Hi, I personally think that the "-v" solution as you mentioned is good enough, cause most users don't care these "recommend" or "suggestion" messages, even "installed" things, they just want a "out of box" application. For people want specific information, like me, "-v" would be a smart thing.
At this moment, it seems that I have to apply this patch manually to get a "custom aptitude". Thanks for your information and patch ! 2014-02-09 9:05 GMT+08:00 Manuel A. Fernandez Montecelo < manuel.montez...@gmail.com>: > forcemerge 647474 720074 > severity 647474 minor > owner 647474 ! > tags 647474 + patch moreinfo > stop > > Hi, > > The problem was introduced here in 2007, after a feature request which > was previously implemented in Ubuntu: > > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=452202 > > > The basic problem is that showing the recommends in command line mode > depends on the option "Quiet" (which can be set by "-q" in the command > line, or "-o Quiet=integer"). > > It turns out that in src/main.cc, this option is set up automatically > to a positive number if the output is not a tty (the case of pipes or > redirections), even if the user didn't request it through the command > line explicitly: > > int curr_quiet = aptcfg->FindI("quiet", 0); > if(seen_quiet) > aptcfg->SetNoUser("quiet", quiet); > if(quiet == 0 && !isatty(1)) > aptcfg->SetNoUser("quiet", std::max(curr_quiet, 1)); > > > The code above cannot be "fixed", since the "progress" operations > depend on this to work correctly. > > I created the patch, attached. The solution is not very good for my > taste, but Daniel Burrows solved it in this way, attaching this > message to the "Quiet" option (which doesn't happen for any other > output than progress-like), so this is the quick and dirty fix chaning > behaviour minimally and fixing this problem. > > I say minimally because I don't think that users will rely (and thus, > complain if changes behaviour) on something as subtle as setting the > option -q explicitly while using pipes/redirection for other reasons, > to specifically avoid printing this message. > > My prefered solution would be to do like with > Suggests-Will-Not-Install, that is, only show > Recommended-Will-Not-Install if verbose>0 (as the patch from Ubuntu, > but Daniel Burrows didn't like the solution, and I don't know why). > Or alternatively, show it inconditionally (which probably users didn't > like, and that's why Ubuntu implemented a solution). > > Opinions? > > -- > Manuel A. Fernandez Montecelo <manuel.montez...@gmail.com> >
