1 |
Grant Goodyear wrote: |
2 |
>>I was wondering if switching to CVS would only update the ebuilds or |
3 |
>>also update (relax) the maskings. In other words, it is safe to |
4 |
>> sync the portage tree with gentoo cvs ? |
5 |
> |
6 |
> |
7 |
> Yep. The anoncvs and rsync servers return exactly the same files. |
8 |
|
9 |
So, if I did emerge rsync and the acroread bug was still there, |
10 |
the bug was obviously not fixed (or I didn't get the last release). |
11 |
The bug fixes should differ in -r? . |
12 |
|
13 |
The revision I have is acroread-5.05-r3.ebuild. |
14 |
|
15 |
It contains the following code snippet: |
16 |
|
17 |
for i in Browsers Reader Resource |
18 |
do |
19 |
insinto ${INSTALLDIR} |
20 |
doins ${i}/* |
21 |
done |
22 |
|
23 |
It should copy recursively the Browser, Reader and Resource |
24 |
subdirectories in ${INSTALLDIR}, but instead a warning is issued. |
25 |
(reasonable, since it uses "/bin/install" and it doesn't copy recursively). |
26 |
|
27 |
I have fixed it adding: |
28 |
|
29 |
cp -r Browsers Reader Resource ${D}/${INSTALLDIR} |
30 |
|
31 |
just below. |
32 |
|
33 |
Can you explain me why is preferred to use doins instead of cp ... |
34 |
${D}/${INSTALLDIR}? If I have sources of a package it is safe to use a |
35 |
"emake install" (configuring it previously with something like |
36 |
--prefix=${D}/usr) ? |