According to Lachlan Andrew: > How does the following sound: > 1. Target 3.2.0b6 for the end of May > 2. This will be basically a bug-fix / optimisation from 3.2.0b5 > 3. It will include the following bug fixes from Joe's archive > - config_parser.1
My one concern about config_parser.1 is that the conf_parser.cxx was previously built by Neal with a more recent version of bison than what I'm running. I included my build so Joe and others could apply the patch, but I'm concerned that some stuff, e.g. Win32 compatibility, may have been lost in the process. We could really use Neal's input at this point, and maybe a new build of this by him. > - cygwin.0 > - DESTDIR.0 > - exclude_perform.1 These are fine by me. To answer your concerns about DESTDIR.0, it looks fine to me, as far as the installdir/Makefile.am is concerned. The derivative Makefiles would need to be built. The idea behind DESTDIR is to support an alternate root directory at build-time (e.g. BuildRoot in RPM .spec files), so that the files are temporarily installed in a subdirectory, but eventually relocated back to the actual root directory. DESTDIR should never find its way into config files or shell scripts. I haven't checked for similar errors in other Makefiles yet (and we likely ought to) but that shouldn't hold up committing this one fix, the only one I think is needed for installdir. > - extension_filter.0 (already committed) > - operator[].0 (already committed) > - robots.0 robots.0 is already committed, albeit in a modified form. > - TMPFILE.0 The .spec file that TMPFILE.0 fixes could use a lot of other reworking, but it's not a critical file as far as the distribution is concerned, as most (all?) RPM packagers use their own anyway. Having it as a good, reliable, and up-to-date example of how to do it wouldn't hurt, though. > - without-zlib.0 (already committed) > 4. We don't include fileSpace, since it isn't really a bug fix. I'd personally vote to include this one. Though not technically a bug fix, it is a simple fix for a problem that's bothered several users. It just needs a corresponding defaults.cc entry with docs. > 5. For those who want 3.1.6 functionality with minimal overhead, > I suggest the attached patch, which adds a "store_phrases" > attribute. If that is false, the DB size is cut to 1/3 by > storing only the first occurrence of each word in the text of > a document. (Keywords, links etc can still have reduncancy.) > 6. Before we put out 3.2.0b6, we run it past everybody who has > commented on the speed of 3.2.0b5, to make sure that there are > no configurations which cause it to be particularly slow (like > having very many 'exclude's did). > 7. After that, we resume testing for 3.2.0rc1 > > Does that sound reasonable? Works for me. My only concern is the end of May deadline, as I can't personally offer much of my time these days. You guys (especially you, Lachlan) seem to be well on track, though, so it strikes me as achievable with or without my help. -- Gilles R. Detillieux E-mail: <[EMAIL PROTECTED]> Spinal Cord Research Centre WWW: http://www.scrc.umanitoba.ca/ Dept. Physiology, U. of Manitoba Winnipeg, MB R3E 3J7 (Canada) ------------------------------------------------------- This SF.net email is sponsored by: The Robotic Monkeys at ThinkGeek For a limited time only, get FREE Ground shipping on all orders of $35 or more. Hurry up and shop folks, this offer expires April 30th! http://www.thinkgeek.com/freeshipping/?cpg=12297 _______________________________________________ ht://Dig Developer mailing list: [EMAIL PROTECTED] List information (subscribe/unsubscribe, etc.) https://lists.sourceforge.net/lists/listinfo/htdig-dev
