Arfrever Frehtes Taifersar Arahesis wrote:
> 2009-02-09 16:22:36 Hyrum K. Wright napisał(a):
>> Hyrum K. Wright wrote:
>>> Arfrever Frehtes Taifersar Arahesis wrote:
>>>> 2009-02-09 03:38:18 Hyrum K. Wright napisał(a):
>>>>> Arfrever Frehtes Taifersar Arahesis wrote:
>>>>>> Author: arfrever
>>>>>> Date: Sun Feb 8 18:36:08 2009
>>>>>> New Revision: 35748
>>>>>>
>>>>>> Log:
>>>>>> Use '.sql_h' filename extension for headers generated from '.sql' files.
>>>>> Yes, but Why?
>>>> This filename extension (or .sql.h) allows to easier distinguish headers generated from .sql files.
>>> Does the text "This file is automatically generated from foo.sql" not accomplish
>>> the same goal? Why not encode this redundancy in the source file, or even the
>>> source file name (e.g., rep-cache-db.sql) instead of in *two* different
>>> locations in the build system? It seems a bit redundant and cryptic.
>>>
>>> Btw, I had this conversation with somebody in IRC (and maybe should have
>>> recorded it here) when the change was made from .sql.h to .h. I originally felt
>>> that we should munge the output filename somehow, but further consideration has
>>> led me to be against such.
>> One more problem: we lose automatic syntax highlighting with the munged
>> extension.
>
> :set filetype=c
>
>> Can you please further justify this change, or revert it?
>
> This change allows to genericize the 'fast-clean' target in Makefile.in.
Well, I seem to be the only one to cares about painting this bikeshed purple, so
I'll concede. But I do so under protest. :P
-Hyrum
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1129936
Received on 2009-02-09 19:31:37 CET