Wedge

Public area => The Pub => Off-topic => Topic started by: emanuele on October 21st, 2013, 02:35 PM

Title: git hiccups
Post by: emanuele on October 21st, 2013, 02:35 PM
Quote from Nao on October 20th, 2013, 04:26 PM
Fact is, all he's been doing at SMF 2.1 is adding features that are taken STRAIGHT from Wedge. In fact, some of his commits (look at his 05be2b92 revert) have Wedge code in them, such as wesql::query, if proof was needed that he didn't just take his 'ideas', he was taking his code, too.
Wrong subject:
https://github.com/SimpleMachines/SMF2.1/pull/691/files#diff-cfc225eff5aeb11b2e42d1a705c3ae63R476
Title: Re: git hiccups
Post by: forumsearch0r on October 21st, 2013, 02:50 PM
Quote from Nao on October 21st, 2013, 12:50 PM
- Indenazi hack. Only executed if used in a skin.
Sounds neat. What is it?
Quote from Nao on October 21st, 2013, 12:50 PM
The SMF codebase was fantastic and very modern in 2003.
Ah, good old UBB, YaBB and, ehm... PHP-Nuke days.
Title: Re: git hiccups
Post by: Nao on October 21st, 2013, 06:21 PM
Quote from emanuele on October 21st, 2013, 02:35 PM
Quote from Nao on October 20th, 2013, 04:26 PM
Fact is, all he's been doing at SMF 2.1 is adding features that are taken STRAIGHT from Wedge. In fact, some of his commits (look at his 05be2b92 revert) have Wedge code in them, such as wesql::query, if proof was needed that he didn't just take his 'ideas', he was taking his code, too.
Wrong subject:
https://github.com/SimpleMachines/SMF2.1/pull/691/files#diff-cfc225eff5aeb11b2e42d1a705c3ae63R476
That page shows a commit by Arantor. Why does it also say live627:ssi? Is this some kind of gitty trick?
If it's a commit by John, why is Pete credited for it..?
And if it is, why does he use Wedge terminology, when he's made very clear (by his silence mostly!) that he chose not to keep working on the Wedge codebase?
Title: Re: git hiccups
Post by: Nao on October 21st, 2013, 06:24 PM
Quote from forumsearch0r on October 21st, 2013, 02:50 PM
Quote from Nao on October 21st, 2013, 12:50 PM
- Indenazi hack. Only executed if used in a skin.
Sounds neat. What is it?
Nothing more than a vanity hack... It allows the skinner to fine-tune tabs (spaces) per-block in the HTML source code. I wrote it for use within skins that would want to move around the sidebar, like Warm did, but it's lost some of its value by now. Plus, I'm the only one around who cares about HTML formatting, so, whatever... Not worth discussing, really, it's only for the anally retentive like me ;)
Quote
Ah, good old UBB, YaBB and, ehm... PHP-Nuke days.
Ah, PHP Nuke, and its 'successor' PostNuke... I actually tried to build a website using those, back in the day...! (I ended up using SPIP, which had the better codebase; the website is still live, surprisingly.)
Title: Re: git hiccups
Post by: forumsearch0r on October 21st, 2013, 06:27 PM
HTML formatting is a great Plus when you have mixed contents (like HTML and PHP) which sometimes is hard to avoid IMO.
Title: Re: git hiccups
Post by: Nao on October 21st, 2013, 06:41 PM
Well, it's in the final page, not before... (?)
Title: Re: git hiccups
Post by: forumsearch0r on October 21st, 2013, 06:50 PM
Hm, OK. At least that's nice for debugging. :D
Title: Re: git hiccups
Post by: TE on October 21st, 2013, 07:13 PM
Quote
That page shows a commit by Arantor. Why does it also say live627:ssi? Is this some kind of gitty trick?
If it's a commit by John, why is Pete credited for it..?
Nope, it's a commit from live627 .. his branch was named "ssi". Arantor was the one who merged it.

