> On 16 Dec 2017, at 17:55, Stephane Ducasse <stepharo.s...@gmail.com> wrote:
> 
> Hi Esteban
> 
> Now how can I maintain my fork up to date?

there is a “synchronise repositories” option there, at the side of “create 
branch from fogbugz”.

> This is not in the tutorial and to me it was a major problem and a big
> source of pain.

it is not there because you do not need it.
when you “create branch from fogbugz”, it will take your image version and 
branch from there. Then, when you push to your remote, you will push correctly. 
There is no need to have syncd the forks to contribute… 

Now, if you do  “synchronise repositories”, it will take what is in 
pharo-project/pharo/development branch and it will sync that with your 
pharo/development branch. But again… there is no need to do this to contribute.

Esteban

> Stef
> 
> On Sat, Dec 16, 2017 at 5:39 PM, Stephane Ducasse
> <stepharo.s...@gmail.com> wrote:
>> I committed my code for issue 20861 and I do not understand what I see
>> 
>> I do not understand why I get local changes with empty packages. :(
>> 
>> Stef
>> 
>> On Sat, Dec 16, 2017 at 5:32 PM, Stephane Ducasse
>> <stepharo.s...@gmail.com> wrote:
>>> I did a little pass on the wiki to make the flow clearer
>>> 
>>> 
>>> On Sat, Dec 16, 2017 at 5:26 PM, Stephane Ducasse
>>> <stepharo.s...@gmail.com> wrote:
>>>> In the tutorial:
>>>> 
>>>> - Put a little heading before
>>>> 
>>>> "You need to add pharo repository as a remote
>>>> (g...@github.com:pharo-project/pharo.git)."
>>>> 
>>>> On Sat, Dec 16, 2017 at 5:25 PM, Stephane Ducasse
>>>> <stepharo.s...@gmail.com> wrote:
>>>>> I double clicked and it did a massive amount of stuff and finally told
>>>>> me that it is up to date.
>>>>> 
>>>>> 
>>>>> On Sat, Dec 16, 2017 at 5:24 PM, Stephane Ducasse
>>>>> <stepharo.s...@gmail.com> wrote:
>>>>>> OK so I restarted everything from scratch:
>>>>>> - deleted my fork
>>>>>> - reforked
>>>>>> - clone pharo again
>>>>>> - here is some feedback
>>>>>> 
>>>>>> In the tutorial add /pharo + src in the screenshot
>>>>>> 
>>>>>> 
>>>>>> Then when I add the local repository I get uncommited changes and I do
>>>>>> not understand why?
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> On Sat, Dec 16, 2017 at 9:57 AM, Esteban Lorenzano <esteba...@gmail.com> 
>>>>>> wrote:
>>>>>>> 
>>>>>>> 
>>>>>>> On 16 Dec 2017, at 09:42, Alistair Grant <akgrant0...@gmail.com> wrote:
>>>>>>> 
>>>>>>> Hi Esteban,
>>>>>>> 
>>>>>>> On 16 December 2017 at 09:05, Esteban Lorenzano <esteba...@gmail.com> 
>>>>>>> wrote:
>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>>> On 15 Dec 2017, at 17:37, Alistair Grant <akgrant0...@gmail.com> wrote:
>>>>>>> 
>>>>>>> Hi Esteban,
>>>>>>> 
>>>>>>> I had no problems following the process (Ubuntu 16.04,
>>>>>>> Pharo7.0-32bit-e175bc2.image, fogbugz 20872). :-)
>>>>>>> 
>>>>>>> I guess that you have already thought of this, but...  Is there any
>>>>>>> reason why we can't just put up a dialog asking for the user's github
>>>>>>> credentials and fogbugz issue number and then automatically clone the
>>>>>>> repository, configure the upstream remote and create the issue branch.
>>>>>>> That would remove most of the remaining manual steps.
>>>>>>> 
>>>>>>> I realise that it only works for option 1, although where people
>>>>>>> configure a common pharo-local, it could check for a pre-existing
>>>>>>> clone and use that one.
>>>>>>> 
>>>>>>> 
>>>>>>> "I realise” means you tried and it didn’t work?
>>>>>>> because in my tests it worked as good as the first one (I tested on
>>>>>>> windows), but that may need to be “re-validated” :)
>>>>>>> 
>>>>>>> Esteban
>>>>>>> 
>>>>>>> 
>>>>>>> The contribution process works fine (even on linux :-)).
>>>>>>> 
>>>>>>> The "I realise" paragraph is a comment on my suggestion to try and
>>>>>>> reduce the number of manual steps required (and is actually wrong).
>>>>>>> Just to rephrase (and extend) the suggestion, I think we could create
>>>>>>> a single dialog that currently covers the following steps (from your
>>>>>>> instructions):
>>>>>>> 
>>>>>>> 1. Clone a fresh repository, or point to an existing repository.
>>>>>>> 2. Tell Iceberg about pharo-project
>>>>>>> 3. Create a new branch from the fogbugz issue
>>>>>>> 
>>>>>>> 
>>>>>>> ah, I got lost in translation ;)
>>>>>>> 
>>>>>>> Esteban
>>>>>>> 
>>>>>>> 
>>>>>>> Cheers,
>>>>>>> Alistair
>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>>> Cheers,
>>>>>>> Alistair
>>>>>>> 
>>>>>>> On 14 December 2017 at 13:19, Esteban Lorenzano <esteba...@gmail.com> 
>>>>>>> wrote:
>>>>>>> 
>>>>>>> Hi!
>>>>>>> 
>>>>>>> I’m working on simplifying the contribution process, after collecting
>>>>>>> opinions/experiences last couple of months.
>>>>>>> As you know, Pharo contribution process is still WIP and we aim to have 
>>>>>>> it
>>>>>>> as smooth as possible for Pharo 7.0 release. Now, after observe the 
>>>>>>> idea of
>>>>>>> the “system repositories” was a bad idea because it introduced extra 
>>>>>>> and non
>>>>>>> standard “path” to contribution, I managed to remove that to reestablish
>>>>>>> “the regular way”: you will now need to add pharo repository just as any
>>>>>>> other repository you add, by cloning or adding local repository.
>>>>>>> 
>>>>>>> I took Guille’s doc and moved it to pharo project (it does not has 
>>>>>>> sense to
>>>>>>> have it living in a contributor’s repository when is so important). You 
>>>>>>> can
>>>>>>> find it here:
>>>>>>> 
>>>>>>> https://github.com/pharo-project/pharo/wiki/Contribute-a-fix-to-Pharo
>>>>>>> 
>>>>>>> This document is also updated to reveal this new process, please read 
>>>>>>> it.
>>>>>>> 
>>>>>>> How to update your startup scripts?
>>>>>>> Some people has added startup scripts to easy the first part of
>>>>>>> contribution. Instead enabling system repositories, etc. you now need to
>>>>>>> replace that with this:
>>>>>>> 
>>>>>>> (IceRepositoryCreator new
>>>>>>> location: '/path/to/pharo-project/pharo' asFileReference;
>>>>>>> subdirectory: 'src';
>>>>>>> createRepository)
>>>>>>> register
>>>>>>> 
>>>>>>> PLEASE, PLEASE, PLEASE… take a moment to read and try the document. Is 
>>>>>>> very
>>>>>>> important that document reflects new process and works reliable in 
>>>>>>> different
>>>>>>> scenarios (I validated it on macOS and Windows, and assumed it worked 
>>>>>>> fine
>>>>>>> on linux but you know… bad assumptions is the base of failure ;) )
>>>>>>> 
>>>>>>> I’m eager to hear your feedback and continue enhancing the process.
>>>>>>> 
>>>>>>> (yes, Stef, I know UI is still cumbersome… I’m working on that :) )
>>>>>>> 
>>>>>>> cheers!
>>>>>>> Esteban
>>>>>>> 
>>>>>>> 
> 


Reply via email to