[Rpm-maint] [PATCH 2/5] Only build bundled fts if system has a bad version that doesn't handle LFS

[Panu Matilainen]

On 03/24/2017 12:13 PM, Mark Wielaard wrote:
> On Fri, 2017-03-24 at 11:15 +0200, Panu Matilainen wrote:
>> On 03/23/2017 08:21 PM, Gleb Fotengauer-Malinovskiy wrote:
>>> diff --git a/configure.ac b/configure.ac
>>> index bdcb741..687d58c 100644
>>> --- a/configure.ac
>>> +++ b/configure.ac
>>> @@ -717,6 +717,9 @@ AC_CHECK_FUNCS(
>>>     [mkstemp getcwd basename dirname realpath setenv unsetenv regcomp lchown utimes],
>>>     [], [AC_MSG_ERROR([function required by rpm])])
>>>
>>> +AC_CHECK_HEADERS([fts.h])
>>> +AM_CONDITIONAL([USE_BUNDLED_FTS_KLUDGE], [test "$ac_cv_header_fts_h" = no])
>>> +
>>>  AC_LIBOBJ(fnmatch)
>>>
>>
>> How exactly does this ensure the system fts.h is LFS compatible?
>
> The configure.ac has an AC_SYS_LARGEFILE fairly early. Which will define
> _FILE_OFFSET_BITS 64 on 32-bit systems that need it. So on such systems
> when AC_CHECK_HEADERS([fts.h]) is ran it will have that define in the
> conftest.c and the test compile will fail because old fts.h have an
> explicit check (error out on) -D_FILE_OFFSET_BITS==64.

Ah... thanks for the explanation.

That does deserve a comment though, it's not exactly obvious :)


	- Panu -