| 1 |
Peter Volkov wrote: |
| 2 |
> В Птн, 30/05/2008 в 20:28 -0700, Brian Harring пишет: |
| 3 |
>> Either way, basically it's coming down to if gentoo wants to follow |
| 4 |
>> the definition of 'academic' right, or 'pragmatic' right. Exempting |
| 5 |
>> ciaran, vote seems to be pragmatic. |
| 6 |
> |
| 7 |
> Well, although I've asked about problems with having --as-needed by |
| 8 |
> default, I'd better go with academic. C++ is quite common language to |
| 9 |
> ignore its design problems and in the end it's not hard to define |
| 10 |
> LDFLAGS in make.conf. |
| 11 |
> |
| 12 |
|
| 13 |
To clarify: |
| 14 |
|
| 15 |
- static initializers (as in __attribute__((constructor), so no, it |
| 16 |
isn't a C++ only feature) have nothing wrong with --as-needed. |
| 17 |
|
| 18 |
- ugly code that refers to undefined symbols that are resolved to ones |
| 19 |
from the main binary and written in the constructor is broken already in |
| 20 |
systems not allowing undefined refs. |
| 21 |
|
| 22 |
- you don't have guarantees about the order in witch the .init sections |
| 23 |
are parsed and constructor function are called, they can be called in |
| 24 |
parallel and you have no means to have a predictable behavior, all you |
| 25 |
know is that everything will be called right before main() or as the |
| 26 |
first thing in dlopen(). |
| 27 |
|
| 28 |
- doing such stuff is uncommon since it isn't the simplest thing to do, |
| 29 |
doesn't work in every place, you have to be particular perverse and |
| 30 |
convoluted even to think about this. |
| 31 |
|
| 32 |
- making such thing go away is good for security, maintainability and |
| 33 |
sanity. |
| 34 |
|
| 35 |
lu |
| 36 |
|
| 37 |
-- |
| 38 |
|
| 39 |
Luca Barbato |
| 40 |
Gentoo Council Member |
| 41 |
Gentoo/linux Gentoo/PPC |
| 42 |
http://dev.gentoo.org/~lu_zero |
| 43 |
|
| 44 |
-- |
| 45 |
gentoo-dev@l.g.o mailing list |
Archives