Ken Moffat wrote:
> > Hi Torsten, > > I think you're overlooking a couple of things Sad but true most times. Well, it's always worth a try :-) > - editors upgrade packages and do any testing them with the current > book. Nobody, AFAIK, is testing package updates against the last stable > book, other than to fix identified problems and vulnerabilities. Even > then, changes might still introduce unexpected results. This rises the question: How much are the packages tested the errata page advises to use? > - the svn book is by definition unstable. Not questioning that! That's why I only use it on my spare pc to keep up with development. > Sometimes, it can take days, > or weeks, for the problem to show up (e.g. past issues with bison and > strip) and the fix might be to revert the version. > Well, point taken, but one last try: I think there should be some explaining on the to-be updates page regarding this issue so that people could decide wether to stick with the plain stable releases or to go with the updates. Anyway (and I know this is wishfull thinking) the version numbering of packages should hint toward wether there was a _major_ change in code/API or a _minor_ fix/improvement. Also the changelog should indicate if there is trouble on the horizon. This could be an argument in favor or against my scheme so maybe I have to stick with updating to new packages and hoping they work as proposed. Torsten. -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
