Re: [Fink-devel] strict-dep checker script
At 20:10 Uhr -0800 24.03.2002, Kaben wrote: Re: interactive packages such as Qt: free/netbsd ports/pkgsrc have a mechanism for handling this. `interactive' patches have variable `IS_INTERACTIVE' set true; build mode can be unset in which case everything is built, `BATCH' in which case interactive packages are skipped, or `INTERACTIVE' in which case only interactive packages are built. a useful feature fink could swipe. Re: alternate dependencies: would it be easy to use all combinations of dependencies if desired? combinatorially, I mean, via graph traversal. you would get more thorough testing that way... A depends on B or B'; B depends on C or C'; B' depends on D; build C-B-A; start over, build C'-B-A; start over, build D-B'-A. The problem with this approach is that it exponentiates the required build time, i.e. I think it just won't be feasible if we have many packages with alternate dependencies, given the time available in the avarage lifespan of a human being. Right now it still might be possible, though, since few packages use this. Also, it wouldn't work well for another reason: the system-* packages. But I plan to exclude them from testing from the start. It would be nice to test them, but that would mean I'd now also have to repeated (de)install 3rd party applications... I have to draw the line at some point, after all, I have a real life, too. Max -- --- Max Horn Software Developer email: mailto:[EMAIL PROTECTED] phone: (+49) 6151-494890 ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] strict-dep checker script
At 13:35 Uhr +0900 25.03.2002, Peter O'Gorman wrote: I have one suggestion for this. When you have built everything with nothing installed, you should build again with everything installed. Building twice like this will find both missing Depends and Conflicts... I think it is a great idea. The idea is great, but sadly, reality is a bit different. For example, how do you define everything installed ? In fact, it is techincally impossible to install every Fink package at once, since some conflict. Don't even suggest to work through all possible combinations, as I said already in my previous mail to Kaben: we have 1 week for this, not an entier life time. Max -- --- Max Horn Software Developer email: mailto:[EMAIL PROTECTED] phone: (+49) 6151-494890 ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] strict-dep checker script
On Monday, March 25, 2002, at 08:14 PM, Max Horn wrote: At 13:35 Uhr +0900 25.03.2002, Peter O'Gorman wrote: I have one suggestion for this. When you have built everything with nothing installed, you should build again with everything installed. Building twice like this will find both missing Depends and Conflicts... I think it is a great idea. The idea is great, but sadly, reality is a bit different. For example, how do you define everything installed ? In fact, it is techincally impossible to install every Fink package at once, since some conflict. Don't even suggest to work through all possible combinations, as I said already in my previous mail to Kaben: we have 1 week for this, not an entier life time. Good point, I still think a reasonable set of packages could be made that would deal with your technical impossibility, but I don't have a solution for the build times... Wait, why don't you use all the money that fink has generated in sales to buy a couple of high-end Macs? :) Peter ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] strict-dep checker script
Re: interactive packages such as Qt: free/netbsd ports/pkgsrc have a mechanism for handling this. `interactive' patches have variable `IS_INTERACTIVE' set true; build mode can be unset in which case everything is built, `BATCH' in which case interactive packages are skipped, or `INTERACTIVE' in which case only interactive packages are built. a useful feature fink could swipe. Re: alternate dependencies: would it be easy to use all combinations of dependencies if desired? combinatorially, I mean, via graph traversal. you would get more thorough testing that way... A depends on B or B'; B depends on C or C'; B' depends on D; build C-B-A; start over, build C'-B-A; start over, build D-B'-A. On Mon, 25 Mar 2002, Max Horn wrote: . . Of course some stuff can't be tested easily with this. For example anything requiring user interaction, however, that is a bad thing for Fink packages normally. QT requires it (to confirm the license; is this really necessary?). And Atlas requires it too (makes sense in this case, though). . . Finally, I am not sure yet how to handle alternate dependencies, e.g. (emacs|xemacs) - maybe I'll just use the first, or whatever. Merry Christmas -- K ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] strict-dep checker script
I have one suggestion for this. When you have built everything with nothing installed, you should build again with everything installed. Building twice like this will find both missing Depends and Conflicts... I think it is a great idea. Peter On Monday, March 25, 2002, at 10:20 AM, Max Horn wrote: For the stable bindist, I plan to write a special batch mode strict dependency checker, and I'd love to hear your comments on this idea, and possibly suggestions on how to improve upon it. ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
