Recent

Author Topic: NewPascal plans, Generics.Collections and FPC trunk  (Read 45975 times)

Thaddy

  • Hero Member
  • *****
  • Posts: 14204
  • Probably until I exterminate Putin.
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #30 on: May 08, 2018, 09:46:51 pm »
Maciej, stop it. Work it out. These kind of comments aren't very helpful.
You know I respect and test your contributions.
Specialize a type, not a var.

hnb

  • Sr. Member
  • ****
  • Posts: 270
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #31 on: May 08, 2018, 11:06:00 pm »
Maciej, stop it. Work it out. These kind of comments aren't very helpful.
You know I respect and test your contributions.
I am opened to continue my work for community, not for Michael but I have ban from Michael which is also boss of FPC foundation. How can I contribute bug fixes for compiler, updates for my libraries without access to trunk and core mailing list?

I am really forced to continue NewPascal way. For sure we need new more open minded foundation with place for everyone (Michael is also welcome).

This is all Thaddy, I am not angry anymore I just want to work on modern Pascal in friendly atmosphere.
Checkout NewPascal initiative and donate beer - ready to use tuned FPC compiler + Lazarus for mORMot project

best regards,
Maciej Izak

Blaazen

  • Hero Member
  • *****
  • Posts: 3237
  • POKE 54296,15
    • Eye-Candy Controls
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #32 on: May 08, 2018, 11:17:20 pm »
Hi,

on ML you mentioned: "I have ban for FPC core mailing list (no more "Daily test suite diffs" for me)".

What are "Daily test suite diffs"?

Thanks
Lazarus 2.3.0 (rev main-2_3-2863...) FPC 3.3.1 x86_64-linux-qt Chakra, Qt 4.8.7/5.13.2, Plasma 5.17.3
Lazarus 1.8.2 r57369 FPC 3.0.4 i386-win32-win32/win64 Wine 3.21

Try Eye-Candy Controls: https://sourceforge.net/projects/eccontrols/files/

HeavyUser

  • Sr. Member
  • ****
  • Posts: 397
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #33 on: May 09, 2018, 01:19:28 am »
Maciej, stop it. Work it out. These kind of comments aren't very helpful.
You know I respect and test your contributions.
I am opened to continue my work for community, not for Michael but I have ban from Michael which is also boss of FPC foundation. How can I contribute bug fixes for compiler, updates for my libraries without access to trunk and core mailing list?
use the bugtracker, attach the diff of the files to a bug report and walk away, if you really want to do. If not then as long as your repo is online and accessible by the team then they can pull what ever they need.

marcov

  • Administrator
  • Hero Member
  • *
  • Posts: 11383
  • FPC developer.
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #34 on: May 09, 2018, 03:03:42 am »
on ML you mentioned: "I have ban for FPC core mailing list (no more "Daily test suite diffs" for me)".

What are "Daily test suite diffs"?

It summarizes the differences in the nightly testsuite runs on https://www.freepascal.org/testsuite/cgi-bin/testsuite.cgi


bee

  • Sr. Member
  • ****
  • Posts: 393
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #35 on: May 09, 2018, 04:49:32 am »
On one side, everybody agrees that both FPC and Lazarus need more and more volunteers, especially from younger generation to ensure these great projects will have a bright future. But on the other side, sometimes new people aren't very welcome to contribute if this new people doesn't conform to the seniors way of thinking. This Maciej's case is the latest example, but I've seen similar cases since I joined FPC/Lazarus.

On one side, everybody knows that we need to get popular so this project could attract both volunteers and fundings, especially from younger generation. But on the other side, I don't see much effort has been done to overcome this problem. The core devs barely think about it, because they mostly think about the technical things. I don't blame them because it's their focus and concern, but if nobody would tackle and handle these non-technical problems, then we'll getting nowhere. We'll be forever a niché in software development. Nobody cares about Pascal but a bunch of old people who still maintaining their projects in their niché market.

Now I think I understand why CodeTyphon chose its way. Rather than contributing directly to the project, Sternas (CT's author) simply took the project with him and do it his way so he didn't need to bother with core devs approvals and agreements. Although he did some nasty things with the licenses, but look… FPC/Laz devs don't mind, even support him. So, I think Maciej's decision to work on his own FPC fork is correct. As long as he keeps following FPC development and base his work on it, I'll support him.

Maybe it's time for FPC and Lazarus to have a new "competitor" other than Delphi so they will start to be more open minded accepting new ideas from new people. If Maciej keeps continuing his work and regularly release stable New Pascal with new features that offers something more than the orginal FPC, I'm sure New Pascal would have its own users and fans. Just like CodeTyphon.
-Bee-

A long time pascal lover.

engkin

  • Hero Member
  • *****
  • Posts: 3112
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #36 on: May 09, 2018, 06:41:50 am »
Now I think I understand why CodeTyphon chose its way. Rather than contributing directly to the project, Sternas (CT's author) simply took the project with him and do it his way so he didn't need to bother with core devs approvals and agreements.
I think this conclusion is not accurate.

Although he did some nasty things with the licenses, but look… FPC/Laz devs don't mind, even support him.
Are you sure they don't mind? Who supports him?

So, I think Maciej's decision to work on his own FPC fork is correct. As long as he keeps following FPC development and base his work on it, I'll support him.
Sincere question: Why do you want him to base his work on the work of "a bunch of old people"?
« Last Edit: May 09, 2018, 07:05:33 am by engkin »

hnb

  • Sr. Member
  • ****
  • Posts: 270
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #37 on: May 09, 2018, 09:20:45 am »
use the bugtracker, attach the diff of the files to a bug report and walk away, if you really want to do. If not then as long as your repo is online and accessible by the team then they can pull what ever they need.

I was thinking about using FPC bugtracker as normal user (this is how was looking my work before). What I can say : In many cases filling bug report, preparing *.patch takes more time than fix and commit to trunk. Redundant work (especially for minor fixes). Someone needs to assign bug report, apply the patch and commit (some patches waits few years !sic so filling bug reports for sport is not practical). Because of someone's "strange" decision we need to waste many of the important time. Doing this in current situation is also demotivating. Probably I should continue detailed list of changes like :

http://newpascal.org/compass.html

but without bugreports in FPC mantis.
Checkout NewPascal initiative and donate beer - ready to use tuned FPC compiler + Lazarus for mORMot project

best regards,
Maciej Izak

hnb

  • Sr. Member
  • ****
  • Posts: 270
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #38 on: May 09, 2018, 09:21:57 am »
Sincere question: Why do you want him to base his work on the work of "a bunch of old people"?

Free bug fixes for NewPascal!
Checkout NewPascal initiative and donate beer - ready to use tuned FPC compiler + Lazarus for mORMot project

best regards,
Maciej Izak

marcov

  • Administrator
  • Hero Member
  • *
  • Posts: 11383
  • FPC developer.
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #39 on: May 09, 2018, 10:10:44 am »
I was thinking about using FPC bugtracker as normal user (this is how was looking my work before). What I can say : In many cases filling bug report, preparing *.patch takes more time than fix and commit to trunk. Redundant work (especially for minor fixes). Someone needs to assign bug report, apply the patch and commit (some patches waits few years !sic so filling bug reports for sport is not practical). Because of someone's "strange" decision we need to waste many of the important time. Doing this in current situation is also demotivating. Probably I should continue detailed list of changes like :

For major work and restructures, yes this can take some while. For short patches it is significantly quicker.

valdir.marcos

  • Hero Member
  • *****
  • Posts: 1106
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #40 on: May 09, 2018, 10:14:16 am »
https://www.firebirdsql.org/en/snapshot-builds/
ftp://ftp.freepascal.org/pub/fpc/snapshot/v31/

Quote
https://www.firebirdsql.org/en/server-packages/
Seems to be regular release. We have that.
ftp://ftp.freepascal.org/pub/fpc/dist/3.0.4/

Quote
https://www.firebirdsql.org/en/development/
https://www.freepascal.org/testsuite/cgi-bin/testsuite.cgi
Any help increasing coverage and frequency welcome, but it is all basically there.


I was thinking of solutions similar to:
https://jenkins.io/
https://travis-ci.org/
https://www.gocd.org/

To easy and improve environments such as:
Whenever I need to compile a new version of my cross platform tools, I do it like this :

Copy code to OSX system. Compile and build.
Copy code to Windows virtual machine. Compile and build.
Copy code to Linux 32-bit distribution virtual machine. Compile and build.
Copy code to Linux 64-bit distribution virtual machine. Compile and build.

Upload binaries, executable and OSX app file to web server. 

As you might imagine, that is a bit of a faff!! And can be frustrating, especially if you get to the end of even half way through and realise I've forgot to update a form caption or something and so have to repeat it all!

Someone mentioned that if code is on GitHub, the build process can be automated for all three systems using continuous integration.
....
I realize in the project settings you can change the target CPU and so on, but I've never yet successfully managed to create, for example, a Linux binary on a Windows system. So I've always just done it as stated in the steps above.

Any guidance welcome.

For example if I want to get automatic development builds of a program made with lazarus with one of the CI tools available online (appveyor). There's a tutorial on how to do it?

It for example needs some packages installed in the IDE to work.

I'm intrigued since I recently seen an appveyor file in New Pascal.

And if is not possible don't mind. Is a specific question, I know I can automate that in my machine too, but I'm asking about online services.

I've been working on an fpcunit listener that writes fpcunit test results to a database instead of an XML file: handy to be able to compare results between runs and detect regressions.

My aim was to make the code cross-database friendly (using portable SQL) and normalize the database so it can be used for a lot of test results (e.g. in Continuous Integration server kicked off by SVN commits).

