Daniel Macks <[EMAIL PROTECTED]> wrote: > It does seem like we need a tri-state thing here. It feels funny to > overload a boolean to accomplish that, but I can't think of a cleaner > way, and this allows us to easily find these special-case packages and > do .deb validation. > > The .info could use these three states: > [usual boolean-true values | usual boolean-false values | blank/not-present] > (note drm: just accepting "false" for the new special case makes this > a four-state system) and (assuming dpkg makes a distinction between a > blank value for a control field and that field not being listed) to > get a third "value" for BDO while still being distinct from a .deb > constructed from an "old" fink that did not encode this thing at all.
Ah, I see. I should use "True", "False", and "Undefined" in the .deb (extracted from the three states in the .info file). Then the absence of the field in the .deb just indicates an old .deb on which the test should not be run. The test would be: "if ($BDO =~/Undefined/), and the headers and dylibs are present, then issue the warning." This is good, because the values used for $BDO in fink 0.20.3 and 0.20.4 were only True and False, so those .debs are also treated as "old on which the test should not be run." > > > The BuildDependsOnly field should only be set if the package contains > > header files, installed into /sw/include. > > Wording quibble: I don't think our parser currently makes a > distinction between "present but blank" and "not present" for .info > fields, and "set" sounds a lot like "boolean-true". Maybe something > like "...should be _listed_..." would be clearer? > Yes, something like that would be good. Maybe "should only be declared"? -- Dave ------------------------------------------------------- This SF.Net email sponsored by Black Hat Briefings & Training. Attend Black Hat Briefings & Training, Las Vegas July 24-29 - digital self defense, top technical experts, no vendor pitches, unmatched networking opportunities. Visit www.blackhat.com _______________________________________________ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
