corepkg - a new core package and updates manager

[roberts]

First of all the SQL was rejected because of size. Not only in the base but the database itself.
Second having to maintain a second database on the server is overhead that is not needed.
Third quick updates are already in 4.3 by using a 96k database. Keyword has been in the infrastructure since 4.0.

Martin had written that if this was rejected he would consider it as "enjoyed creating it as a bit of an intellectual exercise". Instead I get a threat of forking. Then an announcement of his new forked desktop on that webpage was no attribution to Core and was in violation of the GPL. I was not talking about corepkg but the forked desktop remaster.

Is this the way to react when rejected. Force your way in? Your way or else? Are these winning tactics to gain favors?

Finally we have rules that everyone else has had to follow and respect. So, given the forceful nature of Martin, we should ignore the rules that we have?

Martin has had PMs with me and I was not sharing it with everyone. They are confidential.
But Martin comes across as only trying to make a contribution? Oh please.

And now Guy, has to, again, personally attack me. It makes me sick.

[martin]

Martin had written that if this was rejected he would consider it as "enjoyed creating it as a bit of an intellectual exercise".

I don't remember wording it like you say in the former, but the latter is true. I found it a great little piece of problem solving to engage it.

Then an announcement of his new forked desktop on that webpage was no attribution to Core and was in violation of the GPL.

I PM'ed you about this with no less than 3 major places on the website where Core was attributed and credited. I may have unintentionally violated the GPL, which I attempted to fix, and I realised it (creating a spin-off) is not what I really wanted to do anyway. So deleted it all.

Finally we have rules that everyone else has had to follow and respect. So, given the forceful nature of Martin, we should ignore the rules that we have?

I still have not been shown the rules I have broken. But if there were any, I apologise.

But Martin comes across as only trying to make a contribution? Oh please.

The spirit of my well intentioned efforts have always been to make a contribution. I went over-dramatic in one place because I was convinced my concept introduced a great improvement, and created a "fork" because noone seemed wholly interested, but it really not what I wanted to do.

and Roberts is still acting hostile and cynical towards me and I cannot really grasp the motivation behind it. Apologies if I tread on anyone's toes here, and I mean ANYONE.

So, something happened between "hey, I made some extensions for you guys" and "hey, I got a great idea to improve handling of system updates and stuff" which set Robert off and I can't find what or where.

[roberts]

You still cannot accept that it has been rejected for the reasons I have repeatedly stated.
I have no plans to bloat up the base or bog down the server with unecessary additional databases.

[martin]

You still cannot accept that it has been rejected for the reasons I have repeatedly stated.

No, I have accepted the reasons why you don't want to bundle it with core.gz and/or Core Plus and I have no problem with that. At all.

I have no plans to bloat up the base

Not asking you to.

bog down the server with unecessary additional databases.

You don't need to. It all runs on my server.

I simply want to provide it as an extension that people can use. They download the package database from MY server (via corepkg, it is completely invisible to them unless they know the source code), and when packages are fetched, the mirror is read from the usual place in the system. That's it.

edit: clarifying last part.

[Guy]

Just one question Robert: What part do you consider to be a personal attack?

Now a decision has been made. The idea is history. Let's all move on and be happy.

[roberts]

Then submit corepkg via normal channels.

[martin]

Then submit corepkg via normal channels.

Will do. I have been testing the polishing up of v0.82 all evening. It's bedtime here in Oz, so tomorrow I will package it up, update the code on sourceforge and await. Do you have any objections if I were to also post a direct link to the tcz hosted on my personal home page from here directly? Noone has to see the site, just get the file.

If you have a problem with that, I can probably wait for it to be uploaded.

[roberts]

It is not I but forum rule 2 specifically states:
http://forum.tinycorelinux.net/index.php/topic,7738.msg41259.html#msg41259
2. No attachments/links of binary extensions

[martin]

oh right, yes.

[bmarkus]

I'm sad.

On one side, there is Tiny Core a great toolkit to build special systems, distributions, remasters (you can name them as you want). Flexible, easy to customize, efficient. A great tool to make your own system.

On the other side there is a person who invest time, energy to use it for such purpose and to implement what he think is useful and to create a new desktop system. I have never tried it, but all my symphaty goes to those creating something new, somethin different.

Now his project is dead according to its WEB site:

The Nucleus Desktop is no more. We recommend TinyCore Linux instead.

I'm sad. I'm really sad to see a project to die before its birth.

[Guy]

Martin

quick updates are already in 4.3 by using a 96k database.

Before going ahead, I suggest you have a good look at the official apps, and see if yours offers a genuine advantage. I have not looked into this myself. If your app does not offer any genuine advantage, give up the idea.

If yours does offer a genuine advantage, can this be done using the official database. If so use it, and don't host your own.

[Rich]

Hi Martin C
I believe I've spotted one flaw in what you are doing.

* You grabbed the TCZs, but you didn't get the MD5's. CorePkg can re-generate the md5's for you.

This defeats the purpose of the MD5 file. By downloading both the tcz and MD5 and verifying them
against each other you ensure that neither file was corrupted.

[martin]

If yours does offer a genuine advantage, can this be done using the official database. If so use it, and don't host your own.

Robert told me via PM once that he was basing his own version of a database on everything that is included in busybox, sed and awk primarily, I believe, but I could be wrong.

My solution is more flexible and more, IMHO, and it will be up to users to tell me if I am wasting my time. The other benefit to CorePkg is the flexibility in the system; I've seen mention of people looking at wanting the possible ability to categorise extensions in the future, CorePkg offers this.

Of course, at the very start, Robert was only hinting at behind the scenes development to address the problem of slow update checks. I had no idea when 4.3 was coming along, so I had a go at it myself and produced what I believe to be a very, very fast and efficient method. It is up to the users to decide which one they prefer.

Like I said before, in some implicit ways CorePkg might be competition, but I rather see it as complementary.

[martin]

Hi Martin C
I believe I've spotted one flaw in what you are doing.

* You grabbed the TCZs, but you didn't get the MD5's. CorePkg can re-generate the md5's for you.

This defeats the purpose of the MD5 file. By downloading both the tcz and MD5 and verifying them
against each other you ensure that neither file was corrupted.

That feature was a quick hack. Under the most strict circumstances, you may be right. In other times, you may know you have a great connection and just get the file. One quick example for this: to be considerate to other network users, I got libreoffice.tcz and sun-jre.tcz in my off-peak time. Just the files. then, during the day, I got the rest using CorePkg's "offline downloader". I then used CorePkg rather than waste a network request to re-create the md5 files for later when doing an update check. I had already tried the extension for myself so I knew that it worked and I wasn't too concerned at the time of verifying it manually. It's just a matter of choice; There is more than one way to do it.

If noone ever, ever uses that feature for themselves, I wont be offended. Like I said, a quick hack. It's not even a primary feature that I cared about even if it works flawlessly. It was just another part of the CorePkg toolbox.

[martin]

A follow-up:
* Can anyone tell me the right way to link to a set of easy to read instructions for CorePkg to give people an idea of how it works so that the link wont get deleted? The Rules here http://forum.tinycorelinux.net/index.php/topic,7738.0.html aren't clear on it, nor why my previous links were deleted (which were not commercial in any way, and the only links IT contained were to sourceforge and tinycorelinux.com).
* I'm looking for 1 or 2 testers for CorePkg before submission. Let me know here or via PM if you're interested.

edit: ahh stuff it, I'll just include the text file as it is and hope people know where to look. still looking for testers.