Re: [Fink-devel] unstable and wrong links
hi alex I forgot all about dryrun. This sounds like a great thing for somebody to run every so often and post the bad connections to the maintainers. wow, just fyi. the situation is not that bad. i have run a dryrun and tested most of the urls and there are about 90 out of 3000 packages which results in a 404 and about 25 with a 503. there also some other error messages like 30x and other 50x, but in total these are only 180 or so. seems to be a fulltime job to inform the maintainers of the packages. :) anyway, you are doing a great job, gottfried --- The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA. http://www.eclipsecon.org/osdn ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
[Fink-devel] unstable and wrong links
hi girls and guys, maybe this question has been already answered in the past. but i cannot find a definite answer for this in the archives or in the FAQ. i have already tried the user-list, but with no success til now. since i am using unstable i have always problems with the downloads. i mean that the download server has version X and the info file shows version Y, where X Y. so the download obviously does not work. ok, the FAQ recommends to search for the package and to place is it into the /sw/src/ directory. sure this works, when you are happy to find the correct version of the file. but there are some drawbacks like: - it is not working all the time (sometimes i cannot find the right version) - it is boring, to search for so many packages in the net. - ease of use, because i want to start the update, go for a beer and come back in the morning, when everything is done. i understand that unstable means unstable and i expect some problems. why not, maintaing the lot of packages is really tough. isnt there a way of automatic link checking and checking the consistence of the info-files with the download locations? is there such a tool? if the problems could be found in advance i think this would also improve the confidence into fink - even for unexperienced users. is there an automatic testing of the info files? mabye this is the wrong location for the post, but any comments and corrections are highly appreciated. gottfried --- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] unstable and wrong links
hi alex so, for me this is quite fine, because i believe to know what was happening and i solved the problem for me. for unexperienced user this could a reason to not to use fink any more, because it could be frustrating. sure, noone forces them to use unstable, but who is not using the latest version of the software on a desktop system? It's possible, though it seems like most people post to the mailing list--which means that the problem gets known. yep, thats correct. but this requires the user either to subscribe to a mailing list or to register at SF to report a bug. both a really huge burden for average mac os users. i think that they are a little bit lazy. :) and even in the case that the user is using stable, there is same problem (you have mentioned it at the beginning of the mail). this means that fink and the packagers relies on the feedback of users, if a package does no longer exist (huh, user feed back, is this really working?). i mean, one way to improve the confidence of normal users into fink is to ensure that at least the stable tree is consistent. Every so often somebody does a check by running fink fetch-all, which tries to download every package. This could be done more regularly--it's tedious, though, because I think there's about 1000 packages in the stable tree, and twice that in unstable. i have tried a different approach which does not download all the packages, because just checking for the existence of the file on the mirror is enough. i created a list of downloadable files with the help of fetch-all and the dryrun option which prints a list of urls. i always took the first url in the list (dryrun prints the name of the package, checksum and a list of download locations) and checked with a HEAD command (curl supports this, option -I ) and in combination with a proxy the existence for file. this brought up some non-working locations (503, 404, and time outs occured). so, this produces a not so high traffic (eg 5kb/package, which means about 15meg for unstable with about 3000 packages) and also, sorting the files by server allows curl to combine requests to the same server and reduces the cost of connection-setup. cu, gottfried --- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
