Re: Just did a gitpull got an update diff shows 1464 lines?

On 4/29/26 09:08, tito via busybox wrote:
> On Wed, 29 Apr 2026 23:18:16 +1000
> "Michael D. Setzer II via busybox" <[email protected]> wrote:
> 
>> Didn't see a message about an update, but git pull found difference.
>> compared the 4/13 snapshot with current after git pull.
>> Result was 1464 lines, so has to be some update.
>> Perhaps others would know more.

Denys has been showing up a couple times a month to shovel out the backlog.

$ git log --stat --pretty=fuller | grep CommitDate | less
CommitDate: Wed Apr 29 06:36:39 2026 +0200
CommitDate: Wed Apr 29 06:31:55 2026 +0200
CommitDate: Sun Apr 12 16:20:09 2026 +0200
CommitDate: Mon Mar 16 14:49:42 2026 +0100
CommitDate: Sun Mar 15 16:35:32 2026 +0100
CommitDate: Sun Mar 15 16:12:23 2026 +0100
CommitDate: Fri Mar 13 00:46:19 2026 +0100
CommitDate: Thu Mar 12 13:23:48 2026 +0100
CommitDate: Thu Mar 12 07:25:38 2026 +0100
CommitDate: Mon Mar 9 15:23:59 2026 +0100
CommitDate: Thu Mar 5 21:46:04 2026 +0100
CommitDate: Sat Feb 28 11:25:05 2026 +0100

The March 16 to April 12 gap was longer than the April 12 to April 29 
gap. Not particularly alarming.

>> Found this with diffstat
>> cat diffstat

Did you run the "git diff" output through diffstat, or did you run the 
_diff_ output through diffstat and get piles of noise from .git internal 
files?

$ git diff 2f3b16fc1b46..master | diffstat
  archival/libarchive/Kbuild.src         |    6 +++---
  archival/libarchive/data_extract_all.c |    4 ++++
  archival/tar.c                         |    4 ++++
  libbb/ioloop.c                         |    1 +
  networking/httpd.c                     |    6 +++---
  scripts/randomtest                     |    4 ++++
  scripts/randomtest.loop                |    5 +++++
  shell/hush.c                           |    8 ++++----
  testsuite/cpio.tests                   |   12 +++++++-----
  testsuite/start-stop-daemon.tests      |    2 ++
  testsuite/tar.tests                    |    2 +-
  util-linux/fdisk.c                     |    4 +++-
  12 files changed, 41 insertions(+), 17 deletions(-)

Seems reasonable?

>> No change in latest snapshot yet.
>>
>> Just interesting

Looks to me like Denys just tackled his backlog. All commits on 
wednesday and thursday by him and signed-off-by him:

> commit 25b756f789e7dda8048f69f2ab23f9cdb5d78d6f
>     cpio.tests: fix false positive
> commit 3b3d0fc5352ac036520bac74073b9b205a21b7c3
>     cpio.tests: do not use "ls" to detect that file does not exist
> commit 48c7b48f95bbbe5c0a104e906a1260fc7c6280d9
>     tar: unicode fix
> commit 4f11a4b25d5a1093695653d1826ba69eecc9d304
>     cpio.tests: fix false positive when -o support is not enabled
> commit 6908229d4f7f9d3f199d831e09cacd48bb08a97b
>     cpio: fix a bug (wrong free() address) in !FEATURE_TAR_LONG_OPTIONS config
> commit fb10ad3b5cee445418f9422709fd8d2abac23a61
>     start-stop-daemon.tests: fix false positive if longopts are not enabled
> commit c8eb8614068a662271a5187d4876df1d8439a0ae
>     Fix "undefined reference to 'skip_unsafe_prefix' build failure
> commit fdca28aaf4e8ab38381dc5db336aba7bd88ebd62
>     fdisk: fix build failure on glibc
> commit a0a98f070066cf7615082363b997a6fd0d103098
>     Fix randomtest to not select TC on glibc
> commit 7e2b9a4d9c23de1f6e651a5dedd6c6e8e17d4c1e
>     httpd: fix build for !FEATURE_HTTPD_CGI config
> commit a588fb9703a7f1885cc2943729d374800fa0e4fd
>     Fix randomtest to never select a debug config
> commit a9e7466e5cbca5e2bc1402bd801537e1818b9d41
>     telnet: fix build for !TELNETD config
> commit de26b74b1d6597a4b593084b309caf8b4d7fcaec
>     httpd: fix build for !FEATURE_HTTPD_CONFIG_WITH_SCRIPT_INTERPR config
> commit 0519429ea166949f2d8bee9cdbd70bf986b7b683
>     hush: fix build for !HUSH_CASE config

Repository's still at the same place on git.busybox.net, right where 
https://busybox.net/source.html says to clone it from.

> Hi,
> there seems to be some updates on the website, gitweb
> now points to https://github.com/vda-linux/busybox_mirror
> and works correctly and seems also that Denys is alive and well.

Busybox's cgit implementation got taken down because LLM crawlers were 
paralyzing the server. Happened back in january according to the date on 
the directory, dunno if Denys did it or the host (osuosl).

Various people poked me earlier this week about somebody with advanced 
https://en.wikipedia.org/wiki/Chatbot_psychosis spamming the list and 
causing problems, and I asked osuosl to add the moderation bit to that 
account until Denys got back to deal with it. They did so and said they 
emailed him about it, and he apparently gave the project some time 
shortly thereafter.

I also emailed Denys to suggest echo '<meta http-equiv="refresh" 
content="0; url=https://busybox.net/source.html" />' > 
/var/www/git.busybox.net/htdocs/index.html on the server (my ssh key 
still works but I forgot my sudo password, permission denied doing it 
myself) so git.busybox.net would redirect instead of giving a 404 error. 
Doesn't look like he's chosen to do so yet. :)

If he wants to point the GUI at his own mirror on microsoft github so 
they tank the crawlers, that makes sense. (Although personally I'd 
suggest codeberg instead, ala https://codeberg.org/landley/toybox .)

And... where did that change happen? https://git.busybox.net/ still 
gives me the 404 error...

> Ciao,
> Tito
Rob