Bug Busters

[marcov]

(Blush). Searching for a "save layout" and saving it makes it at least go away.

No multimonitor setup, just fresh installs on fullhd screens.

[Mike.Cornflake]

This week is apparently a learning week.

Just discovered that if you raise an issue in Mantis and the developers subsequently set it to resolved, you then have the ability to Close the issue (if you're happy with, and have tested, the solution).

Sigh - which means I've just closed a slew of issues. 

Sharing this as looking at the number of issues marked "resolved" I suspect I'm not the only one who didn't know about that last step...

[nightrider]

Did you solve the problem with Windows Mobile/CE screen cleaning management?

Greetings from Sao Paulo - Brazil

Ricardo

[JuhaManninen]

Did you solve the problem with Windows Mobile/CE screen cleaning management?

Who are you asking from? Some sections of Lazarus code are not very actively maintained which is a problem.
We must find a way to get patched in those areas committed, too.

[Basile B.]

Is it possible to fork on GitHub, to modify locally, to push the fork and then to propose a pull request ? or Laz and Fpc are purely svn and git only for visibility across the www ?

[JuhaManninen]

Is it possible to fork on GitHub, to modify locally, to push the fork and then to propose a pull request ? or Laz and Fpc are purely svn and git only for visibility across the www ?

To whom would you send the pull request?
There is no difference between Git or SVN in this case. You can send a "pull request" by e-mail to anybody you want to make sure your patch is notified. However it will not help if the maintainer is busy with other things. The maintainer is finally responsible for the commit and he must check and test the patch before committing it. This is the essential part of the process, the revision control system in use is irrelevant.

I know there are patches waiting for WinCE. The maintainer Felipe is busy (which is OK naturally) and I was thinking of a system where I would commit a patch without testing it (!) after I get confirmation from 2+ WinCE users that the patch works. Those people would then be "semi-developers" in Lazarus project, partly responsible of the quality.
Surely it would be better than ignoring the patches. The worst case scenario is that a bug is found later and the commit must be reverted which is not so awful either.

About Git support in Lazarus development: It is fully supported! Either create patches from your local branch:
  http://wiki.freepascal.org/Creating_A_Patch#Creating_a_patch_using_Git
or develop in a distributed manner:
  http://wiki.freepascal.org/Creating_A_Patch#Using_a_forked_Git_repository_directly
Nobody has yet used the distributed way, and I am rather happy for that. It would make things more difficult for me and for other people, compared to patches uploaded to bug tracker.
The diff format and the "patch" command are very good tools, and the bug tracker is very good for managing all the various issues and discussions about them.

[Basile B.]

Is it possible to fork on GitHub, to modify locally, to push the fork and then to propose a pull request ? or Laz and Fpc are purely svn and git only for visibility across the www ?

To whom would you send the pull request?

bah...It's quite obvious that's the PR will be made to the repository I'd initially forked.

[JuhaManninen]

bah...It's quite obvious that's the PR will be made to the repository I'd initially forked.

I don't know how it would work when the project has many subsystem maintainers.
Maybe the pull request can be assigned to a specific person in GitHub.

BTW, I personally use Git, not because of its distributed aspects but because:
1. Local branches let me maintain experimental code and new features easily without hassle.
2. Local history and a tool, Gitk, to view it. There are also features like "git bisect" made possible by the local history.

Many companies at least in here have started to use Git but they also configure a server into their system. No distributed development there either.

P.S.
Returning to the first post of this thread, we have resolved some 150 reports since then. It is a big amount. Some reports were closed because trunk already had a fix for them or because they were not valid issues, but we also got many patches and developers fixed other issues by themselves.
I think this was a success.
In future I try to make sure patches are not ignored. It gives a bad impression of the project for those few people who actually take the effort and create code for it.

[Basile B.]

bah...It's quite obvious that's the PR will be made to the repository I'd initially forked.

I don't know how it would work when the project has many subsystem maintainers.
Maybe the pull request can be assigned to a specific person in GitHub.

BTW, I personally use Git, not because of its distributed aspects but because:
1. Local branches let me maintain experimental code and new features easily without hassle.
2. Local history and a tool, Gitk, to view it. There are also features like "git bisect" made possible by the local history.