https://github.com/SimpleMachines/SMF2.1/pull/691
Title: Re: git hiccups
Post by: emanuele on October 21st, 2013, 08:57 PM
Quote from Nao on October 21st, 2013, 06:21 PM
That page shows a commit by Arantor. Why does it also say live627:ssi? Is this some kind of gitty trick?
If it's a commit by John, why is Pete credited for it..?
And if it is, why does he use Wedge terminology, when he's made very clear (by his silence mostly!) that he chose not to keep working on the Wedge codebase?
The commit you posted was by Arantor to fix live's commit (that I linked) with Wedge-like syntax.

Ops, didn't see TE's post.
Title: Re: git hiccups
Post by: live627 on October 22nd, 2013, 12:21 AM
Wait, now I can't submit fixes to SMF?

I get the feeling that Nao thinks everyone has turned against him and that someone, somewhere has declared war. Are we schoolboys that fight at every chance we get?
Title: Re: git hiccups
Post by: Nao on October 22nd, 2013, 09:45 AM
Quote from TE on October 21st, 2013, 07:13 PM
Quote
That page shows a commit by Arantor. Why does it also say live627:ssi? Is this some kind of gitty trick?
If it's a commit by John, why is Pete credited for it..?
Nope, it's a commit from live627 .. his branch was named "ssi". Arantor was the one who merged it.

https://github.com/SimpleMachines/SMF2.1/pull/691
<sigh>
From what I saw on the Github site, because Arantor 'merged' the commit, and he had his name in bold (while the live627:ssi branch is in small-size Courier New characters), it wasn't hard to imagine that he committed it.

Let's see the path that I followed...

https://github.com/SimpleMachines/
Click SMF 2.1. Brings us to...

https://github.com/SimpleMachines/SMF2.1
Search (ctrl+f) for 'wedge' in the page. The last commit on SSI.php has it, so it's plain in your face (and the reason why I mentioned this thing at all, because I'm afraid I don't spend a lot of time on either Elk and SMF codebases, ahem.)
Click the link. Brings us to...

https://github.com/SimpleMachines/SMF2.1/commit/aec98628cbc78d30a38f2dcf68dfc747740a899f
"Arantor authored 14 days ago"
Click on "1 parent: (hash)", top right of the page. Brings us to...

https://github.com/SimpleMachines/SMF2.1/commit/444bdeac2797b058b2eac26cf756d7c57f2053d9
In the very same place, it says "Arantor authored 14 days ago".

So, believe what you think, but to me, it seemed obvious that Arantor made both commits... It's what Github insists on saying after exactly THREE clicks from the SimpleMachines homepage.

Is Github that fucked up, or is it Pete who did the wrong move and found himself crediting himself for someone else's work?
Title: Re: git hiccups
Post by: TE on October 22nd, 2013, 09:53 AM
Just wanted to make sure Arantor isn't blamed for something he isn't responsible for...
Fact is:
This commit by Arantor:
https://github.com/SimpleMachines/SMF2.1/commit/aec98628cbc78d30a38f2dcf68dfc747740a899f
reverts this commit by Live627:
https://github.com/SimpleMachines/SMF2.1/commit/02e097e3b87dd22529a491ce9c7b357d9d41168b

Live627 should tell us his version..

 
Title: Re: git hiccups
Post by: Nao on October 22nd, 2013, 10:07 AM
Quote from live627 on October 22nd, 2013, 12:21 AM
Wait, now I can't submit fixes to SMF?
Who said you couldn't?
Quote
I get the feeling that Nao thinks everyone has turned against him and that someone, somewhere has declared war. Are we schoolboys that fight at every chance we get?
Where do you get this feeling? Just because I check on people's work and mention things that are of concern to me, I'm suddenly paranoid?

I'm only asking that anyone with access to the Wedge codebase doesn't use/distribute my code anywhere.
You can do whatever else you want.

