1 |
Daniel, answering you in this mail... Yes, probably sometimes we have to |
2 |
discuss basic things in this mailing list but since developers do this |
3 |
basic errors we have to. Technical problems in our portage tree are |
4 |
perfectly valid for this mailing list. |
5 |
|
6 |
В Пнд, 05/01/2009 в 13:55 +0530, Nirbheek Chauhan пишет: |
7 |
> >> SRC_URI="http://www.porcupine.org/forensics/${PN}-1.01.tar.gz" ~~~~ |
8 |
> |
9 |
> Missing MY_PV I presume? |
10 |
|
11 |
Hardcoded version makes version bumps harder. I'm not sure why ebuild |
12 |
was versioned differently from tarbal but yes, in this case it's better |
13 |
to use versionator eclass define PV. |
14 |
|
15 |
> >> RDEPEND="virtual/libc" |
16 |
> > ^^^^^^^^^^^^^^^^^^^^^^ |
17 |
> |
18 |
> Useless deps which are already in @system? |
19 |
|
20 |
Dependency on libc is completely useless since every usable system have |
21 |
C library. |
22 |
|
23 |
> >> S=${WORKDIR}/${PN}-1.01 |
24 |
> > ~~~~ |
25 |
> |
26 |
> Quotes and ${MY_PV} missing? |
27 |
|
28 |
You don't need quotes in assignments. It hardcoded version again. |
29 |
|
30 |
> >> src_compile() { |
31 |
> >> cd ${S}/memdump-1.01 |
32 |
> > " " ~~~~ |
33 |
> |
34 |
> Quotes, and ${PN}-${MY_PV} ? |
35 |
> Although, should this even be required since ${S} is already set correctly? |
36 |
|
37 |
src_compile initial working directory is S so this line could (and |
38 |
should) be just dropped. |
39 |
|
40 |
> >> einfo "testing" |
41 |
> >> if [ "`./memdump -s 344 | wc -c`" = "344" ]; |
42 |
> >> then |
43 |
> >> einfo "passed test" |
44 |
> >> else |
45 |
> >> die "failed test" |
46 |
> >> fi |
47 |
> |
48 |
> Here, why not use |
49 |
> |
50 |
> einfo "testing" |
51 |
> [ "`./memdump -s 344 | wc -c`" = "344" ] || die "failed test" |
52 |
> einfo "passed test" |
53 |
|
54 |
Yup. It's shorter and does same thing. |
55 |
|
56 |
-- |
57 |
Peter. |