[PATCH v2 0/6] "am" state file fix with write_file() clean-up

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

 



"git am" was recently reimplemented in C.  While the implementation
was done conservatively and followed the original logic in the
scripted version fairly faithfully, the state files it left in the
$GIT_DIR/rebase-apply directory were made slightly different by
mistake---they lacked the final LF, leaving their last line
incomplete.

The patch [1/6] is Peff's idea to consolidate callers in "am", in a
more concrete form.

The patch [2/6] is the fix to the state files with incomplete lines.

The workhorse helper function that implements "we have this (short)
body of text; create a new file that contains it" has a "fatal"
parameter, to which 1 was passed by almost all callers, but to
casual readers, it was unclear what that 1 meant.  The patch [3/6]
splits it to write_file() and write_file_gently() and drops this
parameter that looks mysterious at the callsites.  A common helper
function write_file_v() is introduced to implement these two as thin
wrappers of it.

The patch [4/6] updates write_file_v() so that it does the "we are
writing a text file.  Make sure it does not end with an incomplete
line" logic that [2/6] added only to builtin/am.c, thusly reverting
what was done to builtin/am.c in [2/6].

The patch [5/6] stops all callers that creates a single-liner file
using write_file() and write_file_gently() from including the final
LF to the format they pass.  This should not change the behaviour,
but it probably makes it conceptually cleaner.  You have the contents
to be placed on a single line, and the helper turns the contents
into a proper "line".

The patch [6/6] drops the final LF from the parameter to create a
multi-line file; while this does not hurt in the sense that the
callee will add a necessary LF back, I do not think it should be
applied.  Conceptually, if you have a buffer that contains a bunch
of lines and throw it at a helper to create a file, you'd better
have the terminating LF yourself before asking the helper to put
them in the file.

Junio C Hamano (6):
  builtin/am: introduce write_state_*() helper functions
  builtin/am: make sure state files are text
  write_file(): drop "fatal" parameter
  write_file_v(): do not leave incomplete line at the end
  write_file(): drop caller-supplied LF from calls to create a one-liner
    file
  write_file(): drop caller-supplied LF from multi-line file

 builtin/am.c       | 69 ++++++++++++++++++++++++++++++++----------------------
 builtin/branch.c   |  4 ++--
 builtin/init-db.c  |  2 +-
 builtin/worktree.c | 10 ++++----
 cache.h            |  5 ++--
 daemon.c           |  2 +-
 setup.c            |  2 +-
 submodule.c        |  2 +-
 transport.c        |  2 +-
 wrapper.c          | 36 ++++++++++++++++++++++++----
 10 files changed, 88 insertions(+), 46 deletions(-)

-- 
2.5.0-568-g53a3e28

--
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]