FreeOrion

It's been almost 10 years since I posted for the last time in the forums and it's fantastic to see that things are progressing nicely (and Geoff is still leading this crazy thing ).

I'm still reading some forums but I noticed a couple of things:
- The commit rate is really strong. There's an average of 2 commits per day, including weekends.
- The last release was 10 months ago and there's no nightly/weekly snapshots.

Would it be any interest on those? I think that it would really drive up testing and bug squashing as people could test the latest and greatest (in whichever state it was).

My only concerns are:
- Back in the day when I packaged this for Debian the compilation consumed insane amounts of memory during the final linking phase, so this would probably be a job for a desktop machine rather than a cheap VPS in some datacenter.
- Is there any way to cross-compile this on e.g. OS X? If I recall correctly there's none and you have to run Xcode on Apple hardware, but I wanted to make sure.

Jo-Con-El wrote:...there's no nightly/weekly snapshots.

Vezzra provides weekly Windows and OSX builds:
viewtopic.php?p=84588#p84588
viewtopic.php?p=84587#p84587

Builds for various Linux distributions might get some minor interest, particularly if there's a way to submit them to the distro's own servers / package manager as experimental / test / dev builds.

I set up to compile from source the day after I got my Linux box, I'm not really sure how to compile snapshots &c, we tend to recommend Linux users just compile because it's setup to be dead easy these days.

We're working on a new Release at the moment, the only thing stopping it is dev time right now, a few people are busy with real world stuff, but yeah, this last year in particular (ie since we moved to GitHub and picked up loads of new contributors) has been particularly good for work done and a release is due.

Geoff the Medio wrote:

Jo-Con-El wrote:...there's no nightly/weekly snapshots.

Vezzra provides weekly Windows and OSX builds:
viewtopic.php?p=84588#p84588
viewtopic.php?p=84587#p84587

Builds for various Linux distributions might get some minor interest, particularly if there's a way to submit them to the distro's own servers / package manager as experimental / test / dev builds.

Oh! I'm very sorry, I went to the compile wiki page but I missed completely the link in the front page to the test area.

This is fantastic news! Thanks.

How far are you from a release? Are you accepting only bugfixes at this point? I asked back in 2007 why FO didn't have gettext support and it's still not in the code, so I might convert all the string files into proper .po's and add gettext support, but I don't know if it's too disruptive at this point.

Too late for this release, especially as it sounds like a fairly major conversion task, we're in the "finish all declared projects and tidy up loose ends" stage.

However, if it's a good idea (I have no clue about backend technical stuff), getting it done early in the next cycle is a good thing as it gives plenty of conversion time.

We're hoping to have another release later this year, there's another major project we deliberately put on hold as it couldn't be finished in time (see the Fighters branch on GitHub for more) and if we can get both the AI able to cope with it AND the costs and tech side balanced that'll be the core of the next release. Initial playtesting is very cool.

Jo-Con-El wrote:I asked back in 2007 why FO didn't have gettext support and it's still not in the code, so I might convert all the string files into proper .po's and add gettext support, but I don't know if it's too disruptive at this point.

It's been suggested a couple times, but I haven't seen an convincing explanation why it would be better than the existing stringtables. As far as I could tell, gettext is designed to solve a problem that FreeOrion doesn't need solved, and is designed around assumptions about translatable text used that don't fit how FreeOrion handles much of its translatable text. If you want to discuss it again, start another thread, or look up the old one(s) about it, though.

Geoff the Medio wrote:

Jo-Con-El wrote:I asked back in 2007 why FO didn't have gettext support and it's still not in the code, so I might convert all the string files into proper .po's and add gettext support, but I don't know if it's too disruptive at this point.

It's been suggested a couple times, but I haven't seen an convincing explanation why it would be better than the existing stringtables. As far as I could tell, gettext is designed to solve a problem that FreeOrion doesn't need solved, and is designed around assumptions about translatable text used that don't fit how FreeOrion handles much of its translatable text. If you want to discuss it again, start another thread, or look up the old one(s) about it, though.

I asked it in the mailing list. Good times.

I'll start a post in the Programming section.

Jo-Con-El wrote:How far are you from a release?

Currently, 15 open issues/PRs on github

The major obstacle right now are a few important adjustments to the AI made necessary by several changes/additions to game mechanics. And most of our AI dev force is MIA...

Are you accepting only bugfixes at this point?

It's probably not quite that strict, but contributers are currently encouraged to finish/polish/playtest/fix things instead of starting ambitious projects, otherwise we never get that release out.

Jo-Con-El wrote: I asked back in 2007 why FO didn't have gettext support and it's still not in the code, so I might convert all the string files into proper .po's and add gettext support, but I don't know if it's too disruptive at this point.

Hi, as the FO french translator, I find a plain text file way more convenient to handle (compare, search, proofreading, regex) than a .po file. I worked one time with a .po file (with poedit then on a localization server) for the translation of a website, and it's a little bit tedious.