[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#634052: transition: glew



On Sat, 16 Jul 2011 15:18:20 +0200
Niels Thykier <niels@thykier.net> wrote:

> On 2011-07-16 15:15, Joost Yervante Damad wrote:
> > On Sat, 16 Jul 2011 14:48:57 +0200
> > Niels Thykier <niels@thykier.net> wrote:
> > 
> >> [...]
> >>
> >> I have checked the 1.5.8-3 -> 1.6.0-2 diff plus the news from
> >> upstream[1] and I get the feeling the SONAME bump is unneeded.
> >>
> >> in config/version:
> >>   SO_MAJOR = $(GLEW_MAJOR).$(GLEW_MINOR)
> >>   [...]
> >>   LDFLAGS.SO = -shared -Wl,-soname=$(LIB.SONAME) [...]
> >>
> >> in config/Makefile.linux:
> >>   LIB.SONAME = lib$(NAME).so.$(SO_MAJOR)
> >>
> >> I suspect that upstream got the SONAME part wrong here and
> >> $(GLEW_MINOR) is not supposed to be in the SONAME.  Could you
> >> please ask your upstream if they really intended to do a SONAME
> >> bump for "minor bug fixes and some new extensions"?
> > 
> > There indeed are two symbols removed besides the usual additions:
> > 
> > -__glPNTrianglewesfATI
> > -__glPNTrianglewesiATI
> > 
> > That could be a valid reason I guess.
> > 
> > Joost
> > 
> 
> Thanks for checking up on that.  Usually symbols like that are
> intended as private symbols; perhaps upstream or you could work
> upstream on hiding these symbols so they can rename/remove them
> without having to bump the SONAME.
> 
> I believe gcc's -fvisability or a linker script ought to do the trick.

Indeed, upstream confirmed that a symbol type fix was reason to
change the soname.

Joost

-- 
Joost Yervante Damad - http://damad.be/joost/



Reply to: