Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
games-strategy/0ad
View unanswered posts
View posts from last 24 hours

 
Reply to topic    Gentoo Forums Forum Index Gamers & Players
View previous topic :: View next topic  

Are you interested in support for the 0ad game?
Yes
90%
 90%  [ 10 ]
No
9%
 9%  [ 1 ]
Total Votes : 11

Author Message
Navar
Guru
Guru


Joined: 20 Aug 2012
Posts: 353

PostPosted: Mon Apr 28, 2014 4:12 am    Post subject: games-strategy/0ad Reply with quote

EDIT: ebuild here

Hey folks. I had this excellent pre-release quality real time strategy game, 0ad installed from someone's overlay (I think it may have been hasufell but no longer certain versus 4-5 outdated others floating out there). Today after needing rebuilds due to lib(png/jpeg) updates, things broke due to (positive) changes in the git repo.

I see no productive change or evidence in sunrise overlay (it seems non-existent) for overall being current and functioning since this bug.

Anyway, I jumped in to try and resolve this a few hours ago. Would be willing to be a proxy maintainer via gentoo-sunrise overlay if a dev is willing to work with me. I did inquire for interest of same on IRC tonight but ran into crickets. Current build works with all system libraries instead of bundled with local USE options added to support. The Gentoo way.

I will be posting an amd64/x86 keyworded ebuild that is current, makes repoman happy and so on with a link here as soon as I make a follow-up post in unsupported software (the 0ad-99999.ebuild, Manifest and metadata.xml that you can, for now, copy into a local overlay). I've not spent the time to be an ebuild expert, so go easy. It'd be nice to see Gentoo showing more functional and up-to-date on the support page from upstream given the efforts they seem to have done for us.

Hopefully this will be of interest to others. If you enjoyed Age of Empires, etc. this is a very similar, modern and high quality open source game.
Back to top
View user's profile Send private message
hasufell
Retired Dev
Retired Dev


Joined: 29 Oct 2011
Posts: 429

PostPosted: Tue Apr 29, 2014 10:24 am    Post subject: Reply with quote

Not sure if I understand that reasoning. What is wrong with the current 0ad ebuild (which is current). We will import it into the tree as soon as it reaches beta status.

Spreading ebuilds across multiple overlays is not a great idea.
Back to top
View user's profile Send private message
Navar
Guru
Guru


Joined: 20 Aug 2012
Posts: 353

PostPosted: Wed Apr 30, 2014 3:50 am    Post subject: Reply with quote

I was unclear? Those are 4 months old and requirements have changed (as in build failure, go look, have you tried?). They're not the only 4 month old ebuild variant floating out there. They all had issues. From the bug report and your responses there, it seemed more or less an attitude of abandonment (or your own personal overlay, etc.).

As far as spreading to multiple overlays, where are you saying I've done that? This was in response to working on getting functionality for the current build from upstream into sunrise (or main, whatever).
Back to top
View user's profile Send private message
hasufell
Retired Dev
Retired Dev


Joined: 29 Oct 2011
Posts: 429

PostPosted: Wed Apr 30, 2014 12:29 pm    Post subject: Reply with quote

Navar wrote:
Those are 4 months old

0.0.15 is the most recent official release, see http://releases.wildfiregames.com/ and http://trac.wildfiregames.com/roadmap
If you want to update the live ebuild, send a pull request (and btw.: if you know about multiplayer games and desyncs, then you know that running live ebuilds for multiplayer is a bad idea).

Navar wrote:
and requirements have changed (as in build failure, go look, have you tried?).

Yes, 0.0.15 builds fine, I just checked. I don't build the live ebuilds every day, so again: send a pull request for those.

Navar wrote:
They all had issues.

Which issues do my ebuilds have? This sounds too vague.

Navar wrote:
From the bug report and your responses there, it seemed more or less an attitude of abandonment (or your own personal overlay, etc.).

Again: We will import it once it reaches beta status. I use my own overlay for general alpha status ebuilds. I tried to establish an official "games overlay" some time ago, but there was not much interest. In the end, it's just a name. Workflow in sunrise is severely broken and there is no regular games ebuilds contributor (you want to wait 2+ months for a bugfix to get into the main sunrise repo?). Gamerlay is totally unreviewed, so that idea is even worse.

Navar wrote:
As far as spreading to multiple overlays, where are you saying I've done that?

I didn't.

Navar wrote:
This was in response to working on getting functionality for the current build from upstream into sunrise (or main, whatever).

Which functionality? Send a pull request.
Back to top
View user's profile Send private message
hasufell
Retired Dev
Retired Dev


Joined: 29 Oct 2011
Posts: 429

PostPosted: Wed May 21, 2014 6:14 pm    Post subject: Reply with quote

0ad is now in the tree. If you previously had any overlay ebuild, you should re-emerge 0ad and make sure that it doesn't get overwritten by an overlay ebuild, since the versioning has changed (latest version is 0.0.16 now).

so run:
Code:
emerge -av 0ad::gentoo
Back to top
View user's profile Send private message
hasufell
Retired Dev
Retired Dev


Joined: 29 Oct 2011
Posts: 429

PostPosted: Wed May 21, 2014 6:16 pm    Post subject: Reply with quote

if any1 feels like it... you can add a 0ad item here: https://wiki.gentoo.org/wiki/Games/strategy
Back to top
View user's profile Send private message
Navar
Guru
Guru


Joined: 20 Aug 2012
Posts: 353

PostPosted: Tue Jun 10, 2014 5:36 am    Post subject: Reply with quote

I finally have time to respond.

I understand your points on why it was not in the main tree (which is why there are overlays). If anything, I'm surprised at your sudden inclusion.

I'll stop from updating my posted ebuild (in unsupported software subforum) due to lack of interest as seen in the poll and the fact that you've taken on maintainership of a release into the tree. Since it was not clear to you, my interest was in the goals as stated in the apparently deprecated Sunrise project, including broken links.

After doing the work to resolve a functional build, I thought it would be useful to contribute given the prior state of 0ad to other interested users. Sitting in IRC channels for over 2 hours with zero response from anyone was definitely a waste of time. What I contributed here I did out of goodwill. Posting with explanation here was easy and took seconds from what had already been done.

hasufell wrote:
If you want to update the live ebuild, send a pull request

As in yours? I do not have an account on Github. Nor was it at all clear from the bug report that you were maintainer. Referencing you was in trying to give credit on which ebuild author I had used as a base point. Instead I would have submitted a patch to the bug, as others had in past, if it still seemed relevant to do so.

However, I have went through the effort of submitting patches before only to watch, e.g., as a configure test still asserted wrongly after a vaguely ignored patch submission. (not that it matters, ultimately, since upstream head already had mailing list attitude of sse3+ only soon simply because they couldn't be bothered to check and fix their tests from a few years prior when they did care)

hasufell wrote:
if you know about multiplayer games and desyncs, then you know that running live ebuilds for multiplayer is a bad idea

Checksums and other methods have been used against that for a long time. I've done multiplayer gaming off and on for more than 20 years, but not via Gentoo ebuilds or Linux in general. Though I do not know the intricate details of what the 0ad developer is doing in that regard. Regardless, like you stated, it's still considered alpha status upstream. I view that as, end user--lower your expectations and expect issues aplenty, trunk checkout or packaged.

hasufell wrote:
I don't build the live ebuilds every day [...] Which issues do my ebuilds have? This sounds too vague.

I was referring only to the live ebuild which I thought was obvious. At the time upstream had already done a feature freeze (leading to your release version shortly after). Most seen release ebuilds in the wild were in the sub alpha14 range (outdated).

A diff versus your live ebuild of 7 months ago would be an example, but your changelog and changes you've done on your release vs live indicate enough. I don't believe there's any reason to re-hash this other than your confusion on what I meant which was simply that you were one of at least 5+ authored versions floating out there (in addition to submissions to the bug report) which would no longer build. Simple as that, no prejudice in regards to you or any interest in seeking your attention for that matter, unless it was within the interests of the goals mentioned in Sunrise for proxy maintain.

hasufell wrote:
Spreading ebuilds across multiple overlays is not a great idea.

