Re: local copies of libs

2006-10-22 Thread Moritz Muehlenhoff
Hendrik Sattler wrote: >> libavcodec had several vulnerabilities and without doubt it'll have more in >> the next 30 months after Etch release. So it's absolutely necessary to >> link dynamically. (Many do already, e.g. xine-lib). >> I'll file RC bugs for any packages still embedding or link static

Re: local copies of libs

2006-10-22 Thread Moritz Muehlenhoff
Reinhard Tartler wrote: >> libavcodec had several vulnerabilities and without doubt it'll have more in >> the next 30 months after Etch release. So it's absolutely necessary to >> link dynamically. (Many do already, e.g. xine-lib). >> I'll file RC bugs for any packages still embedding or link stati

Re: local copies of libs

2006-10-06 Thread Reinhard Tartler
Moritz Muehlenhoff <[EMAIL PROTECTED]> writes: > libavcodec had several vulnerabilities and without doubt it'll have more in > the next 30 months after Etch release. So it's absolutely necessary to > link dynamically. (Many do already, e.g. xine-lib). > I'll file RC bugs for any packages still emb

Re: local copies of libs

2006-10-05 Thread Hendrik Sattler
Am Donnerstag 05 Oktober 2006 18:01 schrieb Moritz Muehlenhoff: > libavcodec had several vulnerabilities and without doubt it'll have more in > the next 30 months after Etch release. So it's absolutely necessary to > link dynamically. (Many do already, e.g. xine-lib). > I'll file RC bugs for any pa

Re: local copies of libs

2006-10-05 Thread Moritz Muehlenhoff
Hendrik Sattler wrote: > since I often see that packages keep local copies of libs and use those, I= >=20 > kind of want to object to arguments for such build behaviour. > > The latest one I found is xmms-wma: it uses a local stripped-down copy of=20 > ffmpeg's libavcodec and libavformat. > > The g

Re: local copies of libs

2006-10-05 Thread Sam Hocevar
On Thu, Oct 05, 2006, Frank Küster wrote: > If there's ever been a security update for the library, or it's likely > that a buffer overflow or similar would have security implications, then > I think it's definitely more than just wishlist. For example, it's a > pain to patch all those (subtly di

Re: local copies of libs

2006-10-05 Thread Frank Küster
Hendrik Sattler <[EMAIL PROTECTED]> wrote: > For some, the reason is acceptable, for some it isn't? So what makes it > candidate for a bug report with a severity greater than wishlist? > What is the main opinion among Debian maintainers? If there's ever been a security update for the library, or

Re: local copies of libs

2006-10-05 Thread Roberto C. Sanchez
On Thu, Oct 05, 2006 at 03:40:03PM +0200, Hendrik Sattler wrote: > Hi, > > since I often see that packages keep local copies of libs and use those, I > kind of want to object to arguments for such build behaviour. > > The latest one I found is xmms-wma: it uses a local stripped-down copy of > f

Re: local copies of libs

2006-10-05 Thread Hendrik Sattler
Am Donnerstag 05 Oktober 2006 16:10 schrieb Mikhail Gusarov: > You ([EMAIL PROTECTED]) wrote: > > HS> since I often see that packages keep local copies of libs and use > HS> those, I kind of want to object to arguments for such build > HS> behaviour. > > Do you mean debian packages or upstream o

Re: local copies of libs

2006-10-05 Thread Mikhail Gusarov
You ([EMAIL PROTECTED]) wrote: HS> since I often see that packages keep local copies of libs and use HS> those, I kind of want to object to arguments for such build HS> behaviour. Do you mean debian packages or upstream ones? -- JID: [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL P