gambas-devel — gambas development mailing-list

[Gambas-devel] Git merge request for gb.db.odbc.

[First L. <d4t...@gm...>]

Benoît,

Yesterday I forked gambas' master and edited *main.c* in *gb.db.odbc*:Added
and changed several things I was accumulating through time.
Made a merge request (still pending, and now 3 commits back) on gambas
master for my patch, don't know if I did it right.
If not correct, please led me into knowing how to do it right, so I know
for next time.

Regards,
zxMarce.

Re: [Gambas-devel] Fwd: Cloned and edited. How do I get the changes staged, committed, and then uploaded to GitLab?

[PICCORO M. L. <mck...@gm...>]

2017-08-21 11:35 GMT-04:00 Tobias Boege <ta...@gm...>:

> git lets you go as far as committing only *parts of certain files*, with
>
its true, i forgot that!

i can commit partially set of changes in the same archive, and revert part
of them


> git add -i. Remember: git is the stupid content tracker.
>
> Regards,
> Tobi
>
> --
> "There's an old saying: Don't change anything... ever!" -- Mr. Monk
>
> ------------------------------------------------------------
> ------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Gambas-devel mailing list
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gambas-devel
>

Re: [Gambas-devel] Fwd: Cloned and edited. How do I get the changes staged, committed, and then uploaded to GitLab?

[Tobias B. <ta...@gm...>]

On Mon, 21 Aug 2017, d4t...@gm... wrote:
> I also wonder why I had to GIT ADD a file that was cloned, tho. Shouldn't Git already know the file WAS modded and that it HAD to be committed? Anyway I didn't commit it signed so I clearly messed up.
> 

That's a feature :-) Imagine you have worked furiously all day and in the
evening you realise that you did lots things which you'd rather break up
into smaller logical pieces and commit them individually.

If you assume that it is desirable to be able to revert commits, then you
would want a commit to do just one thing. If you made changes to the IDE
and gb.db.odbc in the same commit, simply reverting the commit because you
messed up in the IDE would undo your changes to ODBC as well. Someone
correcting your mistake would have to separate your changes and just undo
the IDE-related part by hand.

git lets you go as far as committing only *parts of certain files*, with
git add -i. Remember: git is the stupid content tracker.

Regards,
Tobi

-- 
"There's an old saying: Don't change anything... ever!" -- Mr. Monk

Re: [Gambas-devel] Fwd: Cloned and edited. How do I get the changes staged, committed, and then uploaded to GitLab?

[<d4t...@gm...>]

Bruce,

Thanks a bunch for taking the little time you had to explain things. Funny and educational! So, looking some further mails it looks like THIS is THE WAY to submit.

Since I'm new to Git, I'll hold for part II of your instructions. 

I wonder, though, if I broke anything by doing what I already did. Nevermind, I'll clone again (after I save the code I modded).

I also wonder why I had to GIT ADD a file that was cloned, tho. Shouldn't Git already know the file WAS modded and that it HAD to be committed? Anyway I didn't commit it signed so I clearly messed up.

Regards,
zxMarce.



On Aug 21, 2017, 03:25, at 03:25, "ada...@gm..." <ada...@gm...> wrote:
>Try something along the lines of this
>
>	$ <edit/compile/test>	
>	$ cd [your local repository] <1>
>	$ git add * <2>
>	$ git commit -a -s <3>
>	$ git format-patch master <4>
>	$ git send-email --to="<person> <em...@ex...>" 00*.patch <5>
>
>This is what is known colloquially as a "git WorkFlow". In other words
>it's all the stuff that needs to be done to make some "other stuff"
>happen.
>
>The <?> bits are explanations of what is happening here. Don't type
>them.
>The [?] bits are needed see below.
>
><edit/compile/test> As in use the IDE, whatever, to make your changes.
>Try them out, fix, etc, etc. This is not really a command line but is
>part of what you do "every-day", so to speak. Nor is it anything to do
>with git.
>
><1> OK, so the fix is looking good. Now we want to get into "git" mode.
>(This was the hardest thing I had to come to terms with. git is
>actually just a "source code manager" aka "SCM". Some may wish to call
>it a "VCS" (Version Control Manager) which it is - if and only if you
>are the administrator of the pure source base. If you are as old as I
>am perhaps the following may jog some memories "PCVS", "PVCS", etc.)
>"git", as in your locally installed "git" only works at the command
>line level with the current working directory, i.e. what you get with $
>pwd
>So,
>[your local repository] is the directory where you cloned the original
>gambas source down to. 
>We are now somewhere where "git" commands can be used.
>
><2> This is only "really" necessary if you have added any new files to
>your local repository. File deletions and modifications to existing
>files are taken care of automagically in the following. You could also
>use $ git add [my new file] to tell git about a specific new file, but
>"git add *" does take care of any of those little things we may have
>forgotten.  And by the way it doesn't do any damage. If there are no
>new files then it will tell you so. If there are some then it will tell
>you that anyway.
> 
>So, where are we at now.  In short, "git" now knows that there is new
>"stuff" in your local repository.  This "stuff" is in the files that
>EXIST in your local repository.  In other words, if you were to open
>one of those files in a text editor, then what you would see is the
>file exactly as you left it. BUT we now have to start thinking in "git"
>mode.  ...
>
>How can I state this. ... ah, bugger political correctness.
>"git" is like your 16 year old son who can't find his football boots on
>Saturday morning, but he is "sure" they "should" be exactly where he
>left them last Saturday afternoon (i.e. in the middle of the laundry
>floor).
>So we have to tell him, "No, I cleaned them, polished them and put them
>back in your cupboard last Sunday." 
>Likewise, we have to tell "git" the same type of thing.  Hey "git",
>your boots are back in the supboard!
>This is called a "commit". 
>So now we know what it is we will 
>
><3> Then make a commit to your local repository. 
>	-a means that all (sorry, ALL) your local changes will be committed to
>your local repository. 
>-s means that the commit will be signed with your (local or "global"
>email name and address)
>	
>So, now the "son", sorry the "git" now knows that you are serious. In
>it's tiny little mind it compares it's idea of what the file should
>look like to what it actually looks like and comes up with a set of
>"differences".  Sort of like "son" comparing his mental image of the
>laundry floor and the actual contents of the cupboard... thinks "Hmm,
>that's strange, the laundry fairies must have moved them?"
>Which is why the -s is kind of important.
>
><4> 
>
>Woops! I gotta go do stuff.
>I'll finish this ;ater.
>
>b
>-- 
>B Bruen <ada...@gn... (sort of)>
>
>------------------------------------------------------------------------------
>Check out the vibrant tech community on one of the world's most
>engaging tech sites, Slashdot.org! http://sdm.link/slashdot
>_______________________________________________
>Gambas-devel mailing list
>Gam...@li...
>https://lists.sourceforge.net/lists/listinfo/gambas-devel

Re: [Gambas-devel] Fwd: Cloned and edited. How do I get the changes staged, committed, and then uploaded to GitLab?

[<ada...@gm...>]

On Mon, 21 Aug 2017 02:35:40 -0400
PICCORO McKAY Lenz <mck...@gm...> wrote:

> too many words..
> 
> zxMarce has only one problem.. dont have access to git repository ..
> 

No, picorro, you don't understand.
YOU Have much experience with git.
WE don't.

Look at it from the other side of the fence!


-- 
B Bruen <ada...@gn... (sort of)>

Re: [Gambas-devel] Fwd: Cloned and edited. How do I get the changes staged, committed, and then uploaded to GitLab?

[PICCORO M. L. <mck...@gm...>]

Adrien the "CONTRIBUTE" file also must be at the git repository, as all the
git repository suggest..

please added to that!

the "alternate" was due the transition from SVN

Lenz McKAY Gerardo (PICCORO)
http://qgqlochekone.blogspot.com

2017-08-21 2:43 GMT-04:00 Adrien Prokopowicz <adr...@gm...>:

> Le Mon, 21 Aug 2017 08:35:40 +0200, PICCORO McKAY Lenz <
> mck...@gm...> a écrit:
>
> too many words..
>
> zxMarce has only one problem.. dont have access to git repository ..
>
> Benoit must added it.. (later) and then he can commit normally
>
> the alternate way its the cited workflow, lest resume:
>
>    1. fork the repo to your account: zxMarce (will be then at
>    https://gitlab.com/zxMarce/gambas.git/)
>    2. clone: git clone https://gitlab.com/zxMarce/gambas.git gambasdevel
>    3. cd gambasdevel
>    4. edit gb.db/odbc/gb.db.odbc/src/main.c
>    5. git commit -a
>    6. git push
>    7. then now in gitlab interface made a pull request to the gambas repo
>
> Everyone who wants to contribute shouldn't have write access to the
> repository !
>
> The fork / merge request workflow is how you should contribute (not an
> "alternate
> way"), if everybody had write access it would be quite a mess !
>
> I am writing a "how to contribute" guide that will be uploaded to the wiki
> right
> now, explaining in detail how to use Git and GitLab to make contributions
> without
> having write access to the main repository.
>
> --
> Adrien Prokopowicz
>
> ------------------------------------------------------------
> ------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Gambas-devel mailing list
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gambas-devel
>
>

Re: [Gambas-devel] Fwd: Cloned and edited. How do I get the changes staged, committed, and then uploaded to GitLab?

[Adrien P. <adr...@gm...>]

Le Mon, 21 Aug 2017 08:35:40 +0200, PICCORO McKAY Lenz  
<mck...@gm...> a écrit:

> too many words..
> zxMarce has only one problem.. dont have access to git repository ..
> Benoit must added it.. (later) and then he can commit normally
>
> the alternate way its the cited workflow, lest resume:
> fork the repo to your account: zxMarce (will be then at  
> https://gitlab.com/zxMarce/gambas.git/)
> clone: git clone https://gitlab.com/zxMarce/gambas.git gambasdevel
> cd gambasdevel
> edit gb.db/odbc/gb.db.odbc/src/main.c
> git commit -a
> git push
> then now in gitlab interface made a pull request to the gambas repo
Everyone who wants to contribute shouldn't have write access to the  
repository !

The fork / merge request workflow is how you should contribute (not an  
"alternate
way"), if everybody had write access it would be quite a mess !

I am writing a "how to contribute" guide that will be uploaded to the wiki  
right
now, explaining in detail how to use Git and GitLab to make contributions  
without
having write access to the main repository.

-- 
Adrien Prokopowicz

Re: [Gambas-devel] Fwd: Cloned and edited. How do I get the changes staged, committed, and then uploaded to GitLab?

[PICCORO M. L. <mck...@gm...>]

too many words..

zxMarce has only one problem.. dont have access to git repository ..

Benoit must added it.. (later) and then he can commit normally

the alternate way its the cited workflow, lest resume:

   1. fork the repo to your account: zxMarce (will be then at
   https://gitlab.com/zxMarce/gambas.git/)
   2. clone: git clone https://gitlab.com/zxMarce/gambas.git gambasdevel
   3. cd gambasdevel
   4. edit gb.db/odbc/gb.db.odbc/src/main.c
   5. git commit -a
   6. git push
   7. then now in gitlab interface made a pull request to the gambas repo



Lenz McKAY Gerardo (PICCORO)
http://qgqlochekone.blogspot.com

2017-08-21 2:25 GMT-04:00 ada...@gm... <ada...@gm...>:

> Try something along the lines of this
>
>         $ <edit/compile/test>
>         $ cd [your local repository] <1>
>         $ git add * <2>
>         $ git commit -a -s <3>
>         $ git format-patch master <4>
>         $ git send-email --to="<person> <em...@ex...>" 00*.patch <5>
>
> This is what is known colloquially as a "git WorkFlow". In other words
> it's all the stuff that needs to be done to make some "other stuff" happen.
>
> The <?> bits are explanations of what is happening here. Don't type them.
> The [?] bits are needed see below.
>
> <edit/compile/test> As in use the IDE, whatever, to make your changes. Try
> them out, fix, etc, etc. This is not really a command line but is part of
> what you do "every-day", so to speak. Nor is it anything to do with git.
>
> <1> OK, so the fix is looking good. Now we want to get into "git" mode.
> (This was the hardest thing I had to come to terms with. git is actually
> just a "source code manager" aka "SCM". Some may wish to call it a "VCS"
> (Version Control Manager) which it is - if and only if you are the
> administrator of the pure source base. If you are as old as I am perhaps
> the following may jog some memories "PCVS", "PVCS", etc.) "git", as in your
> locally installed "git" only works at the command line level with the
> current working directory, i.e. what you get with $ pwd
> So,
> [your local repository] is the directory where you cloned the original
> gambas source down to.
> We are now somewhere where "git" commands can be used.
>
> <2> This is only "really" necessary if you have added any new files to
> your local repository. File deletions and modifications to existing files
> are taken care of automagically in the following. You could also use $ git
> add [my new file] to tell git about a specific new file, but "git add *"
> does take care of any of those little things we may have forgotten.  And by
> the way it doesn't do any damage. If there are no new files then it will
> tell you so. If there are some then it will tell you that anyway.
>
> So, where are we at now.  In short, "git" now knows that there is new
> "stuff" in your local repository.  This "stuff" is in the files that EXIST
> in your local repository.  In other words, if you were to open one of those
> files in a text editor, then what you would see is the file exactly as you
> left it. BUT we now have to start thinking in "git" mode.  ...
>
> How can I state this. ... ah, bugger political correctness.
> "git" is like your 16 year old son who can't find his football boots on
> Saturday morning, but he is "sure" they "should" be exactly where he left
> them last Saturday afternoon (i.e. in the middle of the laundry floor).
> So we have to tell him, "No, I cleaned them, polished them and put them
> back in your cupboard last Sunday."
> Likewise, we have to tell "git" the same type of thing.  Hey "git", your
> boots are back in the supboard!
> This is called a "commit".
> So now we know what it is we will
>
> <3> Then make a commit to your local repository.
>         -a means that all (sorry, ALL) your local changes will be
> committed to your local repository.
>         -s means that the commit will be signed with your (local or
> "global" email name and address)
>
> So, now the "son", sorry the "git" now knows that you are serious. In it's
> tiny little mind it compares it's idea of what the file should look like to
> what it actually looks like and comes up with a set of "differences".  Sort
> of like "son" comparing his mental image of the laundry floor and the
> actual contents of the cupboard... thinks "Hmm, that's strange, the laundry
> fairies must have moved them?"
> Which is why the -s is kind of important.
>
> <4>
>
> Woops! I gotta go do stuff.
> I'll finish this ;ater.
>
> b
> --
> B Bruen <ada...@gn... (sort of)>
>
> ------------------------------------------------------------
> ------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Gambas-devel mailing list
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gambas-devel
>

Re: [Gambas-devel] Fwd: Cloned and edited. How do I get the changes staged, committed, and then uploaded to GitLab?

[<ada...@gm...>]

Try something along the lines of this

	$ <edit/compile/test>	
	$ cd [your local repository] <1>
	$ git add * <2>
	$ git commit -a -s <3>
	$ git format-patch master <4>
	$ git send-email --to="<person> <em...@ex...>" 00*.patch <5>

This is what is known colloquially as a "git WorkFlow". In other words it's all the stuff that needs to be done to make some "other stuff" happen.

The <?> bits are explanations of what is happening here. Don't type them.
The [?] bits are needed see below.

<edit/compile/test> As in use the IDE, whatever, to make your changes. Try them out, fix, etc, etc. This is not really a command line but is part of what you do "every-day", so to speak. Nor is it anything to do with git.

<1> OK, so the fix is looking good. Now we want to get into "git" mode.  (This was the hardest thing I had to come to terms with. git is actually just a "source code manager" aka "SCM". Some may wish to call it a "VCS" (Version Control Manager) which it is - if and only if you are the administrator of the pure source base. If you are as old as I am perhaps the following may jog some memories "PCVS", "PVCS", etc.) "git", as in your locally installed "git" only works at the command line level with the current working directory, i.e. what you get with $ pwd
So,
[your local repository] is the directory where you cloned the original gambas source down to. 
We are now somewhere where "git" commands can be used.

<2> This is only "really" necessary if you have added any new files to your local repository. File deletions and modifications to existing files are taken care of automagically in the following. You could also use $ git add [my new file] to tell git about a specific new file, but "git add *" does take care of any of those little things we may have forgotten.  And by the way it doesn't do any damage. If there are no new files then it will tell you so. If there are some then it will tell you that anyway.
 
So, where are we at now.  In short, "git" now knows that there is new "stuff" in your local repository.  This "stuff" is in the files that EXIST in your local repository.  In other words, if you were to open one of those files in a text editor, then what you would see is the file exactly as you left it. BUT we now have to start thinking in "git" mode.  ...

How can I state this. ... ah, bugger political correctness.
"git" is like your 16 year old son who can't find his football boots on Saturday morning, but he is "sure" they "should" be exactly where he left them last Saturday afternoon (i.e. in the middle of the laundry floor).
So we have to tell him, "No, I cleaned them, polished them and put them back in your cupboard last Sunday." 
Likewise, we have to tell "git" the same type of thing.  Hey "git", your boots are back in the supboard!
This is called a "commit". 
So now we know what it is we will 

<3> Then make a commit to your local repository. 
	-a means that all (sorry, ALL) your local changes will be committed to your local repository. 
        -s means that the commit will be signed with your (local or "global" email name and address)
	
So, now the "son", sorry the "git" now knows that you are serious. In it's tiny little mind it compares it's idea of what the file should look like to what it actually looks like and comes up with a set of "differences".  Sort of like "son" comparing his mental image of the laundry floor and the actual contents of the cupboard... thinks "Hmm, that's strange, the laundry fairies must have moved them?"
Which is why the -s is kind of important.

<4> 

Woops! I gotta go do stuff.
I'll finish this ;ater.

b
-- 
B Bruen <ada...@gn... (sort of)>

[Gambas-devel] Fwd: Cloned and edited. How do I get the changes staged, committed, and then uploaded to GitLab?

[First L. <d4t...@gm...>]

Ok, no luck. Looks like I have to be authorized first (which sounds
reasonable), but don't know who to ask for auth. This is the output I got
for GIT PUSH:

  Username for 'https://gitlab.com': zxMarce
  Password for 'https://zx...@gi...':
  remote: HTTP Basic: Access denied
  fatal: Authentication failed for 'https://gitlab.com/gambas/gambas.git/'

So, cannot upload today. When someone sees this maybe I get the green light.

Regards,
zxMarce.

On Sun, Aug 20, 2017 at 9:36 PM, PICCORO McKAY Lenz <mck...@gm...>
wrote:

> you forgot the "git add gb.db.odbc/src/main.c" before git commit
>
> its like the commit in git are to "mark what be uploaded to repository"
>
> so then:
>
> git clone https://gitlab.com/gambas/gambas.git gambasdevel
> cd gambasdevel
> git config --global user.email "zxarce-email"
> git config --global user.name "zxMarce"
> editor gb.db.odbc/src/main.c
> git add gb.db.odbc/src/main.c
> git commit -m "[gb.odbc] solve situation of odbc, to the great piccoro
> that really need this artgggg"
> git push
>
> this are in conclusion.. if have any other comment please feel free to
> mail me..
>
>
> Lenz McKAY Gerardo (PICCORO)
> http://qgqlochekone.blogspot.com
>
> 2017-08-20 18:34 GMT-04:00 First Last <d4t...@gm...>:
>
>> Guys, grew a pair and did the following: Got my account (zxMarce) at
>> GitLab. Created and added the necessary SSH Keys.
>>
>> Then, cloned with "git clone https://gitlab.com/gambas/gambas.git
>> gambasdevel".
>> Navigated to "gb.db.odbc/src" and changed the old "main.c" with my
>> changed file from ye olde SF SVN tree (kept changes by bgermann from 3 days
>> ago).
>>
>> Then, I guessed I had to commit my changes to my local repo, so I tried
>> "git commit". But Git surprised me by asking my for name and email (wow!
>> heavier than the NSA!). I supplied my GitLab username and email, then tried
>> to commit again, but could not due to changes not staged (?!). The exchange
>> was thus:
>>
>>    ~/gambasdevel$ git commit
>>
>>    *** Please tell me who you are.
>>
>>    Run
>>
>>      git config --global user.email "yo...@ex..."
>>      git config --global user.name "Your Name"
>>
>>    to set your account's default identity.
>>    Omit --global to set the identity only in this repository.
>>
>>    fatal: unable to auto-detect email address (got 'xxx@yyy.zzz')
>>    ~/gambasdevel$ git config --global user.email "d4t...@gm..."
>>    ~/gambasdevel$ git config --global user.name "zxMarce"
>>    ~/gambasdevel$ git commit
>>    On branch master
>>    Your branch is up-to-date with 'origin/master'.
>>    Changes not staged for commit:
>>        modified:   gb.db.odbc/src/main.c
>>
>>    Untracked files:
>>        gb.db.odbc/src/ml_doit.sh
>>
>>    no changes added to commit
>>
>> The untracked .sh is a script I made myself to quickly compile and
>> install the ODBC component. It should remain so.
>> But I don't know how to "stage" and then "commit" my changed file.
>> Then, I will have to deal with change-logging and uploading, but I'd like
>> to go "baby steps".
>>
>> TIA,
>> zxMarce.
>>
>

Re: [Gambas-devel] Cloned and edited. How do I get the changes staged, committed, and then uploaded to GitLab?

[PICCORO M. L. <mck...@gm...>]

you forgot the "git add gb.db.odbc/src/main.c" before git commit

its like the commit in git are to "mark what be uploaded to repository"

so then:

git clone https://gitlab.com/gambas/gambas.git gambasdevel
cd gambasdevel
git config --global user.email "zxarce-email"
git config --global user.name "zxMarce"
editor gb.db.odbc/src/main.c
git add gb.db.odbc/src/main.c
git commit -m "[gb.odbc] solve situation of odbc, to the great piccoro that
really need this artgggg"
git push

this are in conclusion.. if have any other comment please feel free to mail
me..


Lenz McKAY Gerardo (PICCORO)
http://qgqlochekone.blogspot.com

2017-08-20 18:34 GMT-04:00 First Last <d4t...@gm...>:

> Guy, grew a pair and did the following: Got my account (zxMarce) at
> GitLab. Created and added the necessary SSH Keys.
>
> Then, cloned with "git clone https://gitlab.com/gambas/gambas.git
> gambasdevel".
> Navigated to "gb.db.odbc/src" and changed the old "main.c" with my changed
> file from ye olde SF SVN tree (kept changes by bgermann from 3 days ago).
>
> Then, I guessed I had to commit my changes to my local repo, so I tried
> "git commit". But Git surprised me by asking my for name and email (wow!
> heavier than the NSA!). I supplied my GitLab username and email, then tried
> to commit again, but could not due to changes not staged (?!). The exchange
> was thus:
>
>    ~/gambasdevel$ git commit
>
>    *** Please tell me who you are.
>
>    Run
>
>      git config --global user.email "yo...@ex..."
>      git config --global user.name "Your Name"
>
>    to set your account's default identity.
>    Omit --global to set the identity only in this repository.
>
>    fatal: unable to auto-detect email address (got 'xxx@yyy.zzz')
>    ~/gambasdevel$ git config --global user.email "d4t...@gm..."
>    ~/gambasdevel$ git config --global user.name "zxMarce"
>    ~/gambasdevel$ git commit
>    On branch master
>    Your branch is up-to-date with 'origin/master'.
>    Changes not staged for commit:
>        modified:   gb.db.odbc/src/main.c
>
>    Untracked files:
>        gb.db.odbc/src/ml_doit.sh
>
>    no changes added to commit
>
> The untracked .sh is a script I made myself to quickly compile and install
> the ODBC component. It should remain so.
> But I don't know how to "stage" and then "commit" my changed file.
> Then, I will have to deal with change-logging and uploading, but I'd like
> to go "baby steps".
>
> TIA,
> zxMarce.
>
> ------------------------------------------------------------
> ------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Gambas-devel mailing list
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gambas-devel
>
>

[Gambas-devel] Cloned and edited. How do I get the changes staged, committed, and then uploaded to GitLab?

[First L. <d4t...@gm...>]

Guy, grew a pair and did the following: Got my account (zxMarce) at GitLab.
Created and added the necessary SSH Keys.

Then, cloned with "git clone https://gitlab.com/gambas/gambas.git
gambasdevel".
Navigated to "gb.db.odbc/src" and changed the old "main.c" with my changed
file from ye olde SF SVN tree (kept changes by bgermann from 3 days ago).

Then, I guessed I had to commit my changes to my local repo, so I tried
"git commit". But Git surprised me by asking my for name and email (wow!
heavier than the NSA!). I supplied my GitLab username and email, then tried
to commit again, but could not due to changes not staged (?!). The exchange
was thus:

   ~/gambasdevel$ git commit

   *** Please tell me who you are.

   Run

     git config --global user.email "yo...@ex..."
     git config --global user.name "Your Name"

   to set your account's default identity.
   Omit --global to set the identity only in this repository.

   fatal: unable to auto-detect email address (got 'xxx@yyy.zzz')
   ~/gambasdevel$ git config --global user.email "d4t...@gm..."
   ~/gambasdevel$ git config --global user.name "zxMarce"
   ~/gambasdevel$ git commit
   On branch master
   Your branch is up-to-date with 'origin/master'.
   Changes not staged for commit:
       modified:   gb.db.odbc/src/main.c

   Untracked files:
       gb.db.odbc/src/ml_doit.sh

   no changes added to commit

The untracked .sh is a script I made myself to quickly compile and install
the ODBC component. It should remain so.
But I don't know how to "stage" and then "commit" my changed file.
Then, I will have to deal with change-logging and uploading, but I'd like
to go "baby steps".

TIA,
zxMarce.

Re: [Gambas-devel] Git - more surprises

[Christof T. <ch...@de...>]

Am 18.08.2017 um 03:26 schrieb Benoît Minisini via Gambas-devel:

> Check your EDITOR environment variable. I put "kwrite" in it, and 
> everything went fine.

This can be done also for Git (global or per repository) only:

https://stackoverflow.com/questions/2596805/how-do-i-make-git-use-the-editor-of-my-choice-for-commits


Alles Gute

Christof Thalhofer

-- 
Dies ist keine Signatur

Re: [Gambas-devel] Cannot populate database Connection.Name from ODBC driver.

[PICCORO M. L. <mck...@gm...>]

umm okok, write in the browser event in the console:

2017-08-18 7:22 GMT-04:01 ML <d4t...@gm...>:

> of "CREATE IF NOT EXIST Tabla1...") fails to retrieve at least the DB
> filename (test.sq3) as the Database/Catalog name.
> That particular DB is a file, so I guess the driver should retrieve
> *something* as DB Name but it does not, confusing everybody into


that's the behaviour that i try to explain... a DBMS connection "instance"
don't have to handle a db "instance", that's behaviour found on DBMS's such
like postgres or mysql inherits from the "M$" era but are not mandatory...

but in many propietary DBMS the connections dont "must handle" a db
"instance" and that's correct... in some place the ODBC especification said
that a db name must be provided, but mayor DBMS dont need a db name
provided, that's behaviour are found commonly in Oracle, Firebird, Sybase,
SAP and many other, almost all the propietary DBMS do in this way...



Lenz McKAY Gerardo (PICCORO)
http://qgqlochekone.blogspot.com

2017-08-18 8:51 GMT-04:00 Benoît Minisini <ga...@us...>:

> Le 18/08/2017 à 14:39, PICCORO McKAY Lenz a écrit :
>
>>
>> that's the behaviour that i try to explain... DBMS connection "intance"
>> dont have to handle a db "instance" that's behaviour its copied from M$ in
>> mysql and postgres but DBMS in the connections dont must handle a db
>> "instance" and that's correct... but ODBC M$ especifications said that
>> must.. this king of "rare" conflicts its common in documents where M$ are
>> involved!
>>
>>
> I have no idea what you try to tell us, your english is absolutely
> unreadable.
>
> Please try to make some effort!
>
> Can't you, at least, install an english spelling checker on your mail
> client or your browser? It would be a good start.
>
> As for the grammar, I don't know if there is a grammatical corrector for
> english on Linux.
>
> Regards,
>
> --
> Benoît Minisini
>

Re: [Gambas-devel] Cannot populate database Connection.Name from ODBC driver.

[Benoît M. <ga...@us...>]

Le 18/08/2017 à 14:39, PICCORO McKAY Lenz a écrit :
> 
> that's the behaviour that i try to explain... DBMS connection "intance" 
> dont have to handle a db "instance" that's behaviour its copied from M$ 
> in mysql and postgres but DBMS in the connections dont must handle a db 
> "instance" and that's correct... but ODBC M$ especifications said that 
> must.. this king of "rare" conflicts its common in documents where M$ 
> are involved!
> 

I have no idea what you try to tell us, your english is absolutely 
unreadable.

Please try to make some effort!

Can't you, at least, install an english spelling checker on your mail 
client or your browser? It would be a good start.

As for the grammar, I don't know if there is a grammatical corrector for 
english on Linux.

Regards,

-- 
Benoît Minisini

Re: [Gambas-devel] Cannot populate database Connection.Name from ODBC driver.

[PICCORO M. L. <mck...@gm...>]

2017-08-18 7:22 GMT-04:00 ML <d4t...@gm...>:

> of "CREATE IF NOT EXIST Tabla1...") fails to retrieve at least the DB
> filename (test.sq3) as the Database/Catalog name.
> That particular DB is a file, so I guess the driver should retrieve
> *something* as DB Name but it does not, confusing everybody into


that's the behaviour that i try to explain... DBMS connection "intance"
dont have to handle a db "instance" that's behaviour its copied from M$ in
mysql and postgres but DBMS in the connections dont must handle a db
"instance" and that's correct... but ODBC M$ especifications said that
must.. this king of "rare" conflicts its common in documents where M$ are
involved!

Re: [Gambas-devel] Cannot populate database Connection.Name from ODBC driver.

[ML <d4t...@gm...>]

Benoît,

Thanks a ton. It looks way easier than I tried/thought... As usual.


Piccoro,

First off, I always default my examples to MSSQL because I work with
them, for good or bad. That's mainly what I have to try/test things from
Gambas.
But a single MySQL server -which is mostly free ($$$) IIRC- can also be
home to more than one DB/Catalog. Others may or may not follow suit.
Never tried Sybase, DB/2, or Oracle because I don't have access to
either. I rely on users to report inconsistencies like I had for
Firebird regarding RecordCount.

Never mind RDBMS that don't have a DB Name. The interface should be as
wide/general as possible, and if at least one of them uses or can switch
DB/Catalog names, then I think it should be supported.
Actually, SQLite3 (the one I do have at home, with your supplied example
of "CREATE IF NOT EXIST Tabla1...") fails to retrieve at least the DB
filename (test.sq3) as the Database/Catalog name.
That particular DB is a file, so I guess the driver should retrieve
*something* as DB Name but it does not, confusing everybody into
thinking it's an in-memory DB.

On the other hand, connstrings handle instances -at least with MSSQL-
transparently by specifying the instance in the server name, and this is
separate from the DB name.
So you can have any mix-n-match of servers, instances, and DBs, given
every server/instance has its own connection object.
Then you can switch DBs/Catalogs inside each Connection by running
MSSQL's USE (or the applicable RDBMS-specific) query command.


Regards,
zxMarce.

On 18/08/17 04:37, PICCORO McKAY Lenz wrote:
> hi zxMarce and other, i want to mark a WARNING here.. as i can see too
> many experience with mososoft SQL and must be some notes about some
> other scalar HIGH DBMS..
> sybase DOES NOT NEED A DB NAME TO HAVE A CONNECTION !
> be warning: its not like sqlite that a db "in memory" are present at
> least, here theres no connection, in same way db2 and oracle...
>
> 2017-08-17 21:25 GMT-04:00 Benoît Minisini via Gambas-devel
> <gam...@li...>:
>> const char *new_value;
>> GB.FreeString(&desc->name); // Free the Gambas string containing the
>> old database name
>> desc->name = GB.NewZeroString(new_value); // Create a new Gambas
>> string from the zero-terminated string new_value, and store it.
> its great that can be changed the connection db on the fly for the
> module.. its mandatory..
> ok ODBC need! a connection db name when Established, but its only with
> UnixODBC.. there's more "special behaviours" in same "host connection
> name" there are also "same many db names" that its not the same as "in
> same host are many db names served"
> that strange behaviour are applied to oralce and sybase, that runs "a
> instance" of the software for "one or more than" databases... in the
> same host..
> due that PROPIETARY ODBC SOFTWARE CAN OFFERS A CONNECTION TO GAMBAS
> WITHOUT A DATABASE VALID STRUCTURE, or a "empty valid connection" in
> conclusion.. or more rare as succed to me, a connection but the db are
> not the spected (due UnixODBC freedts can connect to a one db per host
> at same time, as all here knowed. but it not limited to in real ysbae
> or oracle)
>> That's it.
>> --
>> Benoît Minisini

Re: [Gambas-devel] Cannot populate database Connection.Name from ODBC driver.

[PICCORO M. L. <mck...@gm...>]

hi zxMarce and other, i want to mark a WARNING here.. as i can see too
many experience with mososoft SQL and must be some notes about some
other scalar HIGH DBMS..

sybase DOES NOT NEED A DB NAME TO HAVE A CONNECTION !
be warning: its not like sqlite that a db "in memory" are present at
least, here theres no connection, in same way db2 and oracle...

2017-08-17 21:25 GMT-04:00 Benoît Minisini via Gambas-devel
<gam...@li...>:
> const char *new_value;
>
> GB.FreeString(&desc->name); // Free the Gambas string containing the old
> database name
>
> desc->name = GB.NewZeroString(new_value); // Create a new Gambas string from
> the zero-terminated string new_value, and store it.

its great that can be changed the connection db on the fly for the
module.. its mandatory..

ok ODBC need! a connection db name when Established, but its only with
UnixODBC.. there's more "special behaviours"

* in same "host connection name" there are also "same many db names"
that its not the same as "in same host are many db names served"

that strange behaviour are applied to oralce and sybase, that runs "a
instance" of the software for "one or more than" databases... in the
same host..

due that PROPIETARY ODBC SOFTWARE CAN OFFERS A CONNECTION TO GAMBAS
WITHOUT A DATABASE VALID STRUCTURE, or a "empty valid connection" in
conclusion.. or more rare as succed to me, a connection but the db are
not the spected (due UnixODBC freedts can connect to a one db per host
at same time, as all here knowed. but it not limited to in real ysbae
or oracle)

>
> That's it.
>
> --
> Benoît Minisini
>
>
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Gambas-devel mailing list
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gambas-devel

Re: [Gambas-devel] GitLab - notifications

[PICCORO M. L. <mck...@gm...>]

there are levels of "whatching".. the default its only notify if
something are relevat and related to you.. but its you arenot mention
or an isue are not mentioned or related to you.. there's no
notification that's the default..
Lenz McKAY Gerardo (PICCORO)
http://qgqlochekone.blogspot.com


2017-08-17 21:35 GMT-04:00 ada...@gm... <ada...@gm...>:
> I've read the help and thought I had notifications set up properly for gambas/gambas, but nothing ever shows up. The only way I found that there are some new changes was from discussions in this list.
> I get notifications via email from my own GitLab playground, but not from gambas/gambas.
> Is there some magical incantation I need to perform?
>
> tia
> b
>
> --
> B Bruen <ada...@gn... (sort of)>
>
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Gambas-devel mailing list
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gambas-devel

Re: [Gambas-devel] Git - a helpful tutorial

[Laurent C. <lor...@gm...>]

Le vendredi 18 août 2017, 03:30:40 CEST ada...@gm... a écrit :
> https://www.atlassian.com/git/tutorials/svn-to-git-prepping-your-team-migrat
> ion#for-developers
> 
> I found that page especially the "for developers" bit, and indeed that whole
> site, very helpful in understanding what is actually going on.
> 
> hth
> b

And what about an officia	l book?

https://git-scm.com/book/en/v2 and available in various languages.

-- 
Laurent Carlier
http://www.archlinux.org

Re: [Gambas-devel] GitLab - notifications

[Adrien P. <adr...@gm...>]

Le Fri, 18 Aug 2017 03:35:21 +0200, ada...@gm...  
<ada...@gm...> a écrit:

> I've read the help and thought I had notifications set up properly for  
> gambas/gambas, but nothing ever shows up. The only way I found that  
> there are some new changes was from discussions in this list.
> I get notifications via email from my own GitLab playground, but not  
> from gambas/gambas.
> Is there some magical incantation I need to perform?
>
> tia
> b
>

The global "Watch" settings for the Gambas group works for me, but  
personally I've
been using the RSS feed to avoid being flooded with emails :

https://gitlab.com/gambas.atom

It's available on the "Activity" tab of the Gambas group.

-- 
Adrien Prokopowicz

[Gambas-devel] GitLab - notifications

[<ada...@gm...>]

I've read the help and thought I had notifications set up properly for gambas/gambas, but nothing ever shows up. The only way I found that there are some new changes was from discussions in this list.
I get notifications via email from my own GitLab playground, but not from gambas/gambas.
Is there some magical incantation I need to perform?

tia
b

-- 
B Bruen <ada...@gn... (sort of)>

[Gambas-devel] Git - a helpful tutorial

[<ada...@gm...>]

https://www.atlassian.com/git/tutorials/svn-to-git-prepping-your-team-migration#for-developers

I found that page especially the "for developers" bit, and indeed that whole site, very helpful in understanding what is actually going on.

hth
b

-- 
B Bruen <ada...@gn... (sort of)>

Re: [Gambas-devel] Git - more surprises

[Benoît M. <ga...@us...>]

Le 18/08/2017 à 03:21, ada...@gm... a écrit :
> Ok, so I did the pull into our local "gambasbase" clone which is a "single-branch" of the "master". All worked fine and super quick.
> 
> Then, I did a pull from my personal clone of "gambasbase". This one has some branches in it for our (paddys-hill) changes.  That pull automagically started a merge, which is fine, but then....
> 
> it invoked vim on the commit header. That was a surprise!  Not the least of which was trying to remember how the heck to get out of vim.
> 
> Is there a way to turn off this "feature"?
> 
> b
> 

Check your EDITOR environment variable. I put "kwrite" in it, and 
everything went fine.

-- 
Benoît Minisini

Re: [Gambas-devel] Cannot populate database Connection.Name from ODBC driver.

[Benoît M. <ga...@us...>]

Le 18/08/2017 à 02:26, zxMarce a écrit :
> Hi there. This is mainly intended to be tackled by Benoît, but anyone with
> the necessary knowledge (Toby?) please feel free to butt in.
> 
> The ODBC driver can use Connection Strings since some time now. I also have
> a patched up module that can run successfully non-data-producing queries
> (the SQL_NO_DATA return code is no longer treated as an error), thanks to a
> bug submitted by Piccoro.
> 
> Problem is in the open_database() function. When using connstrings, the
> database name is usually part of this string, and no .Name property is
> expected to be populated from the BASIC program, nor actually used.
> So, I'm trying to populate the property from the ODBC backend. I can
> successfully fetch the DB Name from the lower level database driver in use
> (SQLite3 does not provide a name, tho, be warned).
> What I cannot do, be it via copying char[] or Gambas String buffer pointers,
> is to get the .Name property populated back to the running BASIC program.
> 
> The open_database() call has two parameters: DB_DESC *desc and DB_DATABASE
> *db. My best guess is that I have to fiddle with desc, but not really sure
> if that's correct nor exactly how.
> 
> To further salt the issue, queries exist in big RDBMS that allow to
> hot-switch databases while connected to a server (example: USE <newDBName>
> in MSSQL). So, the connection's .Name property should also be "refetched"
> after a successful query is run.
> 
> Any suggestions?
> 
> 

The open_database(DB_DESC *desc, DB_DATABASE *db) driver function takes 
two arguments:

- DB_DESC *desc : a pointer to a structure that contains all connections 
string properties and an integer option field.

- DB_DATABASE *db : a pointer to another structure that should be filled 
with all necessary information about the connection.

By modifying the desc fields, you are modifying the corresponding 
Connection object properties, as this is where they are stored.

But you must use GB.FreeString() and GB.NewString() to replace a string 
property value.

For example:

const char *new_value;

GB.FreeString(&desc->name); // Free the Gambas string containing the old 
database name

desc->name = GB.NewZeroString(new_value); // Create a new Gambas string 
from the zero-terminated string new_value, and store it.

That's it.

-- 
Benoît Minisini