Thread: AIX check in datetime.h
Can someone please explain why in include/utils/datetime.h (struct datetkn)
there is a check for _AIX that either initializes a char* pointer or a char
array? Is there any advantage of a char-array except for warnings of some
compilers if the initilization string is too long?
Apart from that I doubt that AIX cannot handle token[TOKMAXLEN] because
similar declarations can be found in other headers without the _AIX check.
The struct definition is more than 9 years old and seems to show up first in
Attic/dt.h.
#define TOKMAXLEN 10 /* only this many chars are stored in * datetktbl */
/* keep this struct small; it gets used a lot */
typedef struct
{
#if defined(_AIX) char *token;
#else char token[TOKMAXLEN];
#endif /* _AIX */ char type; char value; /* this may be unsigned, alas */
} datetkn;
Joachim
Wow, that is strange. We could remove it for 8.2 and see how testing goes.
---------------------------------------------------------------------------
Joachim Wieland wrote:
> Can someone please explain why in include/utils/datetime.h (struct datetkn)
> there is a check for _AIX that either initializes a char* pointer or a char
> array? Is there any advantage of a char-array except for warnings of some
> compilers if the initilization string is too long?
>
> Apart from that I doubt that AIX cannot handle token[TOKMAXLEN] because
> similar declarations can be found in other headers without the _AIX check.
>
> The struct definition is more than 9 years old and seems to show up first in
> Attic/dt.h.
>
>
> #define TOKMAXLEN 10 /* only this many chars are stored in
> * datetktbl */
>
> /* keep this struct small; it gets used a lot */
> typedef struct
> {
> #if defined(_AIX)
> char *token;
> #else
> char token[TOKMAXLEN];
> #endif /* _AIX */
> char type;
> char value; /* this may be unsigned, alas */
> } datetkn;
>
>
> Joachim
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 3: Have you checked our extensive FAQ?
>
> http://www.postgresql.org/docs/faq
>
-- Bruce Momjian http://candle.pha.pa.us EnterpriseDB http://www.enterprisedb.com
+ If your life is a hard drive, Christ can be your backup. +
Bruce Momjian <pgman@candle.pha.pa.us> writes: > Joachim Wieland wrote: >> Can someone please explain why in include/utils/datetime.h (struct datetkn) >> there is a check for _AIX that either initializes a char* pointer or a char >> array? > Wow, that is strange. We could remove it for 8.2 and see how testing goes. It looks like a workaround for some ancient compiler problem. [ digs for awhile... ] Very ancient: we inherited that hack from Berkeley, see http://developer.postgresql.org/cvsweb.cgi/pgsql/src/backend/utils/Attic/nabstime.h I bet we can remove it. regards, tom lane
I wrote:
> It looks like a workaround for some ancient compiler problem. [ digs
> for awhile... ] Very ancient: we inherited that hack from Berkeley,
In fact, now that I know where to look, I find the same thing in the
postgres-v4r2 tarball, which means the hack is pre-1994. I don't have
anything older to look at.
regards, tom lane
Tom Lane wrote: > Bruce Momjian <pgman@candle.pha.pa.us> writes: > > Joachim Wieland wrote: > >> Can someone please explain why in include/utils/datetime.h (struct datetkn) > >> there is a check for _AIX that either initializes a char* pointer or a char > >> array? > > > Wow, that is strange. We could remove it for 8.2 and see how testing goes. > > It looks like a workaround for some ancient compiler problem. [ digs > for awhile... ] Very ancient: we inherited that hack from Berkeley, see > http://developer.postgresql.org/cvsweb.cgi/pgsql/src/backend/utils/Attic/nabstime.h > > I bet we can remove it. OK, removed. Let's see if we get failure feedback. -- Bruce Momjian http://candle.pha.pa.us EnterpriseDB http://www.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +
In an attempt to throw the authorities off his trail, pgman@candle.pha.pa.us (Bruce Momjian) transmitted: > Tom Lane wrote: >> Bruce Momjian <pgman@candle.pha.pa.us> writes: >> > Joachim Wieland wrote: >> >> Can someone please explain why in include/utils/datetime.h (struct datetkn) >> >> there is a check for _AIX that either initializes a char* pointer or a char >> >> array? >> >> > Wow, that is strange. We could remove it for 8.2 and see how testing goes. >> >> It looks like a workaround for some ancient compiler problem. [ digs >> for awhile... ] Very ancient: we inherited that hack from Berkeley, see >> http://developer.postgresql.org/cvsweb.cgi/pgsql/src/backend/utils/Attic/nabstime.h >> >> I bet we can remove it. > > OK, removed. Let's see if we get failure feedback. I haven't been monitoring CVS HEAD, but you can be sure this will get tried out when 8.2 gets anywhere vaguely close to relese... -- If this was helpful, <http://svcs.affero.net/rm.php?r=cbbrowne> rate me http://linuxdatabases.info/info/slony.html Signs of a Klingon Programmer - 8. "Debugging? Klingons do not debug. Our software does not coddle the weak. Bugs are good for building character in the user."
Christopher Browne wrote: > I haven't been monitoring CVS HEAD, but you can be sure this will get > tried out when 8.2 gets anywhere vaguely close to relese... > The whole point of having a buildfarm is that we shouldn't have to wait, we should be able to see very quickly if we have broken something. We currently have AIX coverage for 5.2/ppc with both gcc and (I think) IBM cc. If we need more coverage then feel free to add other AIX machines. cheers andrew
Andrew Dunstan <andrew@dunslane.net> writes:
> The whole point of having a buildfarm is that we shouldn't have to wait,
> we should be able to see very quickly if we have broken something.
And in fact kookaburra seems happy ...
regards, tom lane
Martha Stewart called it a Good Thing when andrew@dunslane.net (Andrew Dunstan) wrote:
> Christopher Browne wrote:
>> I haven't been monitoring CVS HEAD, but you can be sure this will get
>> tried out when 8.2 gets anywhere vaguely close to relese...
>>
>
> The whole point of having a buildfarm is that we shouldn't have to
> wait, we should be able to see very quickly if we have broken
> something.
>
> We currently have AIX coverage for 5.2/ppc with both gcc and (I think)
> IBM cc. If we need more coverage then feel free to add other AIX
> machines.
I'd very much to add an AIX 5.3 system; that's awaiting some sysadmin
activity that is more urgent...
--
output = ("cbbrowne" "@" "gmail.com")
http://linuxdatabases.info/info/internet.html
``What this means is that when people say, "The X11 folks should have
done this, done that, or included this or that", they really should be
saying "Hey, the X11 people were smart enough to allow me to add this,
that and the other myself."'' -- David B. Lewis <dbl@motifzone.com>