Many companies at least in here have started to use Git but they also configure a server into their system. No distributed development there either.

P.S.
Returning to the first post of this thread, we have resolved some 150 reports since then. It is a big amount. Some reports were closed because trunk already had a fix for them or because they were not valid issues, but we also got many patches and developers fixed other issues by themselves.
I think this was a success.
In future I try to make sure patches are not ignored. It gives a bad impression of the project for those few people who actually take the effort and create code for it.

It's almost that. So let say, for a reason or another, I preafer GH. So
- I fork https://github.com/graemeg/lazarus (which seems to be the offical stuff)
- I modify some stuffs
- I make a PR

Is it working or not ? is this repository has a kind of credit, a kind of interest ? Can I contribute, as a "noboby", using this way ? (fork-myownthings-pull request) ?

[JuhaManninen]

Is it working or not ? is this repository has a kind of credit, a kind of interest ? Can I contribute, as a "noboby", using this way ? (fork-myownthings-pull request) ?

No it does not work.
Please again read this:
  http://wiki.freepascal.org/Creating_A_Patch#Creating_a_patch_using_Git
and then upload your patches to a bug report.

[Edit] and more...
> - I fork https://github.com/graemeg/lazarus (which seems to be the offical stuff)

Graeme has kindly put up a mirror repository but it is not really "official".
Still, I have promised to support a Lazarus fork in GitHub directly (see the wiki). Then you must communicate with me by e-mails, personal messages, whatever, but not by pull requests.
I made this promise when there was a group of whiners complaining about the "poor development tools" used by Lazarus project. According to them Git especially would make a big difference for everything. Well, Git is supported now but I don't see any big difference except that those whiners are gone. Good.
So, you can use the GitHub fork in any case but IMO creating patches from it is the easier way for both of us.

[wp]

I think this was a success.

Absolutely. Amazing!

In future I try to make sure patches are not ignored.

I am concerned also about ccr where many projects are in a desolate state. Maintainers have left their activities, but things have changed, fpc has changed, Lazarus has changed, and lot of the material is no longer working. I almost gave up posting patches for ccr because nobody takes care of them. I have a couple of patches hanging around, but there are also patches provided by others to which I added a confirming note that they are working. But of course, since the maintainer left the project, nothing happens.

[Mike.Cornflake]

I think this was a success.

Absolutely. Amazing!

+1

[JuhaManninen]

I am concerned also about ccr where many projects are in a desolate state. Maintainers have left their activities, but things have changed, fpc has changed, Lazarus has changed, and lot of the material is no longer working. I almost gave up posting patches for ccr because nobody takes care of them. I have a couple of patches hanging around, but there are also patches provided by others to which I added a confirming note that they are working. But of course, since the maintainer left the project, nothing happens.

Yes, this is a problem.
Are there patches for CCR in the bug tracker? Which issue numbers?

[wp]

Are there patches for CCR in the bug tracker? Which issue numbers?

I added a tag "patch" to ccr items back to 2013. Not sure if all these patches are working, but the following ones definitely are:

0026037: Checkbox and radiobuttons in spkToolbar pixelated at 120 ppi
0025732: Memory Leak when use Actions on Buttons in SpkToolbar
0026590: Cannot click tree items if tree is scrolled down

I know that ccr developers do have access to the entire svn tree, but should modify only code in "their" own project. How about giving each ccr developer the permission to apply a patch in someone else's project, maybe after a grace-period of, say, 3 months?

[JuhaManninen]

I know that ccr developers do have access to the entire svn tree, but should modify only code in "their" own project. How about giving each ccr developer the permission to apply a patch in someone else's project, maybe after a grace-period of, say, 3 months?

Yes, I don't believe anybody has objections for that. Feel free to commit code there.

I had missed the patches because I have Mantis project "Lazarus" selected most of the time.
I assigned SpkToolbar issues now for myself and marked myself as maintainer in wiki. Luiz Americo told already long ago he has too many things to maintain.
Apparently I don't have commit access to the repo. I will commit those patches when I get it.

Most CCR components have neither a wiki page nor a maintainer. It is bad.
I updated the wiki page so that all comps are at least listed. I will write another message about the situation.