Re: [PATCH 03/14] lockfile: remove some redundant functions

2 messages, 2 authors, 2016-06-15 · open the first message on its own page

Re: [PATCH 03/14] lockfile: remove some redundant functions

From: Junio C Hamano <hidden>
Date: 2016-06-15 23:05:15

Michael Haggerty [off-list ref] writes:
Remove the following functions and rewrite their callers to use the
equivalent tempfile functions directly:

* fdopen_lock_file() -> fdopen_tempfile()
* reopen_lock_file() -> reopen_tempfile()
* close_lock_file() -> close_tempfile()
Hmph, 

My knee-jerk reaction was "I thought lockfile abstraction was
fine---why do we expose the implementation detail of the lockfile,
which is now happen to be implemented on top of the tempfile API, to
the callers?"  I guess that was also where my comments on 02/14 "why
do callers have to include tempfile.h separately?" came from.

I'm undecided but would trust your judgement until I read thru to
the end of the series ;-).

Re: [PATCH 03/14] lockfile: remove some redundant functions

From: Johannes Sixt <hidden>
Date: 2016-06-15 23:05:15

Am 10.06.2015 um 19:40 schrieb Junio C Hamano:
Michael Haggerty [off-list ref] writes:
quoted
Remove the following functions and rewrite their callers to use the
equivalent tempfile functions directly:

* fdopen_lock_file() -> fdopen_tempfile()
* reopen_lock_file() -> reopen_tempfile()
* close_lock_file() -> close_tempfile()
Hmph,

My knee-jerk reaction was "I thought lockfile abstraction was
fine---why do we expose the implementation detail of the lockfile,
which is now happen to be implemented on top of the tempfile API, to
the callers?"
Just for the record, I had exactly the same reaction, and I find this 
transition against the spirit of a self-contained lockfile API.

-- Hannes
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help