[PATCH 0/9] db/fetch-pack topic fixes

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



The following series is on top of the db/fetch-pack series that is
currently in Junio's 'pu' branch.  I found a number of cases that
weren't quite matching what git-fetch used to do, or were outright
major breakages in the code.

Even with this 9 patch series applied I am not yet certain that the
db/fetch-pack series is ready for 'next'.  I will continue to test
(and fix) the topic through this weekend.  Right now the test suite
is passing, but I'm not sure we have enough coverage on git-fetch
for that to be an assertion of the topic's readiness.


1/9 Correct builtin-fetch to handle + in refspecs
2/9 Fix off by one bug in reflog messages written by builtin-fetch

  These two were outright bugs that I noticed a couple of days
  ago when I started to test the topic.  The first is a major
  functionality problem while the second is really just a minor
  annoyance as the reflog messages were being truncated.

3/9 Remove unnecessary debugging from builtin-fetch

  The older git-fetch.sh did not output these messages.  Any tool
  that intends to replace it should also not output this level of
  verbosity by default.

4/9 Remove unused unpacklimit variable from builtin-fetch

  This code appears to be dead.  I need to look into the code further
  as I am actually suspecting that builtin-fetch does not honor the
  user's {fetch,transfer}.unpackLimit configuration option because it
  does not appear to properly initialize builtin-fetch-pack's statics
  before use.

  Something like this may need to come back, but most likely in a
  different form.  I'm getting too tired to look into it right now
  and will try to pick it up again after getting some sleep.

5/9 Replace custom memory growth allocator with ALLOC_GROW
6/9 Simplify fetch transport API to just one function

  This code was just really hairy and a lot more complex than
  it needed to be.  ALLOC_GROW and this single fetch function
  cleans things up considerably.

7/9 Refactor index-pack "keep $sha1" handling for reuse
8/9 Remove pack.keep after ref updates in git-fetch
9/9 Always ensure the pack.keep file is removed by git-fetch

  If we are using index-pack to receive a packfile (because it
  is over our unpackLimit) we must arrange for the .keep file
  to be removed after the corresponding refs have been updated.
  This is necessary to allow the packfile to be garbage collected
  at a later point in time.

-- 
Shawn.
-
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]

  Powered by Linux