You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
While Git's release notes still mention, timidly: "Rust support is enabled by default (but still allows opting out) in some future version of Git." and hedges even further in the "Breaking Changes" document: "We will evaluate the impact on downstream distributions before making Rust mandatory in Git 3.0.", Git for Windows needs to prepare for Rust, if only to allow for distribution tools implemented in that language e.g. to retire the Perl version of docx2txt. And Rust raised the requirements for the regular *-windows-gnu targets to Windows 10 in version 1.76.
Therefore, v2.55 (which might, or might not, be the last Git version before Git v3.0) seems like the best version to mark as the last one to support Windows 8.1.
v2.55 (which might, or might not, be the last Git version before Git v3.0)
https://tinyurl.com/gitCal still calls the next cycle "2.56 cycle" while that is not even close hard evidence, it does point towards some level of plans of at least one more 2.x version before 3.0.
v2.55 (which might, or might not, be the last Git version before Git v3.0)
https://tinyurl.com/gitCal still calls the next cycle "2.56 cycle" while that is not even close hard evidence, it does point towards some level of plans of at least one more 2.x version before 3.0.
I wouldn't put too much stock in that. gitCal has changed without any advance notice or explanations before...
There was also some mention of some funny "jump to 2.95 first" business on the Git mailing list. Who the heck knows what will happen, it's all up in the air.
Following the MSYS2 project, on which Git for Windows is based and which dropped support for Windows 8.1, after Git for Windows v2.55 support that Windows version will be dropped.
I've intentionally called it "far from hard evidence", because I know it can and does change, but I have to assume that Junio was somewhat convinced of 2.56 happening when he created 12 weeks of calender entries with that name. There's also no current talk of when 3.0 will happen, what's missing for 3.0, which would be signs of a 3.0 happening soon.
We will evaluate the impact on downstream distributions before making Rust mandatory in Git 3.0
I've also not seen any sign of that planned evaluation, but Randall just came out of the woodworks and gave an estimate of "no earlier than 2027" for his particular downstream being ready.
Again, I don't think we'll nescessarily wait for him, but he might at least have enough pull to slow the inevitable down a little. I'm also not trying suggest that we need to base our timeline on him
Activity
rimrul commentedon Jun 11, 2026
https://tinyurl.com/gitCal still calls the next cycle "2.56 cycle" while that is not even close hard evidence, it does point towards some level of plans of at least one more 2.x version before 3.0.
dscho commentedon Jun 11, 2026
I wouldn't put too much stock in that. gitCal has changed without any advance notice or explanations before...
There was also some mention of some funny "jump to 2.95 first" business on the Git mailing list. Who the heck knows what will happen, it's all up in the air.
dscho commentedon Jun 11, 2026
As to the notice, how about:
Following the MSYS2 project, on which Git for Windows is based and which dropped support for Windows 8.1, after Git for Windows v2.55 support that Windows version will be dropped.
mjcheetham commentedon Jun 11, 2026
I was struggling to read that wording.. how about:
dscho commentedon Jun 11, 2026
@mjcheetham great! Could you
/add relnote blurbthat?mjcheetham commentedon Jun 11, 2026
/add relnote blurb Following the MSYS2 project, on which Git for Windows is based, Windows 8.1 support will be dropped after Git for Windows v2.55.
The workflow run was started
mjcheetham commentedon Jun 11, 2026
git-for-windows/build-extra@7ee5cf2
rimrul commentedon Jun 12, 2026
I've intentionally called it "far from hard evidence", because I know it can and does change, but I have to assume that Junio was somewhat convinced of 2.56 happening when he created 12 weeks of calender entries with that name. There's also no current talk of when 3.0 will happen, what's missing for 3.0, which would be signs of a 3.0 happening soon.
I've also not seen any sign of that planned evaluation, but Randall just came out of the woodworks and gave an estimate of "no earlier than 2027" for his particular downstream being ready.
Again, I don't think we'll nescessarily wait for him, but he might at least have enough pull to slow the inevitable down a little. I'm also not trying suggest that we need to base our timeline on him
git-artifacts: opt out of the Rust bits for now
git-artifacts: opt out of the Rust bits for now (#124)
git-artifacts: opt out of the Rust bits for now
git-artifacts: opt out of the Rust bits for now
dscho commentedon Jul 21, 2026
I added the checkbox to the list to finalize git-for-windows/git-sdk-64#117.
git-artifacts: opt out of the Rust bits for now
git-artifacts: opt out of the Rust bits for now
dscho commentedon Sep 12, 2026
This is well underway now! With Git for Windows v2.56.0, we will no longer support Windows 8.1 (because of UCRT64).