【转帖】problem with code
 

problem with code
problem with code
i have a problem here.
i am iterating the entities, and ascertaining a dummy layer (for now). the idea is to create that layer (or get its id if it already exists) and update the entity to use that layer. and to set the colour as bylayer (currently remmed out).
if i rem out the code that modifes the entity layer and colour, the resulting file loads into autocad and i get all the layers created correctly, with all the entities on their native layers.
but if i put back in the code to actually update the entity, the program runs, but the file will not load into autocad.
it has just occurred to me that polylines would need their vertices updating, but i think the manual says setlayer will cascade layer id down into sub entities).
what am i doing wrong?
code:
if( ucount > 0 )
{
// get modelspace id
idmodelspace = pdb->getmodelspaceid();
// get modelspace entities (for modification)
pblock = idmodelspace.safeopenobject( oddb::kforwrite );
// iterate modelspace entities
for( pentiter = pblock->newiterator(), uentity = 0; !pentiter->done(); pentiter->step(), uentity++ )
{
// update progress complete
ipercent = calculatepercent( uentity + