i'm using fpc-3.2.2 on linux this is the synapse version i got on 12/22/22 svn checkout https://svn.code.sf.net/p/synalist/code/trunk synapse checked out revision 266
i didn't know you were working on a 32bit win machine - this fpc/synapse is really cross os :) i dare ask if your computer is only 32bit also? On Wed, 9 Aug 2023 20:41:52 +0200 Rik van Kekem <r...@graficalc.nl> wrote: > Op 09-08-2023 om 20:08 schreef paste...@gmx.com: > > i am actially using -S2gch (which is -Mobjfpc ??) in /etc/fpc.cfg > > there must be something else in /etc/fpc.cfg i'm not seeing > I normally compile with Lazarus but it uses fpc as follows: > > Q:\dev32\fpc\bin\i386-win32\fpc.exe -MObjFPC -Scghi -Cirot -O1 -gw3 -gl > -gh -l -vewnhibq ^ > -FuQ:\dev32\config\onlinepackagemanager\packages\synapse40.1\lib\i386-win32 > ^ > -otestmariadb.exe ^ > testmariadb.lpr > > ^ is continuation of previous line in a command prompt. > This command works for me too in a command prompt. > Even without the adding of byte() in Chr(Command). > > > revision 5 compiles fine but even with your -Mobjfpc header i still had to > > add the 'byte()' > >>> Buffer := Chr(byte(Command)) + SQL; > Command is an unumerated type and I thought you should just be able to > feed that to Chr(). > > But even without options with the following it doesn't complain about > the Chr(Command). > > Q:\dev32\fpc\bin\i386-win32\fpc.exe > -FuQ:\dev32\config\onlinepackagemanager\packages\synapse40.1\lib\i386-win32 > testmariadb.lpr > > I puzzled why that doesn't work for you. > I'll have to test this under Linux to see if it makes any difference. > BTW. Do you have the latest FPC version? > > -S2 ( -Mobjfpc ) is already at the top of the source file. > h for ansistring is also at the top ($H+} > c for support for C operators I never use > > > you asked if i had any suggestions .... are you planming on making this a > > unit library for the synapse code? > In the source is a TMariaDBConnector class. I could put this easily in a > separate unit. > I would probably make this a github repository then. > > I would need to include the BINARY edition then too. This version only > uses TEXT protocol. > This version uses a memory dataset. It will crash when there are > millions of records (too great for memory). > BINARY protocol has the upside of choosing how many records you want per > call. > That way you can 'step' though the resultset instead of getting it > thrown all at once at you. > But the BINARY protocol is a bit more complicated :) > > Grtz, > Rik > _______________________________________________ synalist-public mailing list synalist-public@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/synalist-public
