Am 20.11.2012 20:12, schrieb Josselin Mouette:
The MATE-maintainers (especially including upstream team, which is growing) said explicitly they are willing to maintain the legacy code as long as the replacement is not there. As bad as the bonobo code may be, it was in Debian for years and we all know that is not the only code in Debian which at least some Debian maintainers / developers think is to the utmost extent horrible and it is nevertheless in debian. At least I can't comment on this, as I am not a coder, not a maintainer nor a developer, I just have open eyes and ears. I am pretty sure your assessment "they don't have the expertise" would fit in e.g. your opinion to many other maintainers. Same goes probably for other maintainers and developers as well.Le mardi 20 novembre 2012 à 18:37 +0100, Michael Schmitt a écrit :One of the biggest problems with MATE so far is that it currently depends on libraries and other things that have been deprecated and removed in Debian and should not be reintroduced to Debian.Can we agree on something more friendly like "it would be not optimal to reintroduce these packages"? Or at least elaborate a bit why they should not be re-introduced. It would not be the first time something like that happened in Debian.Because nobody knows anymore how to maintain libraries as complex as bonobo, for example. And I’m pretty sure the MATE developers don’t have the expertise.
I noticed that, they noticed and acknowledged that, they simply did not know better at that time. Such things happen, even in Debian, but basically everywhere. Human nature. We can't know everything and what one may consider to be important, others may find silly. For example keeping track of a gnome component after the Gnome-devs did go on the (for the MATE devs) horrible and insane gnome3-trek may not seem to be so urgent, especially in the beginning of the fork. No problem for me understanding how something like that could slip through. Now they know that, now they even keep track on other components as well (including finding bugs in current gnome3-code), backport stuff from various gnome3-components and elsewhere.Even worse, they forked stuff like GConf using sed to rename it mateconf, while GConf is still available in a much more modern version in wheezy, 100% binary-compatible while being ported to D-Bus.
Wow, that's kinda harsh. :( But then again, I read statements like that more often than I would like to in almost any technical oriented place, especially on debian IRC-Channels from other maintainers about code or other Debian coders in general. I can't comment on that either (me = no clue about any programming language) much, just that I tend to think "The only code that is ok is the code that I wrote" that kind of attitude may apply to many coders... Let's ignore that imho rude assessment and consider that no one is born as a master in anything. And, let's think about other distros that already include MATE or will include it in their next release. So, Debian would not be the only one. I guess if a serious (security?) bug would be found in the code, those combined forces should be able to fix it.I don’t think we should allow libraries from clueless developers to be introduced into the archive.
I don't know what I should make of that. I am not part of the actual forking / maintaining team, I am just a user. So don't confuse me with them. Actually they helped me to understand the issue better. An issue not at all limited to MATE. Plain fdo issue (I noticed it first with Gnome + KDE or Xfce), Now I think about digging deeper to be able to write a high-quality wishlist bugreport for fdo. No worries, I wouldn't even think of pestering the debian BTS with that, I'll go straight upstream. ;) And... you do know about what issue I am talking here exactly? Or was that just some rude comment my parser did not parse right? ;)Furthermore, work has to be done to make sure MATE co-installs seamlessly with anything from GNOME which isn't that trivial.Apparently it is easy enough, as they got that already done months ago. At least I run MATE on many sid-boxes with Gnome3 co-installed and no issues apart from that damn mime-types-thingy brokenness when one has more than one DE installed.Maybe you can try to understand how XDG mime works. But maybe this is too much to ask to people who forked libraries that are still available.
In general, I have no real idea how the common Debian users will react to wheezy, but I am sure, with MATE, a drop-in-and-feature-complete replacement for the most used desktop environment in squeeze it will be at least on the desktop-front a pleased reaction. In its current state, from a technical point of view it might not be ideally (but I am sure there are even more not-ideal packages in Debian right now), but not THAT bad after all. And from a political / social point of view, just the only sane option to do (if we have the users in mind).
If one likes Gnome2.x or MATE or not is a question of taste, to acknowledge that many users are just mad about Gnome3 is a fact. To offer no real sane upgrade-path for those users is... dunno how to say it in another way, it is just insane! I did invest some time with thinking about all of this including talking to some MATE folks and current squeeze-gnome-users, and again thinking and thinking... imho it must be done everything possible to try to circumvent that imho probable desaster for the desktop users. Especially when I think of the real *average next door users* not the techie guys who may be happy (or at least not cry to loud about it) to invest some time tweaking their Xfce or KDE to behave more or less like their old gnome2 did. And if you really insist, I can give you detailed reports about what kind of problems those users have with the current DEs in Debian including gnome3, but I don't want to bloat this mail anymore. I just can guarantee you now, the panel in gnome3-fallback being black is not one of them. ;)
From my point of view, as said I can't comment much on the technical aspects here. I can just ask those who may have the ability to do so, to do it in a fair way, with as less prejudice as possible. Again, for the users that will have an issue with the current envisioned upgrade-path. For those users not so keen about adding a third-party repo to their sources.list, those users not knowing about the possibility to do so, those users that will get pissed.
regards MichaelP.S.: I know there is already work in progress for MATE for jessie, I just think that is too late. But I don't want to trip on anyones toes here....