**
I've just checked in changes to CVS to eliminate all of the warnings when compiling under MinGW 3.4.5, using the -Wall and -pedantic flags. There are some in the AR API header files that I could not eliminate (all regarding extra semi-colons). Those are harmless - unfortunately there is no way to over-ride the EXTERNFUNCTION macro to fix the extra semi-colon, as it is defined in the same file where it is used (arextern.h).
There is also a new "beta 3" distribution with these changes.
There is at least one warning that was reported to me (regarding an implicit typename) that I don't see using MinGW.
All of these warnings were relatively harmless, but it's good to eliminate them.
Dan
From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of McKenzie, James J C-E LCMC HQISEC/L3
Sent: Tuesday, March 07, 2006 8:56 AM
To: arslist@ARSLIST.ORG
Subject: Re: Announcement phparapi - alpha
Dan:
Also it appears that the version of gcc is very old. I will check my Solaris 9 systems here, but I think gcc should at least be 3.3, if not 4.0. Are you testing with all of these?
James McKenzie
-----Original Message-----
From: Action Request System discussion list(ARSList)
[mailto:arslist@ARSLIST.ORG]On Behalf Of Dan Hardy
Sent: Tuesday, March 07, 2006 9:33 AM
To: arslist@ARSLIST.ORG
Subject: Re: Announcement phparapi - alpha
I should be able to get to these today. I did verify that the first is harmless (because the base class rtl::AssignableT's default constructor and copy constructor are identical). I'm not sure exactly what the second warning below implies, but it certainly looks harmless. The third was an oversight.
Dan
From: Action Request System discussion list(ARSList) on behalf of Maureen Korda
Sent: Tue 3/7/2006 6:40 AM
To: arslist@ARSLIST.ORG
Subject: Re: Announcement phparapi - alpha
I found these warnings to be persisitent, and I already passed them on to
Dan.
The first 2 could depend on the warning level and version of the compiler.
I have not found any of them to cause any problem at runtime.
The warnings don't look good since they are encountered often during
compilitation, due to the nature of templates being held in header files.
solaris 2.8 (8) and gcc version 3.2.2.
ListT.h:79: warning: base class `class rtl::AssignableT
'
should be explicitly initialized in the copy constructor
/>ListT.h:118: warning: `rtl::ListT::constiterator'
is implicitly a typename
FieldLimit.h:168: warning: `class rtl::FieldLimit::CharLimitImpl' has
virtual functions but non-virtual destructor
-----Original Message-----
From: Action Request System discussion list(ARSList)
[mailto:arslist@ARSLIST.ORG]On Behalf Of McKenzie, James J C-E LCMC
HQISEC/L3
Sent: 07 March 2006 14:29
To: arslist@ARSLIST.ORG
Subject: Re: Announcement phparapi - alpha
**
Axton:
I see these errors and they should be cleaned up. Unused variables can
cause problems and the non-virtual destructor is usually an oversight on the
part of the programmer.
My other statements do stand as I tried to clean up the OpenOffice.org code
and ended up making a mess (and wasting about a day doing so.) However, it
was very educational.
James McKenzie
-----Original Message-----
From: Action Request System discussion list(ARSList)
UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org
UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org
20060125This posting was submitted with HTML in it 20060125This posting was submitted with HTML in it
Previous Topic
Index