| 1 |
On Saturday 13 April 2002 13:06, Terje Kvernes wrote: |
| 2 |
> George Shapovalov <georges@×××××××××××.edu> writes: |
| 3 |
> > Just a quick thought: |
| 4 |
> |
| 5 |
> the same thing happens. `ggz-config --data` returns |
| 6 |
> '/usr/share/ggz' whenever it is called. it does _not_ care about |
| 7 |
> $prefix or DESTDIR or anything at all. :/ |
| 8 |
Too bad :(. However this brings me to the idea. May be it is worth updating |
| 9 |
sandbox functionality so that not only it catches all such attempts but also |
| 10 |
(if for example sandbox-redirect is set or something) redirects write attempt |
| 11 |
to the proper place. |
| 12 |
The rationale: |
| 13 |
There will always be outright broken packages which will not honor any |
| 14 |
redirection flags passed to make or configure. |
| 15 |
I realize this is not gonna be trivial and will probably not be done in a |
| 16 |
close future. |
| 17 |
|
| 18 |
Right now I do not see very clean solution other then to do some sed'ing |
| 19 |
(to add ${D} to the output of /usr/bin/ggz-config). |
| 20 |
I do not see Makefile.in or Makefile so I cannot be more specific (and I am |
| 21 |
not very proficient with sed either), but something along the lines of: |
| 22 |
|
| 23 |
sed -e "s:`/usr/bin/ggz-config':${D}`/usr/bin/ggz-config`:" Makefile.orig |
| 24 |
>Makefile |
| 25 |
|
| 26 |
might work (depending on what ggz-config returns of course). |
| 27 |
(Actually you will need to check if ${D} is passed to Makefile and then either |
| 28 |
add ${D} or its value). |
| 29 |
And I agree that patching is not a very good idea. Using sed seems cleaner to |
| 30 |
me in this situation. |
| 31 |
|
| 32 |
George |
Archives