I have built TAP for Sun Solaris 8 for SPARC. Here is what I have done to make it work. 1. Use apache, not apache2 2. Edited config.mk, in particular: LDSHARED=gcc -shared BUILDTYPE+=-g -Wall -pthreads APRLIBS= -lapr -lrt -lm -lsocket -lnsl -lresolv -ldl -lpthread 3. MemHash_Node_new was allocating memory not on the long's boundary, causing segmentation fault on SPARC during startup, during the database import from the data directory. MemHash_Node_new(TAP_KB *kb, const char *oid) { size_t size = sizeof(MemHash_Node) + strlen(oid) + 1; /* as07: ak03 : Alignment fix!!! */ size = (size + sizeof(long) - 1) & ~(sizeof(long) -1); /* end of fix*/ MemHash_Node *ret = (MemHash_Node *) MemHash_calloc(kb, size); strcpy(ret->name, oid); return (MemHash_NodeHandle) MemHash_PtrToHandle(kb, ret); } That was enough to start enjoying Apache on my personal workstation. I found that issuiing an empty request from kbx crashes my local server: desc> use http://mysun.com:8080/data Using KB http://mysun.com:8080/data desc> desc> identify props> causes Segmentation Fault in TAP_Cursor_AddOutputFilter when it calls asp_pcalloc, because curs is NULL. A link to download the database http://tap.stanford.edu/data/tap.rdf returns an empty messsage, as if the file tap.rdf were missing. A download for the node Resource through the database browser skips a lot of information in the database and returns a broken XML. It generates unpaired tags </0> in the attempt to refer to namespaces, as far as I understand. I am pushing for an ambitious project, using semantic web technologies, and trying to build a demo. The amount of data could be potentially very big. So far I am having fun with TAP. Thank you. Alec Segal
This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 22 March 2009 02:45:11 GMT