Closed sub205 closed 5 years ago
Hi @sub205 I don't want to be prescriptive, I don't know the context of your submission and might I suggest that you squash the multiple commits to the same file? In essence a single commit should encompass a single functional change. The idea behind this, although not particularly relevant for you, is that should a problem occur with a change somewhere, it's less revisions to bisect, test and isolate the breaking change. So for your case your textual changes could be a single commit, and your add image might be another. Hope it helps, Andrew
Oh I forgot one thing. If you commit to local branch other than master it's easy to do
git checkout master
git pull upstream master
git checkout your_work_branch
git rebase master
so that you don't have to keep merging in the current upstream which confuses what exactly your changes are.
great! Guys, this is exactly how we should treat eachother!
@andrewbird Thanks for the nice infos. I'm quite new to collaborative working, i only learned that it's better to fork and merge back than working directly in master ;-) I'll promise to do it better next time!
@roelandjansen In germany we have a saying ... it's like "the sound makes the music". When someone is acting like the biggest teacher and between every sentence is something like "you have no clue, i'm the best", that isn't really helping your contribution and mindset.
Nobody is perfect!
another question arises: Which is the preferred build environment for us? It's quite easy for me to do a packaging-script in linux, but wouldn't help much under DOS.
Especially mounting floppy images.
Dosbox offers a way to do this and may be the best solution?
@sub205 yes it can be difficult transition to collaborative working, but you shouldn't worry too much about format etc. Some projects have strict guidelines about submissions, here no guidelines have been set but in general limiting what a potential reviewer has to do will ease patch submission. I myself use Linux exclusively but understand that others may not, particularly as this project is DOS oriented, so I tend to code for the DOS solution, whilst quietly ensuring that it works on Linux.
my base is linux too and for testing I use vmware.
My primary development platform is DosBox on Linux. This allows me to use all of my Linux tools on the files while still in my DosBox session as well as direct access to git.
I use VirtualBox for my actual "running" tests with both DOS 5 and MOS. Now that the date bug is fixed, I don't really have much need for DOS anymore, but I still have it around if I need it. I am able to run the required tools on MOS without much headache.
Once I have "something" I feel is ready for an actual test, I have a MOS "server" that has a couple built in serial ports, a Pentium 4 MOS laptop, and an old Compaq 486 laptop to run tests on.
So it looks like we have quite a broad set of platforms represented as dev/test environments!
Here are fixed v5.01 images and an empty bootable floppy for further automation of image creation, which i'll try to do next.