Again, I don't understand the point of your assertion and you chose not to clarify. Overview of that multi-year bug report indicates a number of things, the least of which that 0ad remained unsupported other than in stray ebuilds by whomever. Upstream still refers to the bug report as where to go for supposedly functional ebuilds. http://play0ad.com/download/linux/#Gentoo

But given the following comment, maybe there is a lack of interest in working with upstream...
hasufell wrote:
# rebuild premake again... this is the most stupid build system

Always sends a nice message.
Back to top
View user's profile Send private message
hasufell
Retired Dev
Retired Dev


Joined: 29 Oct 2011
Posts: 429

PostPosted: Tue Jun 10, 2014 7:49 pm    Post subject: Reply with quote

Navar wrote:
I'll stop from updating my posted ebuild (in unsupported software subforum) due to lack of interest as seen in the poll and the fact that you've taken on maintainership of a release into the tree. Since it was not clear to you, my interest was in the goals as stated in the apparently deprecated Sunrise project, including broken links.

That's a known problem. I have been a sunrise dev for some time (not anymore) and tried to improve the situation by also allowing people to contribute via github and bitbucket. I guess it just wasn't known enough and not many people used these possibilities.

The workflow in sunrise is still too slow and cumbersome for most contributors. No one wants to wait hours in an IRC channel to get a response, correct. IRC is a huge waste of time.

Gentoo in general has a lot of problems with workflow. It's difficult to change old systems.


Navar wrote:
As in yours? I do not have an account on Github. Nor was it at all clear from the bug report that you were maintainer.

First, I did say in that bug report that I am maintaining it, you probably missed it in the comments
https://bugs.gentoo.org/show_bug.cgi?id=278541#c110 (and a lot of follow-up comments)
I probably missed to add "[hasufell-overlay]" to the Whiteboard item in bugzilla.

It's just that I don't think duplicating work makes any sense. So whenever I am looking for ebuilds of something that is not yet in the tree, then I try to contact the guys who have already worked on it.

Also: If you are not very familiar with git, let me explain that you don't need an account anywhere to contribute git patches. You just clone it anonymously, commit locally on your machine and either use git format-patches and send them manually or use git send-email directly.
That's how a lot of people do this.

Navar wrote:
hasufell wrote:
if you know about multiplayer games and desyncs, then you know that running live ebuilds for multiplayer is a bad idea

Checksums and other methods have been used against that for a long time. I've done multiplayer gaming off and on for more than 20 years, but not via Gentoo ebuilds or Linux in general. Though I do not know the intricate details of what the 0ad developer is doing in that regard. Regardless, like you stated, it's still considered alpha status upstream. I view that as, end user--lower your expectations and expect issues aplenty, trunk checkout or packaged.

Some game developer have an even stronger opinion on this and say don't ever compile multiplayer games yourself.

I see live ebuilds for such multiplayer games only as a testing platform. And therefore, I don't put as much care into them as I do for regular releases.

Navar wrote:
hasufell wrote:
I don't build the live ebuilds every day [...] Which issues do my ebuilds have? This sounds too vague.

I was referring only to the live ebuild which I thought was obvious. At the time upstream had already done a feature freeze (leading to your release version shortly after). Most seen release ebuilds in the wild were in the sub alpha14 range (outdated).

A diff versus your live ebuild of 7 months ago would be an example, but your changelog and changes you've done on your release vs live indicate enough. I don't believe there's any reason to re-hash this other than your confusion on what I meant which was simply that you were one of at least 5+ authored versions floating out there (in addition to submissions to the bug report) which would no longer build. Simple as that, no prejudice in regards to you or any interest in seeking your attention for that matter, unless it was within the interests of the goals mentioned in Sunrise for proxy maintain.

Correct, only the live ebuild was outdated, nothing else. For that, I asked you to send a pull request (which does not necessarily involve a github account).

