On 06/22/2012 08:50 AM, illusionoflife wrote: > On Thursday, June 21, 2012 13:39:07 you wrote: >>> IIRC, that was to allow the URL-extraction portion of wget to be built >>> stand-alone, so that it would create a tool that just extract URLs and >>> spit them out, and not as part of some wget run. >> Gah. This of course applies to the STANDALONE for html-parse. >> Presumably, the one in netrc allows it to extract netrc info and then >> spit it out, or something. > > Well, but we still can remove code duplicate. > Or we are trying to spare little bytes and moving all duplicates to single > library is not apporiate? >
We certainly could. Could move the implementation to its own thing, or something. But I guess all the standalone stuff expects to be run as a single "gcc -o standalone some-file.c" thing, or something (which I doubt would even work right now). What it really amounts to, I suspect, is bitrot. Nobody is testing this stuff, nobody is supporting this stuff, and I doubt it even works properly right now, especially with as closely-tied to gnulib as we are now. Probably, the maintainer should make a decision either to actively support these standalone tools, and include them as part of testing, or else to remove all the #ifdef STANDALONE code in there, so that it's not misleading anybody one way or the other. :) -mjc
