if i showed you what i had done for the password auth i think you would laugh - i got as far as using the XOR but my code didn't look at all like yours (which works nicely)
i tried to do the following but i don't seem to get what is going on here - i'd appreciate if you could show me the error of my ways // following https://mariadb.com/kb/en/com_query/ // this code is put above break; // or do something else here Buffer := 'show databases;'; writeln(Buffer); buffer := Buf2Hex(buffer); Buffer := #03 + Buffer + #0; // #03 com_query header #0 <eof> writeln(Buffer); sock.sendstring(buffer); buffer := sock.recvpacket(2000); writeln(Buffer); //get //show databases; // 73 68 6F 77 20 64 61 74 61 62 61 73 65 73 3B missing #03/com_query header and #0/<eof> ?? //!��#08S01Got packets out of order //Done, press enter to exit On Mon, 07 Aug 2023 06:25:50 +0200 Rik van Kekem <r...@graficalc.nl> wrote: > You can use RecvPacket (it is high level and reads from the linebuffer). But > I think what is meant here is that you can't expect RecvPacket to always get > the entire package. The incoming data could be split up into several parts. > > Each package is preceded by a length (in 3 bytes) and a package number. So > when processing your incoming data you need to keep reading the buffer (keep > doing buffer := buffer + recvpacket(2000) until you are sure you have the > complete data package from the server before proceeding with processing that > package. (You can peek or read in the buffer to see the header for checking > the needed length of the buffer). > > Because the packages for authentication are small, I don't think these would > ever be fragmented. But going further with larger data packages you can't > rely on that fact. Then you'll need to use a buffer mechanisme. > > Maybe there is another way of looking at it. Other views and reading > techniques are welcome 🤗 > > Grtz, > Rik > > > paste...@gmx.com schreef op 7 augustus 2023 02:20:54 CEST: > > > >Lukas would you mind expanding on what you said > > > >'It is stream based protocol, not packet based. So, usage of RecvPacket is a > >big mistake here.' > > > >then it seems to be 'contradicted' by > > > >'on the reply packet structure. It using RecvPacket internally, and receive > >multiple packets if it is needed.' > > > > > _______________________________________________ > synalist-public mailing list > synalist-public@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/synalist-public _______________________________________________ synalist-public mailing list synalist-public@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/synalist-public
