Hi Maurizio, Thanks for pointing it out. I've found this has a long history in our ChangeLog.
The hbmisc one is probably more often used despite being part of one of the buggiest part of our codebase, and to be honest for me the rtl HB_FEOF() function is also a very strange beast (suboptimal to be called in loops and also falls far from regular file handling scenarios.) In xhb HB_FEOF means the hbmisc implementation. While for me it's also more natural to rename the hbmisc one, it would also create more problems. Renaming the core function is problematic because we should find a name which describes functionality and adheres with core rules (so here HB_F_EOF() isn't an option). I have no good suggestion for another name, so either we should find one or simply delete this function which I believe isn't very useful nor it's used by too many users. As for the future and hbmisc: Since the HB_F*() functions were simply copied from NF lib (also by changing the copyright holder) and extended, but the still serve similar purpose, (and similarl bugs, or even more), perhaps it would be nice to fix the corresponding NF_F*() functions and offer them as a replacement for current hbmisc functions. This way the bugs would be solved, redundancy would be avoided, and HB_ namespace usage in contrib would be avoided as well. This option of course needs someone who's willing to fix the NF_F*()/HB_F*() code in the first place. Major problems with it: integers type mess, non-MT compatible, not EOL neutral. Brgds, Viktor On Tue, Jul 21, 2009 at 1:39 PM, Maurizio la Cecilia<m.laceci...@gmail.com> wrote: > I've found a problem due to the past name change of the hb_F_Eof() function > in philes.c ( > http://www.nabble.com/ChangeLog-2007-08-31-03%3A25-UTC%2B0200-Przemyslaw-Cze > rpak-%28druzus-at-priv.onet.pl%29-tp12419138p12419138.html ), because the > new hb_FEof() collides with the other one in hbmisc lib: > > libhbmisc.a:hb_f.o:000004d0 T _HB_FUN_HB_FEOF > libhbrtl.a:philes.o:00000870 T _HB_FUN_HB_FEOF > > The double name, in effect, is still present in xHarbour, avoiding > collisions when both libs are linked. > I suppose the recent issue signaled ( > http://www.nabble.com/libmisc.lib-hb_feof%28%29-always-.t.-tp17100490p171004 > 90.html ) is due to the wrong function called (the hbrtl one, instead of the > hbmisc as expected). > By the way, i think better to have a hb_F_Eof() in hbmisc and hb_FEof() in > hbrtl, as the last is a more standard name for the function and the one in > hbmisc is related to txt files only. > Best regards. > > Maurizio la Cecilia > > _______________________________________________ > Harbour mailing list > Harbour@harbour-project.org > http://lists.harbour-project.org/mailman/listinfo/harbour > _______________________________________________ Harbour mailing list Harbour@harbour-project.org http://lists.harbour-project.org/mailman/listinfo/harbour
