Puppy Linux Discussion Forum Forum Index Puppy Linux Discussion Forum
Puppy HOME page : puppylinux.com
"THE" alternative forum : puppylinux.info
 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

The time now is Wed 22 Oct 2014, 04:37
All times are UTC - 4
 Forum index » Advanced Topics » Puppy Projects » Next Puppy Development » 4.x
Pre Release Testing Methodologies for Puppy Linux
Post_new_topic   Reply_to_topic View_previous_topic :: View_next_topic
Page 2 of 4 Posts_count   Goto page: Previous 1, 2, 3, 4 Next
Author Message
Lobster
Official Crustacean


Joined: 04 May 2005
Posts: 15117
Location: Paradox Realm

PostPosted: Tue 07 Apr 2009, 06:49    Post_subject:  

Good news from Jeremy, Raffy and Warren
GIt is coming . . .
http://git.puppylinux.ca/
http://www.puppylinux.org/wiki/development/git

Chaos is what makes Puppy fun. Very Happy
Everyones help is needed with Git.

Developers want and sometimes only have time to code.
So what happens is they upload it to the forum - "here new code"

We can not force, insist, red tape programmers into some 'rule of git'
So how to facilitate?

My thoughts are along the line of 'puppy pack'
voluntary puppys who can take uploaded code and add it to the git depository
as soon as they see it appearing

Is this feasible?
because getting all developers to use git
- ain't gonna happen . . . Confused

I am going to learn git
I am going to have my own cutting edge local git repository
I might not code much but even so . . .

Any other ideas to retain the simplicity?

_________________
Puppy WIKI
Back to top
View user's profile Send_private_message Visit_website 
ecomoney


Joined: 25 Nov 2005
Posts: 2183
Location: Lincolnshire, England

PostPosted: Tue 07 Apr 2009, 07:24    Post_subject:  

Hmmm...it sounds as if GIT is some kind of way of handling the boring bits of keeping track of all the bugs/testing/etc. Ive had a quick read but Im pressed for time to get the cybercafe online at the end of the week.

Definetely interesting though Lobster thanks

_________________
Puppy Linux's Mission

Sorry, my server is down atm!
Back to top
View user's profile Send_private_message Visit_website MSNM 
ttuuxxx


Joined: 05 May 2007
Posts: 10821
Location: Ontario Canada,Sydney Australia

PostPosted: Tue 07 Apr 2009, 07:44    Post_subject:  

Guy the way I look at it, We started 4.2 as a 'fix/update release' and what we ended up with was almost a new distro, we replaced so much in such little time it was crazy, well for me it was, I was working day and night either compiling or trying to fix things, building graphics, etc endless work, The only things I would liked seen different is time, for a project of that magnitude it would of been better to have maybe double that time. Then for sure things would of been more evident and less bugs. But it was our first time out as a group and it was a learning phase. Maybe next time we could have a forum section, not subsection, but section that stands out and call it the 'Sand Box' like Barry does, there we could have sub subcategories like software, bugs etc under software we could have a section like the packagers names where they can list the software for people to look at and test. I know what your going to say, what about Git,
well the average user won't even go near Git, so that will take the bulk of testers aways. Git is good in someways and other its not.
How many of you have used Git in the past say like 12 months ago and more? I have but I bet most haven't.
ttuuxxx

_________________
http://audio.online-convert.com/ <-- excellent site
http://samples.mplayerhq.hu/A-codecs/ <-- Codec Test Files
http://html5games.com/ <-- excellent HTML5 games Smile
Back to top
View user's profile Send_private_message Visit_website 
ecomoney


Joined: 25 Nov 2005
Posts: 2183
Location: Lincolnshire, England

PostPosted: Tue 07 Apr 2009, 09:14    Post_subject:  

Good idea about the forum section Ttuxxx, Single linear "thread" for bugs and fixes can mean things get missed. A "Thread per bug" approach would make more sense, then we could put "[SQIUSHED]" or something in the title when its solved and put into the latest version, like in the "howtos".

After bugs are "squished" they would naturally descend to the bottom of the forum section, leaving a forum page of outstanding issues there when testers clicked on "sandbox" (I think we should use a more newb-friendly less jargon term but thats me). This could be a link combined with the "default alpha/beta" seamonkey homepage.

Your right there were a lot of updates to 4.2, from 4.1.2. Was it necessary to attempt do so much in one go? I missed the first part of the development process where these needs were identified so I dont know. Basically the only things that should have been updated were the things that were really needed by our users (i.e. flash, Seamonkey, themes/look and feel). Personally Im sorry that the messenger was removed, but glad transmission was put back in etc etc. At least we have meebo.

Ttuxxx, you did a LOT of work towards 4.2 (I am grateful) and its completely understandable that things got missed considering our lack of collective project management experience and the workload you had. Im glad your able to look back and reflect, and contribute to how we might be even better in the future. Thanks for posting your thoughts.

_________________
Puppy Linux's Mission

Sorry, my server is down atm!
Back to top
View user's profile Send_private_message Visit_website MSNM 
WhoDo


Joined: 11 Jul 2006
Posts: 4441
Location: Lake Macquarie NSW Australia

PostPosted: Tue 07 Apr 2009, 09:21    Post_subject:  

ecomoney wrote:
The problem with the "nike - just do it" approach is that is someone "just doesnt do it" then faults get missed i.e. Abiword. Im not sure what background you have is software development, but testing is an important part, and should always be done methodically...not just "hope for the best". If this type of work isnt to your taste, then pass it this way.


BarryK wrote:
Puppy has never been run as a "proper" project, it has always been ad-hoc. There is no CVS/SVN server, no bug-reporter. For the last couple of years, new puppies have been released every 6 - 8 weeks, a hectic pace, and some would say too hasty for proper bug catching prior to each release. Right up front I would like to state that none of this is likely to change.


Patriot wrote:
I know my active time here is only a few days old but again I reiterate, one of the reason I was attracted to puppy is because its a bit more relaxing here. Too much red tape hinders creativity and I've banged my head against it many times in the real world. It will be great to have some QA/QC methods but don't let it spoil the fun please ...


BarryK wrote:
There is the appearance of chaos in many aspects of the project, but there is intelligence behind my choice to let everything run free. Developers come in and just let their heads go -- maximum creativity. I think that the end result speaks for itself.


ecomoney wrote:
Puppy 4.2 is a vast improvement on previous puppies with just a few rough edges...you should take responsibility for it all.


BarryK wrote:
For me, Puppy is a fun project. If it stops being fun, I will move onto something else. I don't have the motivation for more formality, and I find that trying to please many people is not good for the project in the long run. So I just do whatever I please -- I consider all advice very carefully, and often do follow it, but in the end what it comes down to is I decide what to do next in Puppy. There is no committee, no inner-circle. You may advise me, but I have no obligation to do anything that you say or expect. If you think I've done something dumb, well, that's just tough.


ecomoney wrote:
Not wishing to curb your enthusiasm for moving on to even bigger and better things, puppy has always had .1 and .2 releases. Currently on our main puppylinux.org download page, users are being served with a distro with a "semi-broken" wordprocessor, as well as a few other faults. Shouldnt this be put right as soon as possible before even more new features are introduced?


BarryK wrote:
Anybody is welcome to fork the project and manage it anyway they like. The thing is though, it isn't easy, an incredible amount of commitment is required. If you like Puppy but would like to see more formal management, ... {fill in your own blanks on this one but mine would be "Nike"}


When I took control of the Puppy 4.2 development effort, one of my key concerns was to keep the Puppy ethos. Several examples of the philosophy that engendered that ethos are quoted above in response to various quotes from this thread. I have given my own reasoned arguments, on more than one occasion, as to why I will NOT be pushed into quickly releasing a dotpoint1 update. I won't reiterate those here. Rob, the answers to your questions have been posted many times. Repeating the questions multiple times, using mulitiple methods with multiple presumed justifications for doing so will NOT change the answers. Rolling Eyes

_________________
Actions speak louder than words ... and they usually work when words don't!
SIP:whodo@proxy01.sipphone.com; whodo@realsip.com
Back to top
View user's profile Send_private_message 
ecomoney


Joined: 25 Nov 2005
Posts: 2183
Location: Lincolnshire, England

PostPosted: Tue 07 Apr 2009, 10:01    Post_subject:  

Whodo, weve posted at the same time yet again? I was editing the top post to summerize positive things that people have suggested in the discussion.

Have a read of them, and tell me how doing them would spoil any of the developers fun? How would they change their ethos? It obviously works and I wouldnt want to "endanger" that.

Obviously testing isnt a fun process, but theres no lack of people willing to do it, Im not trying to force you to do it. Im just saying let others contribute here too.

Quote:
I consider all advice very carefully, and often do follow it


BarryK

_________________
Puppy Linux's Mission

Sorry, my server is down atm!
Back to top
View user's profile Send_private_message Visit_website MSNM 
Dougal


Joined: 19 Oct 2005
Posts: 2505
Location: Hell more grotesque than any medieval woodcut

PostPosted: Tue 07 Apr 2009, 15:17    Post_subject:  

WhoDo wrote:
Dougal wrote:
This isn't the first release where some menu items ended up just plainly not working..

Which ones, Dougal? In building each release that was one of the things I tried to do - click on each menu item, especially if it was an upgraded or new package, to see that it actually worked.

Well, I recall a few times in the past that we had such problems, where either really basic problems persisted all the way to the last beta, or the final was released with such "dead" menu items. It appears like it still happens...

_________________
What's the ugliest part of your body?
Some say your nose
Some say your toes
But I think it's your mind
Back to top
View user's profile Send_private_message 
Dougal


Joined: 19 Oct 2005
Posts: 2505
Location: Hell more grotesque than any medieval woodcut

PostPosted: Tue 07 Apr 2009, 15:36    Post_subject:  

ttuuxxx wrote:
Guy the way I look at it, We started 4.2 as a 'fix/update release' and what we ended up with was almost a new distro,

And that is exactly the problem (and why I don't participate in 4.2 etc.): you have no idea what you're talking about. Updating a bunch of gui apps is not what creating a new distro consists of -- that's what puplets are.
What Puppy was in a very bad need of is a new base (what Barry is doing now with T2), as the current filesystem is about 18 months old, or at least a new kernel -- so when we try and help people solve their wireless problems we wouldn't have to tell them "go try Ultrapup, it has a newer kernel". Tempestuous' time would have been much better spent just compiling a new kernel, rather than trying to backport various modules to 2.6.25.

As I mentioned on Barry's blog when he "retired": what is needed is for various people to take responsibility for various parts of Puppy -- the load needs to be spread.
No-one needs to try and do everything -- each person should concentrate on a part that interests them and try and do that as good as possible. The "coordinator" should be the one looking over everything -- and only at the level of telling the relevant persons "Oi, matey! What you got for me today?" and making sure they don't stuff things up.

All this "progress" and muggins still hasn't got the keys to the official repos!

_________________
What's the ugliest part of your body?
Some say your nose
Some say your toes
But I think it's your mind
Back to top
View user's profile Send_private_message 
ecomoney


Joined: 25 Nov 2005
Posts: 2183
Location: Lincolnshire, England

PostPosted: Tue 07 Apr 2009, 16:13    Post_subject:  

I know I didnt see you or a few of the usual contributors around for 4.2 development. Ttuxxx can test my patience too, and im known for it.

The repos still dont contain openoffice, amsn or a recent firefox, most needed by linux newbs, although the packages exist Sad

Why all this possessiveness?

_________________
Puppy Linux's Mission

Sorry, my server is down atm!
Back to top
View user's profile Send_private_message Visit_website MSNM 
MU


Joined: 24 Aug 2005
Posts: 13642
Location: Karlsruhe, Germany

PostPosted: Tue 07 Apr 2009, 16:47    Post_subject:  

4.2 is great work, it allowed the comunity to add many addons or enhancements to the 4.12 base.
It has a conceptual and artistic focus (make Puppy more appealing, look nicer, offer beautifull addons like pwidgets).

A next step - Puppy 4.3 - might be focussing the "technical core" again.

- New Kernel ( UTF-8 )
- updated Glib/Gtk
- Xorg 7.4
- test/enhance the localization, e.g. add languagepacks (german.sfs, french.sfs ...)

- Avoid adding new programs / look / gimmicks

If you concentrate on these 4 points, it should take not too much time to release it.
A final step also might update some of the applications to the newest versions.

The challenging point might be, to offer these "technical" upgrades optionally.
Version 1:
- Keep Kernel 2.6.25.16 and Xorg 7.3 in one version - e.g. to be able to use the old Nvidia 73.x drivers.
Version 2:
- Use Kernel 2.6.29.x and Xorg 7.4 to be able to use e.g. newest Intel graficschips
Version 3:
- also keep a Kernel 2.6.21.7 version, as we meanwhile have so many additional kernelmodules compiled by Dougal and Tempestous, so that even exotic old hardware still works.

Once this fundament is updated, a version 4.4 might look again in adding/replacing software.
E.G. the new glib/gtk would allow, to use the Val(a)ide as important tool for software development. You can use it already now, but it may be unstable, so currently it is the better choice to use Ultrapup to develop vala/genie programs.

Mark

_________________
my recommended links
Back to top
View user's profile Send_private_message Visit_website 
WhoDo


Joined: 11 Jul 2006
Posts: 4441
Location: Lake Macquarie NSW Australia

PostPosted: Tue 07 Apr 2009, 17:10    Post_subject:  

Dougal wrote:
WhoDo wrote:
Dougal wrote:
This isn't the first release where some menu items ended up just plainly not working..

Which ones, Dougal?

Well, I recall a few times in the past that we had such problems, where either really basic problems persisted all the way to the last beta, or the final was released with such "dead" menu items. It appears like it still happens...

Yes, that problem only surfaces in specific conditions so I wouldn't have noticed it in pre-testing just by clicking on the menu entry, as was the case for so many others who tested it in RC4. A longer testing cycle for RemaX would have helped, though, so that's my bad.

_________________
Actions speak louder than words ... and they usually work when words don't!
SIP:whodo@proxy01.sipphone.com; whodo@realsip.com
Back to top
View user's profile Send_private_message 
WhoDo


Joined: 11 Jul 2006
Posts: 4441
Location: Lake Macquarie NSW Australia

PostPosted: Tue 07 Apr 2009, 17:40    Post_subject:  

Dougal wrote:
And that is exactly the problem (and why I don't participate in 4.2 etc.): you have no idea what you're talking about.

Despite occasional appearances to the contrary, ttuuxxx wasn't running the 4.2 project; I was.

Dougal wrote:
What Puppy was in a very bad need of is a new base (what Barry is doing now with T2), as the current filesystem is about 18 months old, or at least a new kernel -- so when we try and help people solve their wireless problems we wouldn't have to tell them "go try Ultrapup, it has a newer kernel". Tempestuous' time would have been much better spent just compiling a new kernel, rather than trying to backport various modules to 2.6.25.

As I mentioned on Barry's blog when he "retired": what is needed is for various people to take responsibility for various parts of Puppy -- the load needs to be spread.

Agreed. My understanding was, from the various threads that were discussing the early development of 4.2, that you would be leading the core development part of the community effort for the 4.2 release. I coordinated the inclusion of whatever I was given, including making decisions about what was in or out. With very little input in the way of material changes to the core, except for some important changes to the init script, there was nothing for me to "coordinate" from that part of the program. Now I think I understand why.

Dougal wrote:
The "coordinator" should be the one looking over everything -- and only at the level of telling the relevant persons "Oi, matey! What you got for me today?" and making sure they don't stuff things up.

All this "progress" and muggins still hasn't got the keys to the official repos!

Yep, that's me alright ... except for being solely responsible for any stuff ups. I tried to ensure that personal projects didn't overlap, certainly, but sometimes changes in one place do break things elsewhere, as happened once or twice with localisations of key scripts. That's where the alpha, beta and rc testers come in, so I don't have to do everything from the packaging outward.

I'm pleased to note that you have recognised the focus of 4.2 was NOT to produce "a whole new distro", and your clear recognition that what we produced was indeed only a "polish", packages and usability update from 4.1.2, as I foreshadowed way back when the project started. That's why I wasn't too worried when guys like you and tempestuous didn't take a whole lot of interest in the Deep Thought project. That's also why I have been careful to bracket the next release as 4.2.1/4.3 in my postings about what will follow Deep Thought.

If there is no serious effort in the underlying core, we will have a 4.2.1 release at the appropriate time - bug fixes folded in, packages updated or rolled back as necessary, etc. No changes to look-and-feel at all. If there is some serious effort to update the core, including a new T2 build from Barry, then 4.3 will be the eventual product and that will require a significantly longer testing cycle, more involvement from core devs like you and tempestuous, and less emphasis on packages and polish. Either way, the packages will be coming from the Git repository and managed through that facility.

Why do you think muggins needs the "keys to the official repos"? If he needs something included there, that can be coordinated through me for the next release. I don't see the problem.

_________________
Actions speak louder than words ... and they usually work when words don't!
SIP:whodo@proxy01.sipphone.com; whodo@realsip.com
Back to top
View user's profile Send_private_message 
WhoDo


Joined: 11 Jul 2006
Posts: 4441
Location: Lake Macquarie NSW Australia

PostPosted: Tue 07 Apr 2009, 17:55    Post_subject:  

ecomoney wrote:
The repos still dont contain openoffice, amsn or a recent firefox, most needed by linux newbs, although the packages exist Sad

Why all this possessiveness?

The official repos do contain ayttm (not amsn) and a working firefox rather than one that remains buggy (as the 3 series), but they will NEVER contain a 250Mb office package in sfs format. That's just not what they're for. The official repos are for the packages likely to be included in Puppy by default and Open Office just doesn't fit the bill; nor does K-Office which I happen to like better and which is much smaller. In a sub-100Mb distribution it is unlikely that huge application suites will ever be found at ibiblio.org but that doesn't mean there aren't ways to make them available through the PETget package manager.

It isn't "possessiveness" either; Rolling Eyes it's "protection" of the community asset for the benefit of all. That will not change with Git, either. There will likely be only ONE person with the rights to commit changes to the packages in the main tree.

_________________
Actions speak louder than words ... and they usually work when words don't!
SIP:whodo@proxy01.sipphone.com; whodo@realsip.com
Back to top
View user's profile Send_private_message 
WhoDo


Joined: 11 Jul 2006
Posts: 4441
Location: Lake Macquarie NSW Australia

PostPosted: Wed 08 Apr 2009, 09:19    Post_subject:  

MU wrote:
4.2 is great work, it allowed the comunity to add many addons or enhancements to the 4.12 base.
It has a conceptual and artistic focus (make Puppy more appealing, look nicer, offer beautifull addons like pwidgets).

Exactly! And don't forget it was also intended to be easier to use for refugees as well. There were quite a number of relatively small usability "tweaks" to that end.

MU wrote:
A next step - Puppy 4.3 - might be focussing the "technical core" again.

- New Kernel ( UTF-8 )
- updated Glib/Gtk
- Xorg 7.4
- test/enhance the localization, e.g. add languagepacks (german.sfs, french.sfs ...)

- Avoid adding new programs / look / gimmicks

If you concentrate on these 4 points, it should take not too much time to release it.

Absolutely agree! And by concentrating in these key areas we can also test the operation of Git, and establish our "best practice" rules for its effective use at the same time, without the complication of lots of different development efforts across many packages.

MU wrote:
A final step also might update some of the applications to the newest versions.

Agreed, with the emphasis on "some" of the applications. Those that are working optimally can be left alone for that release, with only those that are working but not to the best level being updated with later releases, and then only when we have the core exactly where we want it.

MU wrote:
The challenging point might be, to offer these "technical" upgrades optionally.
Version 1:
- Keep Kernel 2.6.25.16 and Xorg 7.3 in one version - e.g. to be able to use the old Nvidia 73.x drivers.
Version 2:
- Use Kernel 2.6.29.x and Xorg 7.4 to be able to use e.g. newest Intel graficschips
Version 3:
- also keep a Kernel 2.6.21.7 version, as we meanwhile have so many additional kernelmodules compiled by Dougal and Tempestous, so that even exotic old hardware still works.

Here is where I would prefer to decomplicate our releases a bit if possible. I have been clear elsewhere that I've no in depth understanding of the kernel building process, including what is possible and what isn't with each version. Having said that, I'd really like to roll all the legacy stuff into a single kernel version. Whether that's k2.6.21.7 or k2.6.25.16 doesn't matter to me from the point of consolidation. We must continue to support legacy hardware, for the sake of our user base, so the "best" of these two kernels at supporting that user community ought to be the base and enhancements from the other rolled into that one wherever possible.

Then we can use the later, "cutting edge" 2.6.29.x kernel for high end and later edition hardware, including enabling SMP support and perhaps Quad Core as well (if that is possible - I said I didn't really know about compiling kernels Wink )

Either way, supporting any more than 2 x kernels for standard releases is just too difficult for any community based edition IMHO.

MU wrote:
Once this fundament is updated, a version 4.4 might look again in adding/replacing software.
E.G. the new glib/gtk would allow, to use the Val(a)ide as important tool for software development. You can use it already now, but it may be unstable, so currently it is the better choice to use Ultrapup to develop vala/genie programs.

I would also hope that by then Woof will be the packaging system of choice, and our devs will then be free to develop for a debian, slackware or purely puppy base using whatever environment is supported by their chosen platform. They would also be using Git to have those developed applications looking like polished, shiny apples sitting on the Git tree just waiting to be "plucked" by those eager packagers building puplets as well as the community for standard releases. If that happens as I suspect, then we will probably be looking at series 5.x Puppy from that point forward.

_________________
Actions speak louder than words ... and they usually work when words don't!
SIP:whodo@proxy01.sipphone.com; whodo@realsip.com
Back to top
View user's profile Send_private_message 
ttuuxxx


Joined: 05 May 2007
Posts: 10821
Location: Ontario Canada,Sydney Australia

PostPosted: Wed 08 Apr 2009, 09:34    Post_subject:  

Just from what I've read WhoDo Quad Core won't be part of that all in one kernel you want, due to the fact that it can cause a loop and sslow down dual core cpu's by a great deal, but quad core cpu's can use dual core OS with added speed over a single core. So basicallt dual core should be the peak kernel. But a better Idea would be having a all in one smp, version and a 64bit kernel, then if Barry was kean enough he could make a woof release where it downloads 64bit packages from say ubuntu because they offer it. It would be a extremely quick way of having a 64bit puppy version.
ttuuxxx

_________________
http://audio.online-convert.com/ <-- excellent site
http://samples.mplayerhq.hu/A-codecs/ <-- Codec Test Files
http://html5games.com/ <-- excellent HTML5 games Smile
Back to top
View user's profile Send_private_message Visit_website 
Display_posts:   Sort by:   
Page 2 of 4 Posts_count   Goto page: Previous 1, 2, 3, 4 Next
Post_new_topic   Reply_to_topic View_previous_topic :: View_next_topic
 Forum index » Advanced Topics » Puppy Projects » Next Puppy Development » 4.x
Jump to:  

Rules_post_cannot
Rules_reply_cannot
Rules_edit_cannot
Rules_delete_cannot
Rules_vote_cannot
You cannot attach files in this forum
You can download files in this forum


Powered by phpBB © 2001, 2005 phpBB Group
[ Time: 0.1412s ][ Queries: 12 (0.0080s) ][ GZIP on ]