1 |
-----BEGIN PGP SIGNED MESSAGE----- |
2 |
Hash: SHA1 |
3 |
|
4 |
Karl Trygve Kalleberg wrote: |
5 |
> Petteri Räty wrote: |
6 |
> |
7 |
> I don't see how symlinks in /usr/include can work. If there are two |
8 |
> users on a system, each with his own prefered VM, how can we export two |
9 |
> sets of symlinks for that system? |
10 |
|
11 |
True. |
12 |
|
13 |
> |
14 |
> Of course you can argue that this is an unlikely situation, but what if |
15 |
> root stops the merge with a ctrl-C? Then the /usr/include/jni.h will be |
16 |
> completely wrong, in this scheme. |
17 |
> |
18 |
|
19 |
That is is out of the question then. |
20 |
|
21 |
> |
22 |
> As you pointed out yourself, some packages try to be smart and find the |
23 |
> jni.h themselves. If they use something like if [ -x ${path}/jni.h ], |
24 |
> putting an -I in CFLAGS won't help much, and the configure script will |
25 |
> probably stop with an error. |
26 |
> |
27 |
|
28 |
This is one the issues to discuss at DevJam. Standards are a good thing. |
29 |
Well you could ensure that -I is the first option to the compiler but I |
30 |
agree that it we should come up with something more bullet proof. |
31 |
|
32 |
> |
33 |
> If the package tries to find the jni.h by compiling with various -I |
34 |
> options, we may get lucky. The only remaining problem then is ensuring |
35 |
> that we in fact do _not_ have a gcj-installed jni.h in /usr/include, |
36 |
> since that may take precedence. |
37 |
> |
38 |
|
39 |
Let's mail upstream and see what happens. |
40 |
|
41 |
> |
42 |
> |
43 |
> -- Karl T |
44 |
|
45 |
Regards, |
46 |
Petteri Räty |
47 |
-----BEGIN PGP SIGNATURE----- |
48 |
Version: GnuPG v1.4.2 (GNU/Linux) |
49 |
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org |
50 |
|
51 |
iD8DBQFDBXfccxLzpIGCsLQRAgYPAJ9tZp7ftQmlU5mRQZTU62BOKEkBEwCghUZU |
52 |
SIgy+kgCrK5cl6fGJAVCHe0= |
53 |
=qpms |
54 |
-----END PGP SIGNATURE----- |
55 |
-- |
56 |
gentoo-java@g.o mailing list |