On Thu, Jan 26, 2006 at 04:33:12PM +1300, Mark Robinson <[EMAIL PROTECTED]> was heard to say: > this is still borken > > piwakawaka:~# aptitude -qq update > Expected a number after -q=, got q > piwakawaka:~# > > which is a lie, there was no -q= > > Here's the relevant code from main.cc in aptitude-0.4.1: > > case 'q': > if(optarg == 0) > ++quiet; > else > { > if(*optarg == '=') > ++optarg; > > if(*optarg == 0) > { > fprintf(stderr, _("Expected a number after -q=\n")); > return -1; > } > > char *tmp; > quiet = strtol(optarg, &tmp, 0); > > if(*tmp != '\0') > { > fprintf(stderr, _("Expected a number after -q=, got > %s\n"), > optarg); > return -1; > } > } > seen_quiet = true; > break; > > Obviously the block dealing with the parameters needs to be defined > contingent on the existance of the "=".
Ew. > The -q=n syntax is anomalous, every other option uses -x <value>. I suggest > that -q= be deprecated and that -qqq or -q n be encouraged instead. I think that would be great, and it's how things originally worked; unfortunately, this led to user complaints (see earlier in this bug log). And, predictably, it appears that adding a weird special case had unforseen side effects (who woulda thunk it!) > A side note: Attempting to compile aptitude throws up a linker error about > not being able to find cppunit. I assume this should be added as a > build-depends ? That should only happen when you run "make check" or do a package build, and libcppunit-dev is a build-depends already AFAIK. Which version are you trying to build and how? Daniel
signature.asc
Description: Digital signature