View previous topic :: View next topic |
Author |
Message |
Ttelmah
Joined: 11 Mar 2010 Posts: 19658
|
|
Posted: Thu Mar 24, 2022 7:40 am |
|
|
Problem is I agree with both of you!...
A project that is actually being directly 'worked' on, should be local not
on a server. However archives obviously need to be elsewhere.
Personally I'd say if you are working on something, a local copy should
be made that is updated to the server when any major change actually
'works'. Doing development directly on a server is slow (the way the
directory accesses are done drinks bandwidth), and means problems
exist on the shared files before they are tested. Not good for anyone.
Even working on projects with multiple people, you download the shared
version, make your changes to the part you are working on, and once it
works or at least compiles, put this back on the server. |
|
|
newguy
Joined: 24 Jun 2004 Posts: 1912
|
|
Posted: Thu Mar 24, 2022 8:03 am |
|
|
Use a repo. Any positive incremental change gets committed. An "I'm happy with this" version gets tagged. Your main trunk then has its version number incremented. Ensure your tag commit follows convention. It helps to have a top level README for anyone NOT involved with development or way better yet, whatever assembly line tools are automated so that they can only grab or use the latest officially blessed release.
Work on your local HDD. Once you get a "positive incremental change", push/commit it.
The remote server or repo is not a sandbox, and it should absolutely not get polluted by minute-to-minute "playing" with a project. |
|
|
temtronic
Joined: 01 Jul 2010 Posts: 9377 Location: Greensville,Ontario
|
|
Posted: Thu Mar 24, 2022 8:14 am |
|
|
I see 'sources on servers' being a real 'can of worms' when anyone can edit the sources.
Say pgmr 'a' changes a few lines in a source, B downloads it, makes a few more changes( just stuff he's supposed to ) then 'c' downloads that source, makes a couple changes and then 'd' grabs it off the server and 'it don't work' like it used to....
Just wondering HOW you get back to 'where it worked fine' version ?
I suppose a lot depends on what these 'source' files are.
Then there's the 'server's down' problem. Unless you have valid, cutrrent copies on your PC, zero work gets done....long coffee break ! |
|
|
newguy
Joined: 24 Jun 2004 Posts: 1912
|
|
Posted: Thu Mar 24, 2022 8:19 am |
|
|
Jay: this issue is addressed if you use a repository. It flags issues like this so that dev B's changes don't clobber dev A's code.
The "it worked fine" version of which you speak is called a tag. |
|
|
Ttelmah
Joined: 11 Mar 2010 Posts: 19658
|
|
Posted: Thu Mar 24, 2022 8:42 am |
|
|
Spot on Newguy. Tis the way all server stored projects I have worked on have
been done. |
|
|
jeremiah
Joined: 20 Jul 2010 Posts: 1368
|
|
Posted: Thu Mar 24, 2022 10:10 am |
|
|
temtronic wrote: | I see 'sources on servers' being a real 'can of worms' when anyone can edit the sources.
Say pgmr 'a' changes a few lines in a source, B downloads it, makes a few more changes( just stuff he's supposed to ) then 'c' downloads that source, makes a couple changes and then 'd' grabs it off the server and 'it don't work' like it used to....
Just wondering HOW you get back to 'where it worked fine' version ?
I suppose a lot depends on what these 'source' files are.
Then there's the 'server's down' problem. Unless you have valid, cutrrent copies on your PC, zero work gets done....long coffee break ! |
Newguy hit on it already, but also depending on what type of repository software you use, you can also get additional protections. Like we use a bitbucket like interface that *prohibits* someone from adding code to the production branch of code without issuing a request (called a pull request) and someone with the appropriate permissions inspects the code and approves it. However for some smaller projects, this level of extra might be overkill and is, again, not necessarily needed.
Now that's a layer of protection on top of everything else and it is optional (not necessary) since as newguy pointed out, repositories come with built in tools / options to handle this. |
|
|
temtronic
Joined: 01 Jul 2010 Posts: 9377 Location: Greensville,Ontario
|
|
Posted: Fri Mar 25, 2022 7:34 am |
|
|
hmm, going back to the 'slow' question....
I'm wondering if the server's settings could be changed to increase the 'priority' or amount of continuous time, the programmers PC has access to the server ? Maybe the actual physical connection speed can be increased ? |
|
|
|