I've been using it with a remote Firebird database server in my Jenkins CI server for the dbtestframework tests (see http://wiki.lazarus.freepascal.org/Jenkins#FPCUnit.2FDBTestframework_tests_with_database_output)

By now, I think I've got the test listener code working well from a functional perspective and I'm now looking at improving performance without sacrificing storage size/normalization.

This hypothetical wiki documented best solution for FPC and Lazarus servers would fit and help greatly many of us.

Pascal

  • Hero Member
  • *****
  • Posts: 932
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #41 on: May 09, 2018, 10:31:05 am »
Maybe we can use Docker to setup a CI-System as starting point for own projects.
I've thought of building my own a few times already but in fact didn't start so far.
laz trunk x64 - fpc trunk i386 (cross x64) - Windows 10 Pro x64 (21H2)

marcov

  • Administrator
  • Hero Member
  • *
  • Posts: 11383
  • FPC developer.
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #42 on: May 09, 2018, 10:41:00 am »
I was thinking of solutions similar to:
https://jenkins.io/
https://travis-ci.org/
https://www.gocd.org/

To easy and improve environments such as:

Ok, so for development of applications WITH Lazarus and FPC, not development ON lazarus and fpc.

Lazarus ran Jenkins for a while, iirc  if after a commit there was no new commit in 5 minutes or so.

Personally, I think it is more stuff for teams than for single developers. Getting deep into these systems also takes time, and for one machine administered by one person a few simple scripts are easier. At least IMHO.

Of course it probably it is less of a problem if you already use some of the techniques, but I don't maintain any LAMP or alternatives (webservers, web serverside scripting languages, and I haven't touched Java seriously in a decade), both experience or maintenance, so for me that would be all extra burdens just to use one of these packages and or develop plugins for them.

Afaik Lacak who does a lot of SQLDB work, has an own testsuite.

valdir.marcos

  • Hero Member
  • *****
  • Posts: 1106
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #43 on: May 09, 2018, 10:44:09 am »
On one side, everybody agrees that both FPC and Lazarus need more and more volunteers, especially from younger generation to ensure these great projects will have a bright future. But on the other side, sometimes new people aren't very welcome to contribute if this new people doesn't conform to the seniors way of thinking. This Maciej's case is the latest example, but I've seen similar cases since I joined FPC/Lazarus.
+1

Quote
On one side, everybody knows that we need to get popular so this project could attract both volunteers and fundings, especially from younger generation. But on the other side, I don't see much effort has been done to overcome this problem. The core devs barely think about it, because they mostly think about the technical things. I don't blame them because it's their focus and concern, but if nobody would tackle and handle these non-technical problems, then we'll getting nowhere. We'll be forever a niché in software development. Nobody cares about Pascal but a bunch of old people who still maintaining their projects in their niché market.
+1

Quote
Now I think I understand why CodeTyphon chose its way. Rather than contributing directly to the project, Sternas (CT's author) simply took the project with him and do it his way so he didn't need to bother with core devs approvals and agreements. Although he did some nasty things with the licenses, but look… FPC/Laz devs don't mind, even support him. So, I think Maciej's decision to work on his own FPC fork is correct. As long as he keeps following FPC development and base his work on it, I'll support him.
So far, I have never used CT, but reading discussions on this forum, it seems that CT's author motivations were more economical than philosophical.

Quote
Maybe it's time for FPC and Lazarus to have a new "competitor" other than Delphi
I know that my opinion is not relevant here since I am neither a core developer nor a patcher, but I still think that work together is better than forking in the open source world.

Quote
Maybe it's time for FPC and Lazarus to have a new "competitor" other than Delphi so they will start to be more open minded accepting new ideas from new people.
Considering recent examples such as PHP x Hack (HHVM), PHP evolved very fast after Facebook competition had created Hack based on PHP.

Quote
If Maciej keeps continuing his work and regularly release stable New Pascal with new features that offers something more than the orginal FPC, I'm sure New Pascal would have its own users and fans. Just like CodeTyphon.
Unfortunately, you are probably right.

hnb

  • Sr. Member
  • ****
  • Posts: 270
Re: NewPascal plans, Generics.Collections and FPC trunk
« Reply #44 on: May 09, 2018, 10:45:47 am »
For major work and restructures, yes this can take some while. For short patches it is significantly quicker.
For short patches generally it works in this way, but this is not the rule. If no one from core is interested, the patch can wait long time (even short patch). If I have ban for trunk without rational reason, I think that my patch can be rejected also without rational reason also my bug report can be simple ignored with malice (what I'm afraid of). The general fact about redundant work from my side is still valid. Could you explain me the sense of redundant work? Why I am forced to fill bug report for every minor thing instead of working on more fixes?
Checkout NewPascal initiative and donate beer - ready to use tuned FPC compiler + Lazarus for mORMot project

best regards,
Maciej Izak

 

TinyPortal © 2005-2018