Re: is it possible to compare your file with the "remote" version

Hi,

thanks for the details about how you've solved this situation, looks 
really interesting.
i will try to study that possibility and see if the managers are keen on 
performing any change at all on this matter.

thank you very much.

En/na Andrew Lowe ha escrit:
> On Thu, 22 May 2008 04:28:19 Jordi Moles Blanco wrote:
>   
>> hi.
>>
>> uhmmm, i've been thinking about the whole repository thing, and one of
>> the reasons why they haven't implemented in our company is because they
>> don't quite see how to use it in our situation.
>>
>> What i can do is to explain our scenario, which i don't thing is "that"
>> special and you can tell me if it would be very difficult to use a
>> repository system, if it's ok with you. I must say that i have no idea
>> about this kind of software and what the advantges are.
>>
>> as i said... we've got a local machine with the website. We edit and
>> create those files from Quanta.
>> A part from that, we've got an "online" server with apache were the
>> final files are stored so that people can see them. We upload the files
>> to the server through the "upload this file" option in Quanta, ftp
>> protocol.
>>
>> Right now, as soon as the file is uploaded, is available to everyone,
>> so... if there's any mistake in the code, everyone will see it. So... i
>> guess that the tricky thing here would be to run an apache web server
>> through a repository system and then we should be able to ask apache to
>> load files from on part or another from the repository, and i guess that
>> this part is the one that they fail to see how it would work.
>>
>> if you could give me some recommendations... i would be really grateful.
>>
>> thanks.
>>
>>     
> Jordi,
> Your situation is very similar to the way I used to run things here.  I was 
> the only developer here, so did development on my local machine, publishing 
> it to either the webserver directly or to the testing machine depending on 
> how urgent it was.  This led to breaking things, and making it hard to fix 
> them quickly.  We put another developer on, and I could see things getting 
> messy very quickly (overwriting each others changes, etc).
>
> Our solution is to run a web server on our own machine for local development, 
> just the version of apache/php/mysql that comes with our distro (I run 
> kubuntu, the other dev a mac.)
> When we are happy with our changes we commit them to subversion.  We then 
> checkout/update onto the testing server which is a mirror of the live server, 
> to make sure any features or php plugins are working.  Once this is done we 
> checkout/update onto the production machine.  We may skip the testing machine 
> if it is a small, quick change.
>
> So it ends up looking like this:
> If I have a copy of the project do a:
> svn update
> in the directory the project lives in
> or if I do not have a copy
> svn checkout https://svn.server.name/svn_web/project/
> then make changes
> test them on my local machine
> then 
> svn commit from the project directory
> on the testing machine
> svn update 
> to update the project or
> svn checkout https://svn.server.name/svn_web/project/
> if the project is new and not on the test server
> check everything works fine,
> then do the same on the production box to checkout/update the project.
>
> If you do not like the command line much kdesvn is a very good tool, and 
> integrates well with kompare.
>
> This way if something stuffs up, just roll back to the last working version, 
> or revert the change if it is really bad.
>
> I do not understand how developers could work in a team without a source code 
> management system.  It is just too easy to break others work if you rely on 
> an ftp repository.
>
> Quanta does not intergrate with subversion (that I know of), but does cvs.  I 
> think Quanta 4 should support subversion due to the kdevplatform supporting 
> it and I cannot wait!
>
> If you need more information, just ask
>
>
>