On Wednesday 25 June 2008 07:56:25 Dean Scarff wrote:
> Ping @ipe6 maintainer.
>
> Any chance that ipe6 is going to get updated in unstable?  The current
> version is broken in several respects (try adding text, and if you get
> that far try running LaTeX).
>
> I've successfully built ipe-6.0pre30 on Tiger; finkinfo and patch at:
> http://scarff.id.au/file/finkinfo/graphics/ipe.info
> http://scarff.id.au/file/finkinfo/graphics/ipe.patch
>
> Note I've renamed it so as not to collide with the official ipe6
> package, but also because renaming to ipe6 doesn't make sense when
> there's no legacy version 5 package to conflict with.

Even though the "6" seems unnecessary, there's no reason _not_ to continue 
using it--anybody who wants to test this can avoid collisions simply by using 
their local tree.  Moreover, since it has a -shlibs splitoff and the 
libraries look to be identical to those in the existing ipe6 package, I'd 
think we'd want to keep the same package name as the existing packaging, 
since by policy the -shlibs from ipe would be forbidden to conflict and 
replace those of ipe6, as that would introduce breakage in anything that 
might happen to depend on ipe6-shlibs, and because none of the libversions 
have changed, you would therefore also need need to install the libraries in 
a location that doesn't overlap with where ipe6 puts them.

Assuming that we just keep the existing name, let's modify the Shlibs: lines 
to be

%p/lib/libipe.1.dylib 1.0.0 %n (>= 6.0pre10-14)
%p/lib/libipecanvas.1.dylib 1.0.0 %n (>= 6.0pre10-14)
%p/lib/libipemodel.1.dylib 1.0.0 %n (>= 6.0pre10-14)

(note that these are not fully spelled out in the existing ipe6 package)

You would want to change the syntax anyway, because we prefer a hardcoded 
version in the Shlibs declarations rather than %v:

$ fink validate ipe.info          
Validating package file ipe.info...
Warning: Non-hardcoded version in dependency "%n (>= %v-1)" 
for "%p/lib/libipe.1.dylib" in field "shlibs" of "splitoff". (ipe.info)
Warning: Non-hardcoded version in dependency "%n (>= %v-1)" 
for "%p/lib/libipecanvas.1.dylib" in field "shlibs" of "splitoff". (ipe.info)
Warning: Non-hardcoded version in dependency "%n (>= %v-1)" 
for "%p/lib/libipemodel.1.dylib" in field "shlibs" of "splitoff". (ipe.info)
Warning: Description contains package name. (ipe.info)


However:  does this even need a -shlibs at all?  Are there headers in the main 
package that would enable other packages to build against this and link to 
its libraries?    

I can't verify the above for myself because, unfortunately, it's going to need 
some additional work to build on 10.5:

...
qmake -macx main.pro
make
cd ipelib/ && /sw/lib/qt4-x11/bin/qmake ipelib.pro -macx -o Makefile
cd ipelib/ && make -f Makefile
c++ -c -pipe -O2 -Wall -W -fPIC -D__USE_WS_X11__ -DQT_NO_CAST_FROM_ASCII 
-I/sw/lib/qt4-x11/mkspecs/darwin-g++ -I. -I../include -I. -o 
../../build/obj/ipelib/ipebase.o 
ipebase.cpp
ipebase.cpp: In member function 'IpeString IpeLex::NextToken()':
ipebase.cpp:310: error: 'uint' was not declared in this scope

(the existing ipe6 has similar issues with regard to uint on 10.5 as well)

If the listed maintainer doesn't respond within a reasonable interval, we 
should assume that he no longer has any interest in maintaining the package.

Attachment: signature.asc
Description: This is a digitally signed message part.

-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Fink-devel mailing list
[email protected]
http://news.gmane.org/gmane.os.apple.fink.devel

Reply via email to