Hi Aitor,

> On Jun 25, 2022, at 3:22 PM, Aitor Santamaría <aitor...@gmail.com> wrote:
> [...]
> How do these logs build up? (I am assuming logs in an user-readable format). 
> Every time someone updates a tool, do we need to fill in somewhere?
> Ideally, the implemented bugs and enhancements from GitLab would go to the 
> list, but I guess the ticketing system is not being widely used. ANother 
> option would be that every time a new version comes, instead of logging 
> somewhere else, someone simply logs a new enhancement ticket in GitLab with 
> the changelog.

The packages are all sourced from the GitLab Archive. A couple other key 
components to the release (like FDI, FD-NLS, RBE, etc) are pulled from their 
own GitHub/GitLab projects. The CHANGES.LOG is created using the commit comment 
messages for each project. 

For example, if KEYB on the GitLab archive has a commit pushed with the comment 
“Fixed double keypress bug”. That comment would appear in the CHANGES.LOG as 
something like “2022-06-29 04:52:11 keyb (Aitor Santamaría): Fixed double 
keypress bug”

Also if for example that same commit was made to a branch (other than 
master/main) as a "test fix”, it would be identified as such in the LOG. By 
default, the interim builds check for an “unstable” branch and will use that 
instead of the master/main branch when creating a build. That would look like 
“2022-06-29 04:52:11 keyb [unstable] (Aitor Santamaría): Fixed double keypress 
bug”

For projects that are maintained elsewhere and just get copied and adjusted 
into the GitLab Archive, they usually just get a version number. However, we 
can include more information in the commit message. For example, I updated 
DOSLFN to v0.41f yesterday. It’s logged as "2022-06-24 08:54:13 doslfn 
(shidel): v0.41f - More extending directory fixes and basic version check”

It’s possible that we may include a compatible CHANGE.LOG for the maintained 
elsewhere projects in their GitLab Archive’s version. But, I’m not sure that is 
necessary or desirable. They get the version number comment and if someone is 
interested they could check that projects change.log. After all, it’s not 
actually changes that we would have made. IDK.

> 
> Finally, what path on ibiblio do you want to use to provide package updates? 
> This has nothing to do with creating the interim build. As you recall, the 
> current RBE uses the GitLab FreeDOS Archive (https://gitlab.com/FreeDOS 
> <https://gitlab.com/FreeDOS>) to create the packages on the release. It can 
> also automatically pull specific branches of those packages based on the 
> release, interim, development or other type of build it is creating. The 
> package update url only applies to where the release / interim build will 
> check for package updates. 
> 
> I think it makes sense to keep it along side the other OS update 
> repositories. There are already repositories for each Release
> 
>       1.1 - 
> https://www.ibiblio.org/pub/micro/pc-stuff/freedos/files/repositories/1.1/ 
> <https://www.ibiblio.org/pub/micro/pc-stuff/freedos/files/repositories/1.1/>
>       1.2 - 
> https://www.ibiblio.org/pub/micro/pc-stuff/freedos/files/repositories/1.2/ 
> <https://www.ibiblio.org/pub/micro/pc-stuff/freedos/files/repositories/1.2/>
>       1.3 - 
> https://www.ibiblio.org/pub/micro/pc-stuff/freedos/files/repositories/1.3/ 
> <https://www.ibiblio.org/pub/micro/pc-stuff/freedos/files/repositories/1.3/>
> I don't know if I am getting this well: this means that in the 1.2 folder you 
> find the packages updated over 1.2? 
> This means that if at this moment there appears a new updated package that 
> hasn't been updated earlier, it should go into ALL of the 1.1, 1.2 and 1.3 
> folders?

Basically, yes. There is a separate software update repository for maintained 
for each of OS release.

This supports the ability for packages under different OS versions to use 
different paths, settings, options, versions or other differences. However, 
generally there is no difference in the actual update packages for each 
release.  The repository management utility has no problem maintaining a bunch 
of separate repos. When an update is made to a package, we can just drop a zip 
in one of a couple directories on the server. The repo management stuff will 
grab it, stick it into the proper repository, generate file-system links, 
versioned webpages, etc. However, it will have at 1 copy of the update zip in 
each of the update repositories.

When I eventually get around to making the next version of the repo management 
utilities, it will be able eliminate multiple cross-repo duplicates and 
conserve wasted disk space. That along with desktop clients to more easily 
manage them, multi-repo comparison, syncing and numerous other improvements. 
But, I’m just one person and it keeps getting put off in favor of other more 
immediate needs. Someday hopefully, I will get to it. 

> 
> AItor

:-)

Jerome

_______________________________________________
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel

Reply via email to