You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Instrumenting the build with SetCPPFLAGS:-MD, I see that both the internal libdatrie (C library) and the python glue code (C code wrapping the C library) are using fink's libdatrie1 headers rather than the ones supplied as part of the internal libdatrie library (but still using the internal *.c). Debian's package has a patch to use the external lib. I filed an upstream datrie-py bug about it (pytries/datrie#56).
The reason for the header misuse is that setuputils appends flags rather than prepending, so the default CPPFLAGS (-I%p/include) take priority over anything setup.py adds. One could also argue that this is a long-standing bug in setuputils.
The text was updated successfully, but these errors were encountered:
Gah, you're right. Stupid python. I tried the debian patches and they do allow it to build using libdatrie1, but there are 6 test failures, while all tests pass with the internal libdatrie. I'll have to see if its version is different from our package.
Instrumenting the build with
SetCPPFLAGS:-MD
, I see that both the internal libdatrie (C library) and the python glue code (C code wrapping the C library) are using fink's libdatrie1 headers rather than the ones supplied as part of the internal libdatrie library (but still using the internal *.c). Debian's package has a patch to use the external lib. I filed an upstream datrie-py bug about it (pytries/datrie#56).The reason for the header misuse is that setuputils appends flags rather than prepending, so the default CPPFLAGS (-I%p/include) take priority over anything setup.py adds. One could also argue that this is a long-standing bug in setuputils.
The text was updated successfully, but these errors were encountered: