From domm at cpan.org Fri Nov 1 17:53:03 2019 From: domm at cpan.org (Thomas Klausner) Date: Fri, 1 Nov 2019 17:53:03 +0100 Subject: [Act] Unifiying Act PSGI forks Message-ID: <20191101165303.GP13441@plix.at> Hi! (cc'ing Wesley Schwengle and Harald J?rg because I'm not sure if they are on the Act mailing list) Today I finally took a look at the various Act branches floating around, with the goal to merge them into one Act to rule them all. I failed, like so many people before me. But I think I made a bit progress, and things are a bit clearer to me now. So let me share: I created a new default branch 'master' for act-testing/Act (because I see no point in having weird branch names instead of master) I merged BooK/Act into act-testing/Act, which was rather easy. AFAIK BooK/Act is the version powering act.yapc.eu (but still running on mod_perl) Then I took a look at waterkip/Act. Uff.. This repo contains 3 tries to get a PSGI version running: 2011, Graham Knop & Rob Hoelz 2016, Ilya Chesnokov 2019, Wesley Schwengle; Harald J?rg (via cherry-picking I guess) The diff between waterkip/Act and act-testing/Act is *huge*: 102 files changed, 7935 insertions(+), 1120 deletions(-) Merging it results in ~25 conflicts. So not something I want to try right away. It seems that a lot of the old commits have been already merged (or cherry-picked?) from waterkip/Act (or other repos) into act-testing/Act (but with different commit ids, eg: 7056d2bd24c84e646d6eb7271b5883a215d14c5b vs 70368e270ce1e325cfd86689cf3d778b84c06727 It seems (to me) that commits to waterkip/Act after Sat Apr 6 12:42:58 2019 +0200 have not been merged. that is, this commit is on both repos: https://github.com/waterkip/Act/commit/4d01f92aa64e26b33660a56c53bfaffdc834150a, but later ones aren't. And when I diff this, I get a slightly less big diff: git diff --stat waterkip_docker/dockerize-act 4d01f92aa64e26b33660a56c53bfaffdc834150.. 24 files changed, 243 insertions(+), 264 deletions(-) It seems that waterkip/Act changed quite a lot (to clean up the forest of symlinks? to clean up docker-compose?) So, Wesley, as you have authored those commits, could you try to integrate them into act-testing/Act? Or are those commits a too big change? Are there any other relevant forks flying around somewhere? I really think we need to deploy the PSGI version ASAP, either to on of the French Perl Mongers servers, or to a new one (if the French Perl Mongers don't want to run the PSGI version). Having a stable and modern version of Act is IMO more important than access to the old database (and thus old user info). Greetings, domm -- #!/usr/bin/perl http://domm.plix.at for(ref bless{},just'another'perl'hacker){s-:+-$"-g&&print$_.$/} From haj at posteo.de Fri Nov 1 19:07:23 2019 From: haj at posteo.de (=?UTF-8?Q?Harald_J=c3=b6rg?=) Date: Fri, 1 Nov 2019 19:07:23 +0100 Subject: [Act] Unifiying Act PSGI forks In-Reply-To: <20191101165303.GP13441@plix.at> References: <20191101165303.GP13441@plix.at> Message-ID: <68a1ee93-6e5e-8984-9cc0-e1b346085968@posteo.de> Hi Thomas, you write: > (cc'ing Wesley Schwengle and Harald J?rg because I'm not sure if they > are on the Act mailing list) I am on that list. I should be, shouldn't I? > Today I finally took a look at the various Act branches floating around, > with the goal to merge them into one Act to rule them all. > > I failed, like so many people before me. > > But I think I made a bit progress, and things are a bit clearer to me > now. So let me share: > > I created a new default branch 'master' for act-testing/Act (because I > see no point in having weird branch names instead of master) Great! I started my work so far under a different branch at a time when it was not at all clear whether and how branches would be merged - after all, it has more resemblance to the psgi branch of BooK's than to the master branch over there. I don't mind getting rid of the evolution branch. > I merged BooK/Act into act-testing/Act, which was rather easy. AFAIK > BooK/Act is the version powering act.yapc.eu (but still running on > mod_perl) AFAICS there's the "stable" branch which powers the server, master is the development branch. > [...] > It seems (to me) that commits to waterkip/Act after Sat Apr 6 12:42:58 2019 +0200 > have not been merged. That's correct. I completely missed that Wesley was still working on his branch after submitting his PR, but fixed some issues I discovered independently in act-testing/Act. A merge needs care. > I really think we need to deploy the PSGI version ASAP, either to on of > the French Perl Mongers servers, or to a new one (if the French Perl > Mongers don't want to run the PSGI version). Having a stable and modern > version of Act is IMO more important than access to the old database > (and thus old user info). Well, testing activity on the demo server (https://act-test.plix.at/demo) is non-existing, so it would be brave to call it stable. But at least we've modernized _one_ of the pain points. -- Cheers, haj From emmanuel at seyman.fr Fri Nov 1 20:00:10 2019 From: emmanuel at seyman.fr (Emmanuel Seyman) Date: Fri, 1 Nov 2019 20:00:10 +0100 Subject: [Act] Unifiying Act PSGI forks In-Reply-To: <20191101165303.GP13441@plix.at> References: <20191101165303.GP13441@plix.at> Message-ID: <20191101190010.GA28897@orient.home> * Thomas Klausner [01/11/2019 17:53] : > > AFAIK > BooK/Act is the version powering act.yapc.eu (but still running on > mod_perl) Indeed. Specifically, we're running the stable branch in production while the master branch is for development. > I really think we need to deploy the PSGI version ASAP, either to on of > the French Perl Mongers servers, or to a new one (if the French Perl > Mongers don't want to run the PSGI version). Having a stable and modern > version of Act is IMO more important than access to the old database > (and thus old user info). We're in the process of deploying a Centos 8 vm on which we'll be running the psgi branch of Book/Act. We'll then be able to offer a choice to conference organisers which branch they went their conference to be run under. Emmanuel From domm at cpan.org Fri Nov 1 20:49:33 2019 From: domm at cpan.org (Thomas Klausner) Date: Fri, 1 Nov 2019 20:49:33 +0100 Subject: [Act] Unifiying Act PSGI forks In-Reply-To: <20191101190010.GA28897@orient.home> References: <20191101165303.GP13441@plix.at> <20191101190010.GA28897@orient.home> Message-ID: <20191101194933.GT13441@plix.at> Hi! On Fri, Nov 01, 2019 at 08:00:10PM +0100, Emmanuel Seyman wrote: > We're in the process of deploying a Centos 8 vm on which we'll be running > the psgi branch of Book/Act. We'll then be able to offer a choice to conference > organisers which branch they went their conference to be run under. I don't think it's a good idea to target the psgi branch of Book/Act! I really want to resolve the two main forks (haj=act-testing, waterkip) and set up a new, clean repo containing only the PSGI version in a stable master branch, which can be run locally via docker. Only then we have a chance to get people to work on the actual issues (eg registration process) Greetings, domm -- #!/usr/bin/perl http://domm.plix.at for(ref bless{},just'another'perl'hacker){s-:+-$"-g&&print$_.$/} From emmanuel at seyman.fr Fri Nov 1 21:04:19 2019 From: emmanuel at seyman.fr (Emmanuel Seyman) Date: Fri, 1 Nov 2019 21:04:19 +0100 Subject: [Act] Unifiying Act PSGI forks In-Reply-To: <20191101194933.GT13441@plix.at> References: <20191101165303.GP13441@plix.at> <20191101190010.GA28897@orient.home> <20191101194933.GT13441@plix.at> Message-ID: <20191101200419.GA32122@orient.home> * Thomas Klausner [01/11/2019 20:49] : > > I don't think it's a good idea to target the psgi branch of Book/Act! Laurent was insistant that we do this but I can't recall the reasons why. I'll check with him if this is really the best branch availible to us. > I really want to resolve the two main forks (haj=act-testing, waterkip) > and set up a new, clean repo containing only the PSGI version in a > stable master branch, which can be run locally via docker. That would be great. Once we have this, we can squash the psgi branch of Book/Act with this code and the problem will be solved. Emmanuel