFECTtriageP4issues@sconsThe output of MSVSProject doesn't depend on all of it's arguments.
82
decision 2.1 p2 sgk; Action() not passed all the needed variables; trivial fix
83
84
Bill
85
David
86
GaryNot enough to go on; asked OP for more info.
87
Greg
88
Jason
89
Jim
90
Ken
91
Sergey
92
Steven2.xp3?ah, good one. We stick the values for a number of construction variables into the generated project/solution files, but I bet the Python function Action doesn't list them in envvars, which indicates which variables the function Action "depends on." Should be pretty straightforward to fix.
93
94
26960ENHANCEMENTtriageP4issues@sconsImprove performance with --duplicate=copy
95
decision 2.1 p3 sgk; create an issue for future cleanup (share common unlink() function)
96
97
Bill
98
David
99
GaryNice improvement. One question: should it also check size as well as timestamp just in case? (That should also be fast, O(1).)
100
GregHas patch. It should work this way by default.
Loading...
 
 
 
Sheet 1