Commits
- Commit:
97ddbd299ba74881d7fda4bc0588da7b494ed1af
- From:
- Omar Polo <op@omarpolo.com>
- Date:
changing the strategy for casts
ok stsp@
- Commit:
17f7cea3a2932bd7836df7a03f41db2224a794ae
- From:
- Omar Polo <op@omarpolo.com>
- Date:
gotd: group the flags for re-exec under -T
ok stsp@
- Commit:
eed4085c0f4916f79731c3fc1a8adeb149b59a69
- From:
- Stefan Sperling <stsp@stsp.name>
- Date:
restore abort() calls in lib/hash.c
In the spirit of commit b808e01085d775b284878dfeea5c06339a9f2ec2, restore
the abort() call in got_hash_final_object_id() and add additional ones to
cover other functions in this file.
ok op@
- Commit:
a08fda08dfda5473c47e3d6d0d01fc62d08dd5ec
- From:
- Mark Jamsek <mark@jamsek.dev>
- Date:
insert wait instruction in tog log mark keymap test
On slower machines, the test may fail due to capturing
the screendump before the base commit marker is drawn.
ok op@
- Commit:
7c4b3c1b4735c79a3c26e1efd4b170865d001eb0
- From:
- Stefan Sperling <stsp@stsp.name>
- Date:
quote the hmac secret in test script
- Commit:
46de6cac87425f5bdb8de5af31a62918f9c0300b
- From:
- Stefan Sperling <stsp@stsp.name>
- Date:
make hmac failures detectable by regression tests
- Commit:
7c51647ef180ad6b67d2ac51c60babb3026f65f2
- From:
- Stefan Sperling <stsp@stsp.name>
- Date:
use a proper secret for tests
- Commit:
7b720cf9c9925da29f85e309de7a4df7e2604c4e
- From:
- Stefan Sperling <stsp@stsp.name>
- Date:
apply a Perl idion tweak suggested by op@
- Commit:
7ddb9e12ddb8b4528b6db5c9105ad83adae17765
- From:
- Stefan Sperling <stsp@stsp.name>
- Date:
fix hmac error logging; suggestions by op@
- Commit:
4615fe93f6be3799737f485f025f1ec9b89586f5
- From:
- Stefan Sperling <stsp@stsp.name>
- Date:
rename gotd HTTP notification signature header; suggested by op@
- Commit:
bd84be8a1a59cfed38bb44a3d921b3b16ffa6469
- From:
- Stefan Sperling <stsp@stsp.name>
- Date:
add support for HMAC digests to gotd HTTP notifications
Works similar to how several Git forges authenticate their webhooks.
Feature requested by dch@freebsd.
Help from tb@ regarding which libcrypto calls to use, thanks!
ok op@ (with some follow-up tweaks suggested)
- Commit:
8cc4eb801418181a7eddf2ad28d85b4e60661ae7
- From:
- Mark Jamsek <mark@jamsek.dev>
- Date:
gotwebd: show commit id prefix on briefs page
tweak + ok op@, "seems good to me" (webpage review) stsp@
- Commit:
09d38fee11b9a19a2baef6995366389723745baa
- From:
- Christian Weisgerber <naddy@mips.inka.de>
- Date:
tog: show shortened "id1 -> id2" as path in diffstat when diffing blobs
This matches the output of "got diff -d".
ok jamsek
- Commit:
78a2ad6b7632ff7b206518772a6aa2f1fcfa7f40
- From:
- Mark Jamsek <mark@jamsek.dev>
- Date:
tog: expand diffstat to all diff views
Suggested by naddy: display diffstat in diff views of arbitrary commits.
For the sake of consistency, show it in tree and blob diffs now too.
Adjust and grow regress to cover the change.
ok stsp@
- Commit:
849e5cb6301f6cbcfc5c43d068c0b7c7507c4e00
- From:
- Mark Jamsek <mark@jamsek.dev>
- Date:
regress for tog diff view 'p' keymap; ok stsp@
- Commit:
bc7dcd6603946972172d2e08412488029d34759a
- From:
- Mark Jamsek <mark@jamsek.dev>
- Date:
tog diff 'p' keymap: show id prefix in diff filename
This helps identify diffs when several have been written.
Suggested by stsp on IRC.
ok stsp@
- Commit:
d6cf760a654df954418c577a08765155985c095d
- From:
- Stefan Sperling <stsp@stsp.name>
- Date:
move pack cache entries to front upon any access, not just when cache is full
- Commit:
e11883f02b06fcd5d68da0aa0b781f93b3e94568
- From:
- Mark Jamsek <mark@jamsek.dev>
- Date:
tog: add diff view 'p' keymap to write the diff to file
Write the current diff view to a temporary patch file in /tmp and report
the pathname to the status line.
discussed with naddy
input and ok stsp@
- Commit:
cd2a3010322415cd187fe81189a32e49cac09672
- From:
- Stefan Sperling <stsp@stsp.name>
- Date:
add a todo item for 'got rebase'; prompted by a question from tb@
- Commit:
e5eb6d2d0fae51188a4650e49d0ad3866d3f3fb5
- From:
- Mark Jamsek <mark@jamsek.dev>
- Date:
zap redundant free
This pointer is owned by the caller and is freed in error paths via
close_diff_view()
ok stsp@
- Commit:
f1f0a7dc342e513a050fee6988a9de61769a26a1
- From:
- Omar Polo <op@omarpolo.com>
- Date:
bump version number
- Commit:
350a84376bb1f7c967234fad40ae5a6a30d0ae33
- From:
- Omar Polo <op@omarpolo.com>
- Date:
CHANGES for 0.102
- Commit:
232556d55e4212db6ceba21158a3acce25c77706
- From:
- Omar Polo <op@omarpolo.com>
- Date:
got_worktree_prepare_path_info -> got_worktree_path_info_prepare
- Commit:
ce725b8f7be2659fcbebc5ec00062bac1ca85d29
- From:
- Omar Polo <op@omarpolo.com>
- Date:
got patch: lock the worktree
Since we may update the fileindex, the worktree must be preemptively
locked exclusively. It's an old thing, been there since the start.
ok stsp@
- Commit:
70cea17ff7df1d1c70cefb074ffba2d8749ca4a3
- From:
- Stefan Sperling <stsp@stsp.name>
- Date:
rename got_fileindex_version() to got_fileindex_get_version()