Eric Auer schrieb: > Hi Michael! > >>> thanks for suggesting an automated compile-and-test-SVN farm >>> of test computers, I guess part of the problem is that you >>> typically compile DOS in DOS and DOS does not run cronjobs. >>> Conceivably a smart Linux DOSEMU solution is possible but in >>> that case the test will not be on raw hardware. Other ideas? > >> - Then a cronjob (TSR) is one of the first things you need. > > You will also need a SVN client for DOS :-). Blair? > > Another thing will be a list of things to test. I am > sure you want more than "fdapm warmboot and see if it > still boots", right? There are so many things you can > do with a DOS kernel, where would you start testing?
Well, things like xmstest, memtest, idecheck, nssi, arachne, ems test, alloc test, dosidle, slowdown, etc. etc. just testing if all starts and ends correctly. Probable such a test would need to include anything. Often I tested some nice TSR (like a clipboard or clock tsr etc.) or something and it worked well for many things but in the end there was something no longer working or unstable. >> The coreboot automated testsystem needs also special >> and expensive hardware for automatic flashing and such. > > I am sure that would be possible with self-made hardware. And I am sure it's for this project pretty unrealistic. >>> www.coreboot.org/Distributed_and_Automated_Testsystem > >> But what to do if you have no more then one person or just a few people? > > Then you cannot call the software stable. You can still > announce it, of course, as long as people know that it > is untested and the code is not reviewed by anybody else > than the one who wrote it... Well, then the next distro must be an unstable/beta distro. >> Look at the changes in cvs, release dates, mailing list >> and download statistics. > > I wonder what the download starts for Rugxulo and other > more up to date distros are ;-). Depends on how well it can be googeld or on how good it is linked. >> Linux is confusing, big parts of the community are elitist > > Maybe, but reading the DOS networking howto by Uli does > not at all give me the impression that DOS networking > would be easy. Sure, running EASY apps in DOS IS easy! DOS networking is indeed hard but normal internet applications with packet driver just luck documentation. >> DOS has other sympathetic strikes, DOS just got boring over >> time and could not comply with the new requirements. > > Which requirements? Probably a matter of taste, but still > an interesting topic to discuss :-). Requirements with new features, pretty guy, easiness and clickmotley. >>> can for example offer more / more open drivers now. > >> Which more open drivers? > > In particular the licenses of dosusb, uide and xgcdrom > make life easier than having to download proprietary > USB and SATA drivers... dosusb doesn't allow commercial use this would mean commercial users wouldn't be allowed to download the cd any further? or them are just not allowed to start that driver? > And there is a port of clamav > which makes downloading f-prot unnecessary. It seems > that not much commercial software remains in use :-) Clamav is not yet better then commercial antivirus in virus findings. -mr ------------------------------------------------------------------------------ _______________________________________________ Freedos-user mailing list Freedos-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-user
