The biggest problem I've found with config.cache is with my "development flock". Sometimes I store cached values to speed up the configure process (the whole point of config.cache, eh?). Sometimes I change the logic in configure.in and discover that my cached values are overriding my new logic. This also bites both other developers as well as users who track distributions via CVS instead of tarballs. I'd almost like to see a "version" number in the cache, which I can "bump" whenever I make a "significant" change to the logic which has the effect of... Just had an idea. If the version number controls what sourcing in config.cache does, we could write the config.cache using something like: case "foo_cache_version" in: 1) # foo_cache_version 1 variables get set below ... ;; esac (I generally use "case" instead of "if" because I remember the old days when "test" was not always a shell builtin). This might also solve the problem folks mentioned about the lack of "state" information when doing subdir configures. H
- Re: config.cache considered harmful Tom Tromey
- Re: config.cache considered harmful Martin Buchholz
- Re: config.cache considered harmful Ian Lance Taylor
- Re: config.cache considered harmful Pavel Roskin
- Re: config.cache considered harmful Lars Hecking
- Re: config.cache considered harmful Olly Betts
- Re: config.cache considered harmful Akim Demaille
- Re: config.cache considered harmful Olly Betts
- Re: config.cache considered harmful Pavel Roskin
- Re: config.cache considered harmful Martin Buchholz
- Re: config.cache considered harmful Harlan Stenn
- Re: config.cache considered harmful Pavel Roskin
- Re: config.cache considered harmful Akim Demaille
- Re: config.cache considered harmful Greg A. Woods
- Re: config.cache considered harmful Tom Tromey
- Re: config.cache considered harmful Greg A. Woods
- Re: config.cache considered harmful Akim Demaille
- Re: config.cache considered harmful Martin Buchholz
- Re: config.cache considered harmful Pavel Roskin
- Re: config.cache considered harmful Akim Demaille
- Re: config.cache considered harmful Martin Buchholz