Still, I thought all you wanted was for Wedge to switch to Git, so that you could contribute more easily, and yet, after I opened the repo, you didn't comment on it. I sent you an invite, because you're always going to have a slot for you anyway, accept it if you want.
But since I have limited user team space on that repo, I'll still need an answer.
Title: Re: git hiccups
Post by: live627 on October 22nd, 2013, 10:32 AM
I sent in that very patch in 2010. That is my only contribution worth porting to SMF.

I'll get a good Git client installed tomorrow, and accept the invite. I have almost no free time nowadays, might get even busier (yikes!) over the holidays.
Title: Re: git hiccups
Post by: Nao on October 22nd, 2013, 10:43 AM
Quote from TE on October 22nd, 2013, 09:53 AM
Just wanted to make sure Arantor isn't blamed for something he isn't responsible for...
Yeah, he doesn't need that to blame himself... ::)
Quote
reverts this commit by Live627:
https://github.com/SimpleMachines/SMF2.1/commit/02e097e3b87dd22529a491ce9c7b357d9d41168b
Okay, I got it: from the list of links I gave, you have to click "691" or whatever in the header, and then click on the commit link that's in the first comment, or something. This adds an extra 2 clicks, but then you get the original commit[1].

So, basically, Github is the culprit here for not being very helpful to people who might read these pages without knowing a thing about Git or Github or whatever. Regarding the argument about the use of 'wesql' in the codebase, this obviously makes it moot, but that still doesn't change the fact that (and even Arantor recognizes it himself) all of his work on SMF 2.1 is to bring Wedge features to it, which *obviously* is reason enough for me to be pissed off by his actions.

Remember, two months ago, he argued that he was leaving the forum software community to focus on making use of the Unity engine and create video games with Asgard. Something I could perfectly understand; in fact, the only times I considered leaving Wedge, was because I figured I needed to do something to make a living. Well, in the end, not only did Pete stay in this little world, but worse, he chose to go to the only place we expected each other never to deal with. The very place that had us create Wedge in the first place. And not only that, but now he's contributing his Wedge code to them (thus removing value from the project he poured his heart into for three years). He's shooting himself in the foot, wasting his time, and hurting me. Had I left Wedge before him, I'd never have done that kind of deal. I have a fucking conscience. All code I wrote for the team, is for use within the team only, at least until the day Wedge uses an actual open source license, at which point, everyone have fun. I don't have any right to reuse it for the competitor we left in the first place because of the way they treated us[2].

All I'm saying is that I feel Pete disrespected me, by systematically doing the opposite of what he assured me he would do. So, the next things I'm expecting, are for (1) his Wedge repos to disappear without warning (reason enough to move them), and (2) for my code to find itself in SMF. "By mistake", of course. But there will always be a revision log to make sure not a single line of my Wedge code finds itself in SMF. So, now what I'm *hoping* for (rather than expecting), is for Pete to take a local copy of Wedge, and only copy-paste code from HIS revisions.

That's all there is to say.
 1. I thought that branches in Git allowed for pull requests to essentially be a merge, which isn't destructive like a rebase, and kept all of the original commits. Looks like it creates an extra reference to the same commit, with a different hash. Oh, I have to read more about that...
 2. Again: I have nothing against the current SMF dev team. I'm referring to the remainders of the 2010 SMF 'management'.
Title: Re: git hiccups
Post by: Nao on October 22nd, 2013, 10:46 AM
Quote from live627 on October 22nd, 2013, 10:32 AM
I'll get a good Git client installed tomorrow, and accept the invite. I have almost no free time nowadays, might get even busier (yikes!) over the holidays.
Well, I'm not asking for anyone to contribute anything; I'm just saying, Wedge has suddenly become easier to contribute to (and be properly credited in the changelog), if you want to.

PS: How come you're on github and not using a git client..? Are you using GitHub for Windows..?

(I'm personally happy with TortoiseGit. Used to hate it, but I reconsidered when I realized that TortoiseHg was too different from what I was accustomed to, in TortoiseSVN.)
Title: Re: git hiccups
Post by: TE on October 22nd, 2013, 10:57 AM
Quote from live627 on October 22nd, 2013, 10:32 AM
I sent in that very patch in 2010. That is my only contribution worth porting to SMF.
Thank you for clarification.. Knowing that earlier could have avoided some misunderstandings..

I can recommend Netbeans / gitbash combination.

You can do almost all things via GUI in Netbeans except the git pull -- rebase (maybe it's possible but I haven't figured how to do it).
Title: Re: git hiccups
Post by: Nao on October 22nd, 2013, 11:03 AM
Quote from TE on October 20th, 2013, 11:01 AM
Last time I checked Wedge (a year ago or so) it wasn't true OOP but using singletons all over the place..
Skeletons are a truly OOP class. That's the only one I ever wrote to be OOP, but guess what, it's also the only one that could benefit from being OOP.
Do you have any other idea what could really benefit from OOP..?!
Quote
Elk is using OOP where it makes sense, yes. We have a coding guideline and we simply follow our own standards:
But SMF is already MVC to begin with...
Quote
The point of following standards is to make it less complicated for developers and easier extendable. And last but not least to use other common standards such as unit-testing and build-testing (travis ci).
I don't know much about these, so I'll pass.
Quote
Nao, I would probably have contributed my code and energy to Wedge but I wasn't allowed to do so..  Sadly you (and Pete) never gave me write access to the main repo, thus I moved on at some point.. Sorry.
We were supposed to release much earlier, which is why we always postponed giving more writing rights to anyone in the team. Really, we simply refused write access to *anyone*, not just you... And now that Pete is gone, I'm working on that. Beginning with the conversion to Git, and support for pull requests (since yesterday.) So, if you want access, there's a topic for that... ;)
I'm afraid I don't remember you ever asking for commit access to the main repo, though...! I don't remember you providing any patches either. Only Shitiz and John did. You wrote the importer, and then you got feverish when one day I confirmed that we were going to go with the original SMF license and were considering a paid version, and you didn't want anything to do with that, and you left. Then you saw that we didn't implement that, and you came back later, and then one day you left again, for no apparent reason.

It just stuck with me that you were more interested in helping Elk than Wedge, is all.
Quote
And please: I'm not up for a battle comparing Elk, Wedge and SMF... We all have wasted more than enough time & energy to complain about SMF, their NPO structure and such stuff. At least I don't have the time nor the passion to start such type of "battle" once again. Elk and Wedge are simply following completely different approaches ..
I don't know, how different are they exactly...?
It's not like they're incompatible, either...
Except for Wedge blatantly breaking mod compatibility, of course! :P
(And Pete never ever documenting any of his codebase changes... :whistle:)
Title: Re: git hiccups
Post by: BurkeKnight on October 22nd, 2013, 11:27 AM
For Git client: SourceTree is supposed to be pretty easy to use.
Title: Re: git hiccups
Post by: Nao on October 22nd, 2013, 01:46 PM
Yeah... Well, I don't know about the general opinion, but I didn't like SourceTree when it was released, and I uninstalled it less than a day after trying it out. Was some time ago, dunno if it has improved in the meantime. Just like SmartGit -- apparently, recently they renamed it to 'SmartGitHg' (well, what a wonderful rename... :lol:), maybe it's improved, but for what I know, I liked TortoiseGit best when I tested them all.
Title: Re: git hiccups
Post by: forumsearch0r on October 22nd, 2013, 02:02 PM
I'd use git on the shell ... :D

... there is a public Wedge repo?
Title: Re: git hiccups
Post by: Nao on October 22nd, 2013, 02:14 PM
Nope, it's still a private repo!
Refer to my roadmap to determine when it'll be made public, I guess.
Title: Re: git hiccups
Post by: forumsearch0r on October 22nd, 2013, 02:15 PM
Damn. OK. :)
Title: Re: git hiccups
Post by: Nao on February 20th, 2014, 01:53 PM
I was just wondering... Did anyone ever receive spam at their git e-mails, following your use of that particular e-mail address as their commit e-mail on public github repos..?