Navar wrote:
hasufell wrote:
Spreading ebuilds across multiple overlays is not a great idea.

Again, I don't understand the point of your assertion and you chose not to clarify. Overview of that multi-year bug report indicates a number of things, the least of which that 0ad remained unsupported other than in stray ebuilds by whomever. Upstream still refers to the bug report as where to go for supposedly functional ebuilds. http://play0ad.com/download/linux/#Gentoo

The overuse of overlays (not the concept of overlays in it's own) is bad for the gentoo ecosystem. These are only some reasons:
* people commit bad, wrong or even dangerous ebuilds that can really give you a nice broken system and a lot of work (this happened with multilib conversions)... mainly because they did not get any review
* people care less about QA, security and portable ebuilds, because it's so much easier to do your own thing than actually contributing to the real thing
* profiles cannot be managed between overlays... this leads to a lot of unfixable problems... masks, ebuild removals, broken deps and whatnot
* inter-dependencies between overlays cannot be expressed easily
* it makes people contribute less (like... actually working with gentoo devs)
* it makes people not think about becoming gentoo devs... they have an established overlay, a userbase and that's it (and I have asked quite a few maintainers of big overlays to become devs, no one cared)
* it decentralizes packaging which is a huge problem and causes duplicated work, a lot of problems for users to figure out the right ebuilds and some of the already mentioned points like total lack of coordination, QA etc

However, there are a few use cases where overlays make perfect sense:
* a testbed for non-trivial conversions like you need with new eclasses
* for alpha stage software (which 0ad technically still is... I was just able to convince the games team to accept this one. Usually, alpha stage games are not allowed)
* for WIP ebuilds that still need a lot of work
* for live ebuilds you don't want to actively maintain, but just toss them out
* testbed for big imports like qt5 etc
... and some more

But, sadly... a lot of people use them because it's just easier for them or they don't care. That makes everyone lose... us gentoo devs, because they didn't actually contribute and work with us... users, because they have to use (often) unreviewed ebuilds and take some time to look for them and try different overlays etc... and the overlay maintainers themselves, because they miss the chance to improve their ebuild skills by getting directly involved.
Navar wrote:

But given the following comment, maybe there is a lack of interest in working with upstream...

This can't be more wrong. I have communicated with upstream directly over those years, a lot. I'v submitted some patches to the build system and filed build failure related bugs. They even pinged me sometimes to test their releases before the rollout (especially since disabling precompiled headers regularly broke stuff). I was hanging out a lot on their IRC channel before I quit IRC... and I still have contact with them (they also hang out in the ingame lobby).
Navar wrote:
hasufell wrote:
# rebuild premake again... this is the most stupid build system

Always sends a nice message.

This is just the truth. If anyone feels offended by this, then he should. Statically compiling a SETTINGS file into a BINARY is just so wrong that I hope premake will ultimately die.
Back to top
View user's profile Send private message
Navar
Guru
Guru


Joined: 20 Aug 2012
Posts: 353

PostPosted: Tue Jun 10, 2014 11:23 pm    Post subject: Reply with quote

Hasufell,

Thanks for the thoughtful response. I don't disagree with your points and they clarified quite a bit which is appreciated. As far as git use goes, I've just utilized the basics for a remote checkout (clone) rather than add->commit->push. After a bit of searching with google prior, I was left with the impression to contribute on Github you had to pay to play. I'll give your suggestions a go.

It's technically my own fault for the time spent on this with regards to the live ebuilds. I view them as a precursor stepping stone to testing newest upstream changes before considering implementing in release builds. I am actually interested on the developer/testing side for 0ad, so I will continue keeping a live build route available locally.

As far as premake, I admit I'm not familiar with all the various cross platform build systems. It does seem like upstream is open to suggestions for improvements. The static compile settings(?) file into binary I presume has to do with the static Lua scripts as static C string buffers? I guess that was an upstream design decision for shipping ease, but I did like how they have mostly allowed us to build with system provided libs instead of packaged. If you played a major part in making that happen that's great. :)
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gamers & Players All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum