Skip to content

Conversation

@nilason
Copy link
Contributor

@nilason nilason commented Oct 20, 2025

Follow up to the de facto update with commit: ec42cfa.

Run with autoreconf -vfi.

(What version we use is only of importance for developers updating the configure scripts, not users/builders which use the generated configure script anyway.)

Follow up to the de facto update with commit: ec42cfa
@nilason nilason added this to the 8.5.0 milestone Oct 20, 2025
@marisn
Copy link
Contributor

marisn commented Oct 21, 2025

Do we have to bump up autoconf requirement? The fact that configure was generated with a more recent version doesn't mean an older version can not be used.

@nilason
Copy link
Contributor Author

nilason commented Oct 21, 2025

Do we have to bump up autoconf requirement? The fact that configure was generated with a more recent version doesn't mean an older version can not be used.

We don’t have to, but using different versions lead to unrelated changes (cluttered history). The version is only of importance for us devs, as it now stands, so whichever is more convenient.

@ninsbl
Copy link
Member

ninsbl commented Oct 21, 2025

Since you are modifying the build system: Did you notice that current builds show up with Large File Support (LFS): no? See e.g.: https://github.com/OSGeo/grass/actions/runs/18658480388/job/53193250341#step:10:275
Here: https://github.com/OSGeo/grass/actions/runs/18549653323/job/52874607114 it was stil enabled by default... Is that a regression that warrants a new issue?

@marisn
Copy link
Contributor

marisn commented Oct 22, 2025

Since you are modifying the build system: Did you notice that current builds show up with Large File Support (LFS): no? See e.g.: https://github.com/OSGeo/grass/actions/runs/18658480388/job/53193250341#step:10:275 Here: https://github.com/OSGeo/grass/actions/runs/18549653323/job/52874607114 it was stil enabled by default... Is that a regression that warrants a new issue?

Yes, that is a regression. LFS detection works, but reporting – doesn't. Here's a PR with the fix: #6537

@marisn
Copy link
Contributor

marisn commented Oct 22, 2025

We don’t have to, but using different versions lead to unrelated changes (cluttered history). The version is only of importance for us devs, as it now stands, so whichever is more convenient.

I was more thinking of downstream packagers as on some distros they discard our shipped configure file and generate their own. Now they will be forced to use 2.72, though 2.71 should work just fine as they don't contribute back their configure files. But probably it doesn't matter that much.

@echoix
Copy link
Member

echoix commented Oct 23, 2025

@nilason Would you want to rerun the autoreconf again, since the LFS fix up PR was merged?

@nilason
Copy link
Contributor Author

nilason commented Oct 29, 2025

We don’t have to, but using different versions lead to unrelated changes (cluttered history). The version is only of importance for us devs, as it now stands, so whichever is more convenient.

I was more thinking of downstream packagers as on some distros they discard our shipped configure file and generate their own. Now they will be forced to use 2.72, though 2.71 should work just fine as they don't contribute back their configure files. But probably it doesn't matter that much.

Browsing on repology.org, the platforms delivering 8.4 all are accompanied by autoconf 2.72, so this shouldn't be a problem for 8.5.

@nilason
Copy link
Contributor Author

nilason commented Oct 29, 2025

@nilason Would you want to rerun the autoreconf again, since the LFS fix up PR was merged?

There wouldn't be any need for it (as it has been run by 2.72), but I tested to be sure.

@nilason nilason merged commit 6f92ad1 into OSGeo:main Oct 29, 2025
33 of 37 checks passed
@nilason nilason deleted the update_autoconf_272 branch October 29, 2025 14:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants