| 1 |
On Thursday 18 September 2008 03:24:47 b.n. wrote: |
| 2 |
> > The only case I can think of where _really_ problems might arise is the |
| 3 |
> > (very rare) situation which I had described: That the ./configure script |
| 4 |
> > of X builds X without errors but also without support for Y if only 1.2.2 |
| 5 |
> > of Y is installed: |
| 6 |
> > Then neither later upgrading of Y nor revdep-rebuild will show anything |
| 7 |
> > suspicious, although X does not behave in the intended way. |
| 8 |
> |
| 9 |
> Very clear, thanks. |
| 10 |
> This is, technically, a bug in the ./configure ,however, isn't it? |
| 11 |
|
| 12 |
That's my thinking too. The correct action would be to patch the ./configure |
| 13 |
and submit the patches upstream so everyone benefits. But it's not always so |
| 14 |
easy. |
| 15 |
|
| 16 |
For example, I build e17 from svn and keep my own custom overlay for it. Most |
| 17 |
packages have --enable- options but some features are automagically detected |
| 18 |
at configure time, which leads to the exact problem described above, and it |
| 19 |
has happened to me. So it's not a theoretical possibility but a real (albeit |
| 20 |
rare) one. |
| 21 |
|
| 22 |
-- |
| 23 |
alan dot mckinnon at gmail dot com |
Archives