2.6x..4

The stable tester's forum for ZeosLib 7.0.x series

Report problems concerning our Delphi 2009+ version and new Zeoslib 7.0 features here.
Post Reply
marcov
Senior Boarder
Senior Boarder
Posts: 95
Joined: 24.06.2010, 09:17

2.6x..4

Post by marcov »

Some FPC updates:

- FPC 2.6.4 is coming near New Year, with a RC1 within a couple of weeks. I saw that Zeos 7.0.3 still assumed TBookmark changes would become active after 2.6.2, this has been postponed to the next major release (of the 2.7 branch).
So the version check tbookmark/trecordbuffer should read >20700 not 20602 or so

- I ported my work Zeos app to lazarus (with 703), and had strange string related crashes with FPC 2.7.x. (in the code that reads out the columns). I had some confirmation that this problem started in the last couple of weeks.
IOW it is possible fpc 2.7.1 breaks zeos atm, use a stable as much as possible.
User avatar
EgonHugeist
Zeos Project Manager
Zeos Project Manager
Posts: 1936
Joined: 31.03.2011, 22:38

Re: 2.6x..4

Post by EgonHugeist »

@Marcov
marcov wrote:- FPC 2.6.4 is coming near New Year, with a RC1 within a couple of weeks. I saw that Zeos 7.0.3 still assumed TBookmark changes would become active after 2.6.2, this has been postponed to the next major release (of the 2.7 branch).
So the version check tbookmark/trecordbuffer should read >20700 not 20602 or so
Ok, we'll change your define to >20700.
marcov wrote: I ported my work Zeos app to lazarus (with 703), and had strange string related crashes with FPC 2.7.x. (in the code that reads out the columns). I had some confirmation that this problem started in the last couple of weeks.
IOW it is possible fpc 2.7.1 breaks zeos atm, use a stable as much as possible.
Yes i've noticed some threads too. Can you tell me what did change in this domain? Also the Parameters seems to be involved inbetween. Actually i'm writing a faster Zeos way for the Fields. But this also meens Zeos has it's own TField descandants, which will break compatibility with 7.1down. The new fields can be activated by a define in Zeos.inc: uncomment the WITH_TZSTRINGFIELDS define. This should also resolve this issue. But i'm wondering why FPC is breaking the Delphi compatiblity here. 7.2 performes imbelievable faster and if these Fields came by default i could also fix the missing TArrayField, TDataSetField, TShortIntField, TWordField etc.
I did start, but there is loads of work left to get a final state...

So what's wrong here or what needs to be changed to keep the compatibility?
Best regards, Michael

You want to help? http://zeoslib.sourceforge.net/viewtopic.php?f=4&t=3671
You found a (possible) bug? Use the new bugtracker dude! http://sourceforge.net/p/zeoslib/tickets/

Image
marcov
Senior Boarder
Senior Boarder
Posts: 95
Joined: 24.06.2010, 09:17

Re: 2.6x..4

Post by marcov »

EgonHugeist wrote:@Marcov
Yes i've noticed some threads too. Can you tell me what did change in this domain? Also the Parameters seems to be involved inbetween.
I think it is simply a compiler bug, which will be found sooner or later.
Actually i'm writing a faster Zeos way for the Fields. But this also meens Zeos has it's own TField descandants, which will break compatibility with 7.1down. The new fields can be activated by a define in Zeos.inc: uncomment the WITH_TZSTRINGFIELDS define. This should also resolve this issue.
I'm not sure what this problem is supposed to fix. Breaking compatibility sounds horrible.
But i'm wondering why FPC is breaking the Delphi compatiblity here. 7.2 performes imbelievable faster and if these Fields came by default i could also fix the missing TArrayField, TDataSetField, TShortIntField, TWordField etc.
I did start, but there is loads of work left to get a final state...
I'm not sure what exactly is broken.
marcov
Senior Boarder
Senior Boarder
Posts: 95
Joined: 24.06.2010, 09:17

Re: 2.6x..4

Post by marcov »

I haven't had time to rebuild and test yet, but it might be that another Zeos user (Mr. Do-wan Kim) found and devised a patch for the problem plaguing 2.7.1

http://bugs.freepascal.org/view.php?id=25269

The routine he says he got the crash has the same name as the one I got a crash in iirc. (but then in postgresql code)
Post Reply