the popular ht://dig search engine NOTE: this is my first port, ever. I know that I missed the .info stuff, and it's possible that I may have missed something else. portlint thought I did all right, though.
[Please note that this is _not_ a full review of your port, nor am I volunteering to commit it. You mentioned that it was your first port submission so I looked-over it quickly.] [The above note is only in existence to prevent you from assuming otherwise, which seems to happen too often. It is a standard disclaimer, only...not at all customized towards you! :-] On Wed, 1 Jul 1998, Bill Fumerola wrote: > Xpost-patch: > X ${RM} -rf ${WRKSRC}/gdbm* It's questionable whether you should be removing files in the *patch rules, even if those files aren't used. Somebody may want to do ``make patch'' and then hack on the port themselves, perhaps using the boxed gdbm. > sed 's/^X//' >/usr/ports/textproc/htdig/pkg/COMMENT << 'END-of-/usr/ports/textproc/htdig/pkg/COMMENT' > Xhtdig is a search engine similar to glimpse COMMENT is not allowed to start with ${PKGNAME}. Somebody should add a check for this in portlint. May I suggest in place, A web search engine for sites with a small number of pages. > Xwww/data/htdig/search.html > Xwww/data/htdig/button1.gif [...] > Xwww/cgi-bin/htsearch [...] > Xwww/htdig/bin/rundig I'm not familiar with how websites are set-up. If throwing these things into www/ is normal, then fine. Otherwise, you may consider share/htdig, libexec/, and bin/, respectively. > X-DEST= /opt/www/htdig > X+DEST= /usr/local/www/htdig MeepMeepMeepMeepMeep! Zzzzwwweeerrrrkkkrrrrpppllluuunk! ;-) If you can get away with changing these to ${PREFIX}/www/htdig then do so. If, in the worst case they have to be hardcoded, then add add a rule, post-patch: /usr/bin/perl -pi.orig -e 's!/opt/!${PREFIX}/!' ${WRKSRC}/CONFIG Try very hard to avoid patching CONFIG twice. Ie., don't have a patches/patch-?? that applies to CONFIG and then also have the above rule unless you really want to risk confusing any spectators. I strongly believe that obeying ${PREFIX} is important. If we want to do away with some variable, let's do away with ${NO_PORT_DOCS}, or wahtever that stupid thing is called. [If this is GNU_CONFIGURE, there is quite possibly an easier way to change these things] > echo x - /usr/ports/textproc/htdig/patches/patch-ab > sed 's/^X//' >/usr/ports/textproc/htdig/patches/patch-ab << 'END-of-/usr/ports/textproc/htdig/patches/patch-ab' > X--- configure.in.old Fri Aug 15 17:32:15 1997 > X+++ configure.in Wed Jul 1 13:05:40 1998 MeepMeep! This wouldn't have to be GNU_CONFIGURE=yes, would it? [That "MeepMeep" thing is just a little noise I hear in the back of my head (sometimes the front)...it is not in any way intended to be condenscending, crude, or stupid. :-] > X AC_PATH_PROG(TSORT, tsort) This really smells like GNU_CONFIGURE=yes... > X+LIBDIRS= -L../htlib -L../htcommon -L../@RX_DIR@/rx -L/usr/local/lib > X+INCS= -I../htlib -I../htcommon -I../include -I../@RX_DIR@/rx -I/usr/local/include -L @PREFIX}/lib -I @PREFIX@/include Other than that, it looks good! [And if you're at this point wondering how much "other than that" there is, don't worry---it's possible to nitpick most of the ports that have been committed, too :-] -- Outnumbered? Maybe. Outspoken? Never! tIM...HOEk
On Wed, 1 Jul 1998, Tim Vanderhoek wrote: > > X+++ configure.in Wed Jul 1 13:05:40 1998 > > MeepMeep! This wouldn't have to be GNU_CONFIGURE=yes, would it? ^^^^^^^^^^ "happen to be" -- This .sig is not innovative, witty, or profund.
State Changed From-To: open->closed Closed per author's request .