On Friday 29 January 2010 18:01:30 Daniel Johnson wrote: > The only dependency I'd hope for would be sh. (and git) My belief is that > if these scripts are kept simple, and basically never change, they won't > be "prone to error". The whole point of being very light wrapper around > git commands is that if the git commands fail, you can see and fix it.
One thing I can't figure out is why people are talking out their ass about kdesvn-build (and to be clear this is not just you). Its "dependencies" are Perl (instead of sh) and $scm (either svn or git) Beyond that kdesvn-build simply does the same exact steps that I'd expect you to have to do manually, and even parrots those steps to you (i.e. in the log files and the --pretend output). The logging is designed to be almost completely foolproof to access (log/latest/$module/error.log always gives you the output of the failed step). If you prefer to see all output then either use --debug and accept super- verbosity or file a feature request since no one has complained about this to me. Although kdesvn-build normally is used with a config file, it is not required, and will still try to do something intelligent even without. (For example, see the output of "kdesvn-build --pretend --rc-file /dev/null qt-copy" to simulate no config file) Let's say you've already checked out qt-copy from your personal repository which you've added as a remote. If you configure qt-copy to be updated from that repository then kdesvn-build will even pick up and use the remote you've added instead of trying to make one of its own (likewise for remote tracking branches). This is in order to operate just like the user would be doing and keep things as close as possible to the output of a user doing it manually. So instead of spouting crap could you instead clarify what exactly kdesvn- build is doing to "take you farther from the software"? kdesvn-build is *NOT* an Enterprisey Perl-Moose-Catalyst-DBus-daemonizing monstrosity. It does not use Perl bindings to either git or svn. It is a single script file, which does not require installation and is upgraded simply by overwriting it. (It *is* about 6.7KLOC of code but since I'm not sure what exactly you're complaining about I'm not sure exactly how to respond). All it does is run down a list of modules provided and clone/updates them one by one and... oh, wait, that's what we're talking about. So sorry to come off as some kind of raving lunatic but I'm seriously not sure where all the misunderstanding/freaking-out is coming from regarding kdesvn- build. Is it just too old (and therefore busted)? Do I need to hurry up and change the name before people will believe it handles git modules? Is it because I bothered to try and document it fully? Is it Perl? (then just use Michael Jansen's build tool) Or is this just the next IRC client, and everyone is going to make their own "build them all" script? Regards, - Michael Pyne
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Kde-scm-interest mailing list [email protected] https://mail.kde.org/mailman/listinfo/kde-scm-interest