I've always been very wary of making my e-mail address public (actually, I even have two addresses, in case one of the two ends up being spammed), because my old e-mail address (nao at kyodai dot com) was spammed to death -- about 30K spams per hour at its peak, meaning I eventually had to give up on that e-mail account. I'm aware that Google has better spam management (I'm already getting spam in both my gmail boxes despite NEVER making them public... Sigh, but at least they're mostly in the spam box), but I'm really really not excited about the idea of giving my e-mail address "like that".
My fake e-mail is nao at wedge, implying the org thing for those who can read a website URL, which in turn redirects to my gmail.

GitHub no longer wants to associate fake e-mails with accounts, even though they accepted them up til now.
So, if I associate the 'implied' e-mail address with my git account, the results are:

- I'll have to rewrite history for ALL of my frigging repos (about a dozen, including my Nao ones)... Seriously, that's f*cked up. I spent 4 days last month rewriting my histories to be as clean as possible, squashing newline-fixing commits together to avoid f*cking with the "lines added/lines removed" statistics, etc... I wanted to be as clean as possible, and then 3 weeks later -- GitHub is telling me they changed their mind and want me to do the same again. What next, then..?

- None of the repos I can't control will be able to update my e-mail address, although it doesn't matter as much.

I'd like to know who's currently using a cloned repo of Wedge, or any of its repos.
Title: Re: git hiccups
Post by: Pandos on February 20th, 2014, 01:58 PM
I do.
It's a 1:1 copy. Even languages and plugins.
Title: Re: git hiccups
Post by: emanuele on February 20th, 2014, 04:02 PM
Quote from Nao on February 20th, 2014, 01:53 PM
I was just wondering... Did anyone ever receive spam at their git e-mails, following your use of that particular e-mail address as their commit e-mail on public github repos..?
I'm using my email just because it was already considered the "garbage" email (i.e. the email I use for almost anything like mailing list, not trustable websites, etc.), I didn't see any peak on spam, I always have 22/23 new emails in gmail spam folder, so it's kind of stable.
Others have seen spam (I seem to remember SleePy ranting about it).
Dunno.
Title: Re: git hiccups
Post by: Nao on February 21st, 2014, 01:07 AM
So what would you in my place, guys..?
Posted: February 20th, 2014, 09:46 PM

I'm just bumping this because there have been many updated topics, so I'm increasing visibility like the bastard I am. :P

See you tomorrow!
Title: Re: git hiccups
Post by: Nao on February 21st, 2014, 12:30 PM
Okay, last night GitHub support replied to me, and recognizes that their anti-fake test is a bit too eager, and probably shouldn't be triggered on a fake e-mail such as mine.

They're working on it.

So I'll give it another few days, and hopefully it'll be fixed by then.

If there's one thing I can never take away from GitHub, it's the quality of their tech support. It's free, and yet it never feels like they're sending stock answers, and they're always ready to listen to suggestions. That's amazing. I'd love to be (seen) as caring as they are!
Title: Re: git hiccups
Post by: Bunstonious on February 21st, 2014, 03:15 PM
Quote from Nao on February 21st, 2014, 12:30 PM
If there's one thing I can never take away from GitHub, it's the quality of their tech support. It's free, and yet it never feels like they're sending stock answers, and they're always ready to listen to suggestions. That's amazing. I'd love to be (seen) as caring as they are!
You are... Well mostly ;)
Title: Re: git hiccups
Post by: Nao on February 24th, 2014, 11:26 AM
Oh, forgot to post about this: the fix was pushed a couple of days ago, and I'm happy to announce I won't be tempted to rewrite any history any time soon... :P

It's a bit tempting to start using their 'dedicated' anti-spam addresses, but OTOH, I like having a non-GitHub specific e-mail for my commits.