Hi Damir, my personal opinion is that Apple makes a mistake to keep on supporting (as only Unix like platform) the .so and .dylib model (i.e. making a distinction between as shared library and a pluging). It makes porting a lot of software a pain (as we noticed with the ROOT port). Anyway, they are stubborn buggers and we should not expect anything to improve in this area. Having read the mails in your url I noticed the mention of the option: DYLD_BIND_AT_LAUNCH=YES to get by this issue. Does this work? Have you tried? Cheers, Fons. On Fri, 2003-06-06 at 15:27, Damir Buskulic wrote: > Hi Rene, > > > The feature of executing static objects when loading a shared lib is > > working on all the other systems (even old-fashioned systems). > > I would be a mistake if the MacOSX development team is not addressing > > this question seriously. Also I do don't like very much these dylib > > files. > > I tend to agree with you. But, just to show you how this may become > smthing like a "religious" war, I compiled a set of messages that were > exchanged on the darwin-development list on the subject. These mails > show how it is difficult to convince people when they think they are > correct "in principle". If you have the courage to look at the 40 or so > messages, here they are : > > http://wwwlapp.in2p3.fr/~buskulic/static_constructors_in_Darwin.html > > So if you have more influence than me... > > By the way, I hope this can be useful to Masa if he sees some solution > to this problem. > > Cheers > > Damir -- Org: CERN, European Laboratory for Particle Physics. Mail: 1211 Geneve 23, Switzerland E-Mail: Fons.Rademakers@cern.ch Phone: +41 22 7679248 WWW: http://root.cern.ch/~rdm/ Fax: +41 22 7679480
This archive was generated by hypermail 2b29 : Thu Jan 01 2004 - 17:50:12 MET