On 4 October 2013 02:35, Nick Treleaven <nick.treleaven@btinternet.com> wrote:
On 03/10/2013 15:37, Lex Trotman wrote:
> >- holding us at an unreasonably old GTK version
>Sorry?  I don't think Windows support has anything to do with the GTK
>version we support.  If anything, it's rather the contrary, providing a
>newer version for Windows installers is far easier than forcing users of
>old GNU/Linux distros to build a new version.
Well, as I understand it, moving the windows packages to GTK 2.24 is the
only thing holding us at 2.16.  As there was a deathly hush on the ML
thread when volunteers were requested to try it, I guess nobody is willing
to support it, even just for that.

I'm not sure that was the main reason, I argued that LTS distros was a reason for keeping older dependencies. I expect there are newer runtimes, I heard gtk 3 on Windows works reasonably well since last year or so.

IIRC the LTS distros were 2.18 at least. GTK3 Geany for Linux is currently quite stable, I use it every day.  GTK3 on windows could be ok, but again we come to the point that nobody is going to do the work of testing it and then changing the release and nightly builder.  

Since the only way Windows is "stopping" any upgrades of the oldest version is the builders using 2.16, maybe stopping windows support is just no longer making those packages, leaving the code inside #ifdef OS_WIN32 there for you to build it with your preferred GTK.


In any case, if we want to up the dependencies, don't deliberately drop Windows support.

> >- hard to maintain due to being hacks on top of hacks
>Again, spawning code (only?)
Again 10% of open bugs.

Just because there are unresolved bugs doesn't mean we should drop a platform.

User: 'Your software doesn't work right'
Dev: 'Forget it, we're going to stop releases for your platform'

Only if there is nobody to actually solve the problems.  And I said "a couple of releases", at our speed thats at least a year.  If we can't find some additions to your efforts by then I would wonder how interested the users are.  Geany is an IDE!!!  A far greater proportion of its users are going to be doing some sort of programming than would be the case for a project like Inkscape for example.  But if we never start the process we won't ever get anywhere.

> >- few of the developers have access to a representative development
> >setup (no WinXP on a VM is hardly representative)
>I indeed don't, but AFAIK Nick and Matthew does
And I should add, its not reasonable if you are the only one either, the
fixes needed to windows will deprive the rest of Geany of any of your
effort.  Its not as if there are no other things to do:)

A slightly broken Windows version is infinitely better than none at all. I use it and don't experience any significant issues that can't be worked around.

But thats not a universal experience, and again there is nobody who is looking at why and how to fix it.  At least the Linux version gets a slow trickle of changes.


>Anyway, although I agree the Windows version isn't as good as it should
>right now, I don't really see the problem it causes to the rest of
>Geany.  And if we indeed clean some things up a little like Matthew
>suggests, it could even be just fine (Windows dialogs, come on).

I don't mind if Matt wants to remove the Windows dialogs. Enrico put them in, I think they're not really necessary.

Again "somebody else do it", Matt having already said he won't.  :)

Again, the issue is, without anybody to do it, all that happens is that the
rest of Geany gets less love.

Lots of people over the years have put forward suggestions, but none of
them have even dropped a PR or patch, let alone assisted with maintenance.
  Its not nice to cut off a platform, but if everybody expects others to do
the work, then we may have no choice or the whole project will collapse.

^^^ *Massive* exaggeration. Most open source projects have unresolved bugs. Unfunded open source projects work mostly because developers fix problems that they themselves experience. I don't see any significant problems with the Windows version that affect me.

And you do work on the problems that affect you, and thank you for that, but "support" is also answering bug reports and questions on IRC, otherwise you get the current situation of Matt and I blundering about in the dark (well Matt might have a small torch, but I'm in the dark).  I'm not criticising you or suggesting you take over the role again, I'm looking for people to help you.


I'm not saying delete anything inside #ifdef OS_WIN32 tomorrow.  But also
in fairness some warning that platform specific code is not being
maintained adequately needs to be given to Windows users, and their help
sought, and this is one way I can think of that makes it clear.  Other
suggestions are welcome:)

You are massively overreacting. If we really cared so much about user experience we would make more bugfix releases, but we don't have the resources for that.

Indeed, and that is the problem I'm trying to address.  If we don't make it clear to people that they need to step up and help, we will get nowhere.


Devel mailing list