NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
Bernward Bretthauer writes: > A bug I reported some time ago seems not to be fixed in netcdf-2.4-beta1: > The following test program stops with "ncattput: extend_array: Invalid > argument" ... Yes, this problem seems to occur only on platforms where realloc(NULL,size) is not the same as malloc(size), such as SunOS 4.1.x. I've incorporated your suggested fix, which will be in the next beta. Thanks for being persistent and reporting the problem again. ======= Dan Packman writes: > I found that installation of this release on an ibm aix 4.1.3 system > failed to create ncgen because STREQ was undefined in ncgentab.c > This, in turn, is created from ncgen.y. I just added > > #define STREQ(a, b) (*(a) == *(b) && strcmp((a), (b)) == 0) > > to the start of ncgen.h and then all compiled fine. Perhaps someone can > explain how this definition is supposed to get resolved. ncgentab.c contains the lines /* get lexical input routine generated by lex */ #include "ncgenyy.c" and ncgenyy.c (generated by lex from ncgen.l) contains the needed STREQ definition. This builds with no problem on our AIX 4.1 system. My guess is that the new distribution might have been unpacked on top of an old distribution, so that an old version of ncgenyy.c was used instead of a new ncgenyy.c built by lex from ncgen.l. If you do a make clean all test in the ncgen directory everything should work, because lex will regenerate ncgenyy.c. If that doesn't fix the problem, please let us know. ______________________________________________________________________________ Russ Rew UCAR Unidata Program russ@xxxxxxxxxxxxxxxx http://www.unidata.ucar.edu
netcdfgroup
archives: