2 - Teach 'got merge' to merge changes into an arbitrary subdirectory
3 of the work tree. This would be nice for merging vendor branches.
4 Say you have a branch 'llvm-12' which intially contains a 12.0
5 release tree as published by the LLVM project, added to the repository
6 with a command such as 'got import -b llvm-12'. On the main branch we
7 would want to merge files from the llvm-12 branch into /usr/src/gnu/llvm
8 instead of the root directory checked out at /usr/src.
9 The next LLVM release 12.1 would later be committed onto the llvm-12
10 branch and then merged into main at /usr/src/gnu/llvm in the same way.
11 - When a clone fails the HEAD symref will always point to "refs/heads/main"
12 (ie. the internal default HEAD symref of Got). Resuming a failed clone with
13 'got fetch' is supposed to work. To make this easier, if the HEAD symref
14 points to a non-existent reference it should be updated by 'got fetch'
15 to match the HEAD symref sent by the server.
16 - If invoked in a work tree, got fetch could default to fetching the work
17 tree's current branch, instead of fetching the remote repository's HEAD.
18 - 'got patch' should be able to detect an already applied patch.
19 - 'got patch' should ideally do more passes if a patch doesn't apply and
20 try fancy things (like ignoring context and whitespaces) only in later
22 - investigate whether it's worth for 'got patch' to memory-map the files to
23 edit. (c.f. Plan A / Plan B in Larry' patch.)
24 - when fetching pack files got should verify that the requested branch tips
25 are present in the pack file sent by the server, before making this pack
26 file visible to readers of the repository
27 - enforce that log messages are encoded in either ASCII or UTF-8; this
28 mostly matters for -portable since OpenBSD doesn't provide other locales
31 - add http(s) transport with libtls, speaking the two Git HTTP protocols
32 (both "dumb" and "smart" need to work) via got-fetch-pack, or a new helper
33 like got-fetch-http; it is fine if HTTP remains a fetch-only protocol, and
34 works only against servers which don't require authentication for fetches;
35 anything beyond this would require a full-featured HTTP client (Git uses
36 libcurl, which we cannot use as it is not in the OpenBSD base system)
39 - make 'tog log' respond to key presses while 'loading...' history; loading
40 can be slow for paths in a deep history if the path has not been changed
41 very often, and 'tog log' blocks far too long in this case
42 - make it possible to view the contents of tag objects
43 - verify signed tag objects
44 - make it possible to toggle the parent to diff against in merge commits
47 - fix COMMITS page for paths that were deleted and/or re-added to the
48 repository. One way would be not to let the commit graph filter paths.
49 As an additional optimization we could keep a tailq or the object-id
50 set for everything traversed in the repo to have fast reverse-lookups.
51 (has the additional requirement to invalidate it when the reference
53 - reply with 404 on some kind of errors ('reference not found' for sure,
54 maybe also tree entry not found?)
55 - support category grouping a-la gitweb/cgit with the gitweb.category
56 config option or via the "category" file in the root of the repo.
57 - consider changing the URL scheme to avoid so many query parameters
60 - ensure all error messages are propagated to clients before disconnecting,
61 there are probably still some cases where such error reporting is skipped
62 - client connection timeout handling needs to be checked by regress tests,
63 and is likely in need of improvement
64 - implement stress-tests to observe and fix misbehaviour under load
65 - listener's fd-reserve limit needs to be reviewed and perhaps adjusted
66 - implement pre-commit checks (in lieu of hook scripts):
67 1. deny branch history rewriting ('got send -f') via gotd.conf [done]
68 2. allow/deny creation/deletion of references via gotd.conf
69 3. deny modifications within a given reference namespace via gotd.conf [done]
70 4. entirely hide a given reference namespace from clients via gotd.conf
71 5. allow/deny addition of binary files to a repo via gotd.conf
72 6. enforce a particular blob size limit via gotd.conf
73 7. optionally reject merge commits via gotd.conf
74 - implement post-commit-event libexec handlers (in lieu of hook scripts):
75 1. commit email notification, plaintext smtp to localhost port 25 [done]
76 2. general-purpose HTTP(s) GET/POST request as commit notification,
77 should use a format string to encode commit-info in the request URL
78 3. perform the equivalent of 'got send' to another repository
79 - keep track of available repository disk space and fail gracefully when
80 uploaded pack files would fill up the disk too much, keeping a reserve
81 - reuse packed non-delta objects directly (without re-deltification) for speed
82 - if no permit statement is given in gotd.conf it is possible to send a branch
83 into an empty repository regardless; this should return EPERM instead
86 - add support for generating git-fast-export streams from a repository
87 - add support for importing git-fast-export streams into a repository
88 - speed up 'gotadmin pack -a' is too slow on repositories with many pack files
