Google Code Page for Puppy Development

Message

Aitch · #21 Post by **Aitch** » Sat 10 Oct 2009, 18:36

Looks good to me, PG

Built-in growth potential, as often asked for, AFAIK

Aitch

mikeb · #22 Post by **mikeb** » Sat 10 Oct 2009, 19:17

Good stuff....2 tiny points...
If you have a googlemail.com ident you can log in with gmail.com...redirects mail too....for the dyslexic amongst us <---- yours truly.
You can use their jabber based messenger network..might come in handy.

mike

technosaurus · #23 Post by **technosaurus** » Sat 10 Oct 2009, 21:48

I agree with pizzasgood about the structure. It should allow for easy review, collaborative development and growth.

Furthermore:
|--branches
...|--woof
...|--multiuser_woof
...|--2.X
............

aarf · #24 Post by **aarf** » Sun 11 Oct 2009, 04:58

from your Archrival http://flyspray.org/

mcalex · #25 Post by **mcalex** » Sun 11 Oct 2009, 07:13

whatis:

http://code.google.com/p/puppy-linux/

disciple · #26 Post by **disciple** » Sun 11 Oct 2009, 07:21

IME google code is a whole lot less annoying to use than launchpad or any of the other major code hosts

BTW I'm using google code without a gmail account (although of course it still means Google have my details). If I try to login at Gmail, it says I don't have an account for it.
My account name actually shows up as my email address, which isn't ideal, although I don't know if it needs to be that way.

technosaurus · #27 Post by **technosaurus** » Sun 11 Oct 2009, 15:06

In an ideal world I would prefer to use egroupware, because it has everything we need... except for a version control system. This forum has been acting as our somewhat inefficient version control system and we all know the effect that has had.

Aitch · #28 Post by **Aitch** » Sun 11 Oct 2009, 16:52

Techno

Is there going to be a user voting priority for most wanted tasks, or dev assigned type, based on ability/availability, as I think clearing this up early will keep it cool, and stop jibes about it not being how some imagine it will be?
Maybe a combination of both?
I, too, think version control essential, for reason stated

Aitch

puppyite · #29 Post by **puppyite** » Sun 11 Oct 2009, 17:55

Aitch wrote:Techno

Is there going to be a user voting priority for most wanted tasks, or dev assigned type, based on ability/availability, as I think clearing this up early will keep it cool, and stop jibes about it not being how some imagine it will be?
Maybe a combination of both?
I, too, think version control essential, for reason stated

Aitch

Good thinking Aitch. IMO your suggestions should enhance the perceived responsiveness of the bug tracker among users.

droope · #30 Post by **droope** » Mon 12 Oct 2009, 21:56

Hi there,

Google is (besides being an non-finaluser-annoying corp) very stable, and I have almost never had a problem accessing my email, nor google.

I'll start getting used to the GUI, even though I am not a developer.

Excelent!

Cheers!
Droope

#31 Post by **Pizzasgood** » Tue 13 Oct 2009, 17:06

From http://www.murga-linux.com/puppy/viewto ... 864#350864:

Lobster wrote:Reporting bugs - easy
checking out code - did not seem easy
- in fact I could add a comment to code but not check out
this was with Jeremys code (Pizzasgood)
- eh did you lock it so crazy cructaceans can not check it out
or am I just incompetent? (probably incompetent)

If I wanted to checkout the trunk of the Edit-SFS program, I would do this:

Code: Select all

svn checkout http://puppy-development.googlecode.com/svn/trunk/edit-sfs/trunk edit-sfs_trunk

That would create the directory edit-sfs_trunk on my local filesystem, containing everything in the /trunk/edit-sfs/trunk/ directory of the repository. If I wanted the entire Edit-SFS project, I would do this:

Code: Select all

svn checkout http://puppy-development.googlecode.com/svn/trunk/edit-sfs edit-sfs

That would create the directory edit-sfs on my local filesystem, containing everything from /trunk/edit-sfs/.

Both of those examples are for people who don't intend to commit changes back to SVN, i.e. users or package makers who just want to grab the latest version. Developers who intend to commit changes can use very similar commands, except that instead of http:// they would use https://, and they would also need to add the --username youremail@gmail.com option. So, if I were a developer with an address of amadeus@gmail.com, I would do this:

Code: Select all

svn checkout https://puppy-development.googlecode.com/svn/trunk/edit-sfs edit-sfs --username amadeus@gmail.com

That would then prompt me for my googlecode password, which google generates randomly for you and is obtainable from your googlecode profile, I believe. You should only need to do that once. From there out, unless you've configured SVN otherwise, it will remember your name and password when working with that working tree, and possibly the entire Puppy SVN repo (not sure about that part, I haven't used SVN much since spring).

People should observe some form of etiquette when committing to somebody else's project, such as making their own branch (in /trunk/<project-name>/branches/) so they don't step on anybody's toes, unless they've talked to those in charge of that particular project and gotten permission to commit to the trunk or official branches. Different people and projects might have different workflows, so contributors should also be aware of that. For example, on Edit-SFS I'm just doing all the work in the trunk, and occasionally I might tag a release version. For more complex projects, I might prefer to use a development branch and consider the trunk the stable branch, so that at any time somebody could checkout the latest snapshot of the trunk and have a working, if unfinished, version of the code.

technosaurus · #32 Post by **technosaurus** » Tue 13 Oct 2009, 17:31

I've been thinking of how to organize this also, not just for source code but also for package maintenance.

---[Extras (pinstall.sh, .desktop, pixmaps, additional scripts.... in a proper tree so it can be directly copied)

---[src (for puppy projects or non compiled programs) Outside projects would have a file with a link to the main project

...Notes_Changelog (best configure options, files to add/modify/delete after compile, other nuances,...)

...config file (contains the compile options used to best compile the program)

...Latest_stable_pet(s)??<-- How fast would we use up 1GB - Google would likely increase the limit but....still maybe just a link

What else is missing? I would like to build a tree with the basic stubs already there.

#33 Post by **Pizzasgood** » Tue 13 Oct 2009, 17:50

Something to be wary of: SVN likes to place a hidden .svn/ directory inside of every directory in the repository. So if you set up a tree expecting to literally copy/paste it into a build environment or compress it into a package, you need to either use something like 'svn dump' (I think, haven't looked at it in a while), or else run a script on a copy of the data to remove all .svn/ directories (don't do that to the originals in the SVN working dir or you'll break it).

That's one of the reasons I organized the Edit-SFS project as I did (no directories, and a Makefile to put everything together) even though it doesn't actually have anything to compile. When I want a directory tree I can just run make dir and it appears, nice and clean. I can use make pet to make a .pet. I can use make clean to clear all that out and restore me to a clean SVN checkout.

Also, IIRC SVN doesn't work with device nodes, and probably not with permissions, so if you make a package that contains dev nodes or files that need special permissions, you'll have to have the pinstall.sh script create/set them or something.

technosaurus · #34 Post by **technosaurus** » Tue 13 Oct 2009, 18:27

Good pointer/hack PG. Can't you also use tar to exclude .svn somehow when making a tarball? I don't recall the option. I will definitely have to test the permissions though - That would not be good. (If so, your Makefile option should get around permissions also, assuming appropriate chmod is set up in the Makefile, correct?)

#35 Post by **Pizzasgood** » Tue 13 Oct 2009, 19:09

tar --exclude-vcs will do the trick. I didn't know about that. Cool.

You can also use rsync instead of cp when copying things to filter out directories matching a pattern, but you have to be a lot more careful about how you write the target and destination names with rsync (it attaches significance to the presence of a trailing / to determine how you want to copy things (renaming, copy into, copy over, etc.)).

I just did some quick testing, and apparently SVN will track the execute bit when a file is added but not after the file exists. The other bits were all ignored.

Aitch · #36 Post by **Aitch** » Wed 14 Oct 2009, 01:14

here's a link needs some comment, please

Many years ago, IBM inspector asked me for our QC manual to approve us as a vendor. We had a copy from Transitron that we have adopted.

Today, Puppy is trying to standardize their filename creation without a QC manual as a reference.

So, its time for Ted T'so to get a copy of QC manual from IBM to run Linux coding. Its never too late, but Linux is a business, you have to run it as a professional.

All these newbies writing codes have no idea even, on how to submit codes to a maintainer? A QC manual will help to have traceability and inspector signature to assure quality of software?

http://www.tuxmachines.org/node/35889

Aitch

mikeb · #37 Post by **mikeb** » Sun 29 Nov 2009, 14:08

Erm no issues on the bugtracker and I had another I had found.....what's happening

mike

technosaurus · #38 Post by **technosaurus** » Mon 30 Nov 2009, 21:33

The direct link still works... The wiki page for "issues" overrides the issue tracker list. This is a good place for a "how to file a good bug report" message. I made a quick temporary edit that points to the actuall issue list

mikeb · #39 Post by **mikeb** » Mon 30 Nov 2009, 21:48

I made a quick temporary edit that points to the actuall issue list

thanks

mike

(old)Puppy Linux Discussion Forum

(old)Puppy Linux Discussion Forum

Google Code Page for Puppy Development