On Sunday 19 January 2003 12:40 pm, Daniel Ockeloen wrote: > >-----Oorspronkelijk bericht----- > >Van: [EMAIL PROTECTED] > >[mailto:[EMAIL PROTECTED]]Namens Pierre van Rooden > >Verzonden: vrijdag 17 januari 2003 17:00 > >Aan: [EMAIL PROTECTED] > >Onderwerp: Re: images eating memory > > > >Rob van Maris wrote: > >> I don't expect this to be handled well in all cases. This way incomplete > >> nodes may be retrieved (i.e. missing certain fields), resulting > > > >in all kind > > > >> of exceptions. > > > >This will not be a problem as long as the retrieval process make sure to > >put the "$SHORTED" keyword in the byte-type fields (as MMBase actually > >assumes that blobs are not retrived from the database).
The current code as i wrote it is located here http://carlit.mine.nu/~keesj/MMObjectBuilder.java getNonByteArrayFields.. The code any't nice .. but I can't help it. I have tested it with mysql and mmbase 1.6 It would be good to test te code with other database and with old mmbase installations (where the builder.xml's don't extend eacht other) MMObjectNode does skip the field at this point but because the result set still contained the byteArray fields this cause the OutOfMemory error.
