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
Copy file name to clipboardExpand all lines: rev_news/drafts/edition-136.md
+36-36Lines changed: 36 additions & 36 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -31,7 +31,7 @@ This edition covers what happened during the months of May and June 2026.
31
31
The work was spurred by a real-world report where `git cherry` jobs were each
32
32
doing hundreds of single-blob fetches, at a cost of around 3 seconds
33
33
each, so that batching those downloads should dramatically speed up
34
-
such jobs. As Elijah put it, he "decided to fix up git grep
34
+
such jobs. As Elijah put it, he "decided to fix up `git grep`
35
35
similarly while at it". The series also corrected a small
36
36
documentation typo he had noticed in `patch-ids.h` (a missing
37
37
trailing parenthesis in a comment), as a preparatory fixup.
@@ -64,10 +64,10 @@ This edition covers what happened during the months of May and June 2026.
64
64
`collect_diff_blob_oids()`, lists the blob OIDs touched by a
65
65
commit's diff. It leaves out files that are explicitly marked as
66
66
binary in the userdiff configuration, because for those files
67
-
patch-ID just hashes the OID with `oid_to_hex()` instead of
68
-
reading the blob, so there is no point downloading it.
67
+
the `patch_id` just hashes the OID with `oid_to_hex()` instead of
68
+
reading the blob, so there is no point downloading them.
69
69
70
-
While git cherry relies on hashmap comparisons, the `git grep` patch
70
+
While `git cherry` relies on hashmap comparisons, the `git grep` patch
71
71
takes an analogous but simpler approach: it adds a preliminary walk
72
72
over the tree (similar to `grep_tree()`) that collects the blobs of
73
73
interest and prefetches them in one go.
@@ -76,8 +76,8 @@ This edition covers what happened during the months of May and June 2026.
76
76
spotted something that did not belong: the series added a 210-line
77
77
`investigations/cherry-prefetch-design-spec.md` file to the
78
78
project. He pointed out that, as a document describing how
79
-
`git cherry` works, "it is vastly lacking", that much of its content
80
-
is the sort of material that would normally go in a commit message,
79
+
`git cherry` works, it was "vastly lacking", that much of its content
80
+
was the sort of material that would normally go into a commit message,
81
81
and that he was "not sure how others would benefit from being able
82
82
to read it" once the series landed. Elijah's reply was short and to
83
83
the point: "Ugh, no, sorry." That stray file had been committed by
@@ -240,7 +240,7 @@ This edition covers what happened during the months of May and June 2026.
240
240
improved documentation. The thread also clarified the boundaries of
241
241
partial-clone friendliness for cherry-pick detection, leaving the
242
242
door open for sharing the new code with `git rebase` and
243
-
`git log --cherry-pick` should someone wish to carry that work
243
+
`git log --cherry-pick`, should someone wish to carry that work
244
244
forward.
245
245
246
246
<!---
@@ -288,7 +288,7 @@ __Various__
288
288
[II](https://replicated.wiki/blog/partII.html),
289
289
[III](https://replicated.wiki/blog/partIII.html))
290
290
and [Beagle: git, URIs and all the dirty words](https://replicated.wiki/blog/uris.html).<br>
291
-
[Beagle SCM](https://replicated.wiki/) intends to be a git-compatible LLM-age source code management system.
291
+
[Beagle SCM](https://replicated.wiki/) intends to be a Git-compatible LLM-age source code management system.
292
292
+[Software Is Made Between Commits](https://zed.dev/blog/introducing-deltadb)
293
293
by Nathan Sobo on Zed editor blog,
294
294
about [DeltaDB](https://zed.dev/deltadb), a version control system (in beta)
@@ -307,7 +307,7 @@ __Various__
307
307
+[How to make best use of git and GitHub for AI-assisted software development](https://blog.jonudell.net/2026/06/02/how-to-make-best-use-of-git-and-github-for-ai-assisted-software-development/)
308
308
by Jon Udell on his blog,
309
309
about [Bram](https://github.com/judell/bram) (Bram runs agents mindfully),
310
-
a desktop app that helps you make best use of git and GitHub
310
+
a desktop app that helps you make best use of Git and GitHub
311
311
for AI-assisted software development.
312
312
313
313
@@ -326,7 +326,7 @@ __Light reading__
326
326
[in-toto](https://in-toto.io/) - which protects the build pipeline; and
327
327
[Sigstore](https://www.sigstore.dev/) - which allows to remove long-lived keys
328
328
and for the developer to authenticate with <abbrtitle="OpenID Connect">OIDC</abbr>
329
-
identity you already have, GitHub Actions or Google, etc.,
329
+
identities you already have, like GitHub Actions, Google, etc.,
330
330
where [Fulcio](https://github.com/sigstore/fulcio) issues
331
331
a short-lived code-signing certificate bound to that OIDC identity valid for ten minutes,
332
332
and the signature and cert go into [Sigstore Rekor](https://github.com/sigstore/rekor),
@@ -393,7 +393,7 @@ __Light reading__
393
393
+[Git merges can be better](https://brandondong.github.io/blog/git_merges_can_be_better/),
394
394
on the trick one can use to ensure that the order of branches in the conflict
395
395
is the same in the (tricked-out) merge as it is in rebase.
0 commit comments