FWIW: I'm away soon for a few days, so you'll have to continue this without me (a fact which is no doubt to your infinite relief :-).
On October 10, 2003 11:27 am, Dimitrie O. Paun wrote:
On Fri, 10 Oct 2003, Geoff Thorpe wrote:
[snip]
I haven't dispatched you to the archives to be rude, but to avoid rehashing the same arguments over and over again. Saving time is important, at least in my books, as I have a chronic lack of it :) My point is that most of the people here, in particular Alexandre, prefers inlined patches for many reasons, some of which I've outlined already.
Sure, I understand that.
I really don't understand why we have this entire argument in the first place. One thing that is certain is that one wine-patches we would like *ideally* to receive patches inlined. As you say, the point of having such a filter is to accomodate as many different "styles" of submission, and we do that by translating all those style (if possible) to the prefered style, and that one is inlined.
I would correct that to "and we do that by translating all those styles (if possible) to a prefered style, and that one is what each reader finds easiest to use."
It's perfectly clear that some people with certain mail setups prefer inlined, and some prefer attached. Any argument with that? Insignificantly perhaps, I prefer attachments because they are decidedly easier for me to use; I can file-browse into my CVS repository, drag-n-drop the diff, and get a multi-panelled color dialog showing me the before-and-after picture. I can't do that with the original email itself, nor can I browse the diff itself with syntax highlighting without first manually extracting the patch from the email. If only one list-server output is possible, inlined patches will be the norm. If more than one is possible, wine-patches will be a better service for more people. Feel free to disagree with that logic, but it seems pretty clear to me. BTW: if delivery of wine-patches to your own address generated inlined patches anyway, would you object on any non-philosophical grounds to *submitting* your patches as attachments anyway? You are expressing a preference after all for how you prefer to receive patches, so would you similarly accommodate other people's tastes if it made no difference to you? No matter what you're receiving/processing list mail with, I think sending with attachments is universally pretty easy, and as I say it is very difficult to extract patches from an email except by passing the raw mail data directly through 'patch', which you may not want or be able to do in some setups.
So we have two orthogonal things here:
- Should we have such a filter?
- Should we dissus the prefered style?
s/preferred/default/
I don't think I'd like to reopen the n-th time the discussion about (2), but if you feel you have good arguments...
Only that there are indisputably situations where attachments are easier and more flexible to use, though they happen not to be your situation nor Alexandre's (using your current setups). The question is; is wine-patches a service *to* as many people as possible, or is it a sevice for a couple of core developers *from* as many people as possible? That's not rhetorical, I think it's an open question.
Open source increasingly dissolves the traditional roles of developers and users and, as far as wine is concerned, I am one of these new byproducts: a "participant". Do I help wine or does wine help me? The answer: neither and both. My preferences count, but to who? and how?
BTW, the purpuse of wine-patches is twofold:
- For Alexandre to apply a patch. He has spoken before, and he prefers inlined patches; (please read the section on style http://www.winehq.org/site/docs/wine-devel/style-notes )
I think we're agreed that one of the major roles of such a filter would be to produce more consistent output, and being able to consistently inline patches would clearly be one of the most valuable features for people like yourself and Alexandre. No argument here.
- For others to review patches, and again, inlined patches are better.
For some, this is indeed true. But could I have them as attachments please? :-)
The case of multiple patch submission is a red herring. In the 7 years I've been with the project, I haven't seen a single case of someone submitting alternate patches. Yes, there were people submitting mutliple (separate) patches as one email, but this is strongly discouraged, as I've already said. And even if that happens, just letting them be would nicely take care of things.
Reality check: dereferencing a NULL pointer is "strongly discouraged", receiving multiple patches in an email is "marginally annoying and should have a damned good reason". C'mon, this is open source, *nix, and the uber-world - when did we all suddently turn into puritans and forget that we can be flexible precisely because we control our systems and tools and have a choice of what we use and how we prefer to operate? Concrete rules and "don't touch" inflexibility, that sounds a lot like Microsoft.
Cheers, Geoff