Hello there,
I just wanted to point you that someone (I didn't have time to check) a broken the MySQL client of pike 7.4 since somewhere between 7.4.10 and today.
You can see the result here : http://pike.ida.liu.se/generated/pikefarm/7.4/284_176/makelog.html (Darwin PowerBook G4 / 667 MHz).
The most strange thing is that 7.5 has MySQL support fixed.
The side effect about this is that the next fink submission I will sent in the next days will loose MySQL support...
/Xavier
That's Marcus Comstedts fixes to get mysql support for windows that breaks, it looks like. I don't know what the right fix is, but a working fix for mac should be to remove all the prototypes in dummy_dum_dum from ldiv to fgets.
/ Martin Nilsson (gitarrmusik)
Previous text:
2003-06-20 09:23: Subject: Pike 7.4 and MySQL ?
Hello there,
I just wanted to point you that someone (I didn't have time to check) a broken the MySQL client of pike 7.4 since somewhere between 7.4.10 and today.
You can see the result here : http://pike.ida.liu.se/generated/pikefarm/7.4/284_176/makelog.html (Darwin PowerBook G4 / 667 MHz).
The most strange thing is that 7.5 has MySQL support fixed.
The side effect about this is that the next fink submission I will sent in the next days will loose MySQL support...
/Xavier
/ Brevbäraren
What prototypes?
Anyway, it seems like "fileno" is defined as a macro, without the structure that the macro depends on being defined. This is a bug in the header files in my opinion, but fortunately it should be simple enough to work around. Just look in the manpage for "fileno" on MacOS X which header files are required for fileno(), and add those to dummy.c.
/ Marcus Comstedt (ACROSS) (Hail Ilpalazzo!)
Previous text:
2003-06-20 12:23: Subject: Pike 7.4 and MySQL ?
That's Marcus Comstedts fixes to get mysql support for windows that breaks, it looks like. I don't know what the right fix is, but a working fix for mac should be to remove all the prototypes in dummy_dum_dum from ldiv to fgets.
/ Martin Nilsson (gitarrmusik)
The idea was interresting, but it doesn't fix anything.... worse it break all xenofarms.
Le dimanche, 22 jun 2003, à 23:40 Europe/Paris, Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum a écrit :
What prototypes?
Anyway, it seems like "fileno" is defined as a macro, without the structure that the macro depends on being defined. This is a bug in the header files in my opinion, but fortunately it should be simple enough to work around. Just look in the manpage for "fileno" on MacOS X which header files are required for fileno(), and add those to dummy.c.
No, not all of them. Just those running FreeBSD. Interresting...
/ Marcus Comstedt (ACROSS) (Hail Ilpalazzo!)
Previous text:
2003-06-23 23:16: Subject: Re: Pike 7.4 and MySQL ?
The idea was interresting, but it doesn't fix anything.... worse it break all xenofarms.
Le dimanche, 22 jun 2003, à 23:40 Europe/Paris, Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum a écrit :
What prototypes?
Anyway, it seems like "fileno" is defined as a macro, without the structure that the macro depends on being defined. This is a bug in the header files in my opinion, but fortunately it should be simple enough to work around. Just look in the manpage for "fileno" on MacOS X which header files are required for fileno(), and add those to dummy.c.
/ Brevbäraren
In the last episode (Jun 23), Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum said:
No, not all of them. Just those running FreeBSD. Interresting...
Because on FreeBSD 4.x, stdio.h has this:
#define fileno(p) fileno_unlocked(p) #define fileno_unlocked(p) __sfileno(p) #define __sfileno(p) ((p)->_file)
, and since none of the macros cast p to FILE*, fileno(0) fails trying to index 0->_file :)
Changing the line in dummy.c to "fileno((FILE*)0);" should work around the problem.
Le mardi, 24 jun 2003, à 06:41 Europe/Paris, Dan Nelson a écrit :
In the last episode (Jun 23), Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum said:
No, not all of them. Just those running FreeBSD. Interresting...
Because on FreeBSD 4.x, stdio.h has this:
#define fileno(p) fileno_unlocked(p) #define fileno_unlocked(p) __sfileno(p) #define __sfileno(p) ((p)->_file)
Same on Darwin / OS X 6.x
, and since none of the macros cast p to FILE*, fileno(0) fails trying to index 0->_file :)
Changing the line in dummy.c to "fileno((FILE*)0);" should work around the problem.
Well let's try :)
/Xavier
pike-devel@lists.lysator.liu.se
-
Dan Nelson -
Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) ＠ Pike (-) developers forum -
Martin Nilsson (gitarrmusik) ＠ Pike (-) developers forum
-
Xavier Beaudouin