TODO: ident/filter is in.
Also we would probably want to use gitattributes more.
Junio C Hamano committed
Apr 30, 2007 at 15:29 UTC
6a6e4d1cc1b4869d2ff48585472ab058c4c2ea72
1 file changed
+34
-14
TODO
+34
-14
@@ -24,7 +24,7 @@ From: Linus Torvalds <torvalds@linux-foundation.org>
24
Message-ID: <Pine.LNX.4.64.0704160931550.5473@woody.linux-foundation.org>
25
26
27
-* Use blame machinery to track a single file (not path) in finer
27
+* Use blame machinery to track a single file (not path) in a finer
28
grained way.
29
30
From: Linus Torvalds <torvalds@linux-foundation.org>
@@ -46,17 +46,6 @@ by dropping git-gui/ directory first and then binding git-gui.git
46
as a subproject there. This needs to wait at least until v1.5.2
47
ships and gets installed at kernel.org and repo.or.cz
48
49
-* Decide what to do with 'filter' and 'ident' attribute patches.
50
-
51
-From: Linus Torvalds <torvalds@linux-foundation.org>
52
-Message-ID: <alpine.LFD.0.98.0704212243080.9964@woody.linux-foundation.org>
53
-
54
-If we were to include it, we would need to make it clear that
55
-this is a long piece of rope and you can strangle yourself with
56
-it unless you are careful, and clarify what is and isn't
57
-considered to be "careful". I have parked updates in 'pu', but
58
-I do not know they are clear enough.
59
-
49
* TODO list management
50
51
From: Daniel Barkalow <barkalow@iabervon.org>
@@ -75,12 +64,16 @@ Repeated requests against git-daemon makes it stuck under --syslog
64
65
[jc: does not reproduce easily for me; has anybody seen it?]
66
67
+
68
* git-mirror (reverse of git-push --all).
69
70
From: Shawn Pearce <spearce@spearce.org>
71
Message-ID: <20060926215745.GC8177@spearce.org>
72
83
-[jc: may want the finished version for inclusion]
73
+Pasky sent an updated version to the list recently, but the
74
+general sentiment on the list seems that existing git-fetch
75
+has enough features to perform most of what 'mirror' does.
76
+
77
78
* AsciiDoc 8 would break our documentation.
79
@@ -103,6 +96,31 @@ not just diff with the first parent for a merge.
96
97
[jc: Jakub is interested in it]
98
99
+* Delegate gitweb part to somebody else.
100
+
101
+* Use gitattributes for more things.
102
+
103
+ - 'precious' files that are not tracked but not
104
+ build-products. Currently people seem to put them in
105
+ .gitignore, but that is not quite right, as .gitignore is
106
+ meant for ignoring things that can be lost (build products,
107
+ editor backup files). "git clean -x" and "git checkout" to
108
+ another branch that has a file where the current branch has a
109
+ directory could lose such 'precious' files.
110
+
111
+ - Customized "diff -p" markers per path (Johannes, on #git
112
+ 2007-04-30).
113
+
114
+ I think it makes sense to give an extra parameter to xdiff
115
+ machinery to affect how "diff -p" markers are constructed (as
116
+ opposed to teach xdiff machinery to read gitattributes -- the
117
+ code does not have path information at that level). The
118
+ simplest interface would be to pass a regexp and have the
119
+ existing code always look for that regexp backwards. A more
120
+ complex one would involve a callback function, but I do not
121
+ know if that kind of complexity is worth it.
122
+
123
+ - Others???
124
125
126
Technical (milder)
@@ -130,7 +148,9 @@ Technical (trivial)
148
149
* git-proxy should be spawned with sh -c 'command' $1 $2.
150
133
-[jc: should it?]
151
+[jc: should it? -- deciding if it should may not be "trivial",
152
+but if it turns out to be the right thing to do, the change
153
+itself is trivial.]
154
155
* Maybe a true git-proxy command that reads the first request
156
pkt-line, and redirects the request to its real destination.