Re: [PATCH] cvsimport: rewrite to use cvsps 3.x to fix major bugs

Junio C Hamano <gitster@pobox.com>:

Er. You still don't seem to grasp the fundamentals of this situation. I'm not imposing any damn thing on the users. What's imposing is the fact that cvsps-2.x and the Perl cvsimport are both individually and collectively *broken right now*, and within a few months the Perl git-cvsimport is going to cease even pretending to work. I'm trying to *fix that problem* as best I can, fixing it required two radical rewrites, and criticizing me for not emulating every last detail and misfeature immediately is every bit as pointless and annoying as arguing about the fabric on the deck chairs while the ship is sinking.

To put it bluntly, you should be grateful to be getting back any functionality at all - because the alternative is that the Perl git-cvsimport will hang out in your tree as a dead piece of cruft. Your choice is between making it easy for me replace it with minimum disruption now and hoping for someone else to replace it months from now after you've had a bunch of unhappy users bitching at you.

So let me be more direct. I think the -M and -m options are sufficiently bad ideas that I am *not willing* to put in the quite large amount of effort that would be required to implement them in cvsps or parsecvs. That would be a bad use of my time.

This is not the case with -o; that might be a good idea if I understood it. This is also not like the 2.x fallback; I thought that was a bad idea (because it would be better for users that the combination break in an obvious way than continue breaking in a silent one), but it was a small enough effort that I was willing to do it anyway to keep the git maintainer happy. The effort to fix the quoting bugs is even easier for me to justify; they are actual bugs.

Those are my engineering judgments; go ahead and call them "subjective" if you like, but neither the facts nor my judgment will change on that account.

OK, thank you. In the future I will try to bear in mind that English is not your primary language when I evaluate statements that seems a bit offensive.

So what's your next bid? Note that you can't increase my friction and hassle costs much more before I give up and let you deal with the consequences without me. I want to do the right thing, but I have more other projects clamoring for my attention than you could easily guess. I need to get git-cvsimport *finished* and off my hands - I may already have given it more time than I really should have.

So give me your minimum list of deliverables before you'll merge, please, and then stick to it. I assume fixes for the quoting bugs will be on that list.