Sourceforge subversion problem

[TurboPT]

Andy, we will eventually remove those entries with other future commits.

However, I don't believe that having that $Id reference would makes much sense in a distributed repository? (vs. centralized, like SVN)

They way that changes can be pulled (or not) across various repositories, at different/various points, I don't know how the $Id would be useful in the 'git realm'?

However, on the flip-side of that, there is history with every file, even across forks.

Just another mindset difference, I guess?
I see that git does have the concept of tagging (kinda reminds me of CVS in a way), but I don't yet have the perspective as to how useful these might be, or not.

[phil]

I messed about with sed to delete the svn $Id lines - destroyed my repository and had to rm -r webERP the whole shooting match
Had to re-clone my suppository to get back up again:-)
Got there in the end ... but much learning!!!

sed is super powerful - I used it to remove all the $db out of DB_query calls some time ago

[TurboPT]

Doh! ... A little overSEDous were you?!

I have removed many other $db that the sed didn't catch, but most likely because many of those were not on the same line and outside the expression.

[phil]

LOL!
Yes way overSEDous ... still a battle for my old head to get around the concepts as Andy noted - it is mind altering but sadly does seem to add about 5 steps over SVN to my usual workflow - at least it seems to work.
It was just a simple

Code:

$svn commit -m 'some changes'

but now we have

Code:

$commit -am 'some changes'
$push

then it is only on my own repository and I have to go into my browser to my repository on github then create a pull request on webERP-team and pull these changes into the master on webERP-team
The pull request then has to be merged into the webERP-team/master
... then I have to pull back down into my local repository

Phew .... it seems quite convoluted!!!
Am I making a meal of this, or is this how it is?

[TurboPT]

I am not 100% sure, for I have not yet used git via the command line.

Tim might have more details on that (and perhaps some shortcuts) as he's been working with it a few years now.

[phil]

I may be wrong but think the loads of $db you removed were not related to DB_query() calls I think these were other functions that previously needed $db passed?
Thought I had got all the DB_query() ones but I stand corrected if I am wrong as happens regularly/mostly :-)

[TurboPT]

Well, there were several lingering $db with DB_query() and other DB_* related calls, so it was not only with these functions alone, but you are right that I've also removed many $db from other functions that still had $db in the parameters, but unused within the function.

There are still other functions to cover, but the SVN trouble kinda cut that short.
I'll get to them eventually, unless removed with future commits by others.

[TurboPT]

Tim,

Before the file gets too divergent...

With SVN, we were updating the Change.log file along with the associated change(s).

I've not yet done that since the git move, so before it goes too far, and easier to update now ... is it still necessary to maintain this file, or is there a way to pull a list of changes from git before releases?

[TimSchofield]

You can create a change log file using the
git log
command and doing some manipulation of the output. As I remember the last time I looked there were tons of examples online

[phil]

I've been starting to write up a bit of a how-to on git on the webERP-team git wiki

https://github.com/webERP-team/webERP/wiki

I am hopelessly under-qualified to do so - but thought I might be able to write up based on my recent learning.... only the embryo of a useful document so far. However, I am not completely sure I am correct and if someone could take a look/embrace and extend that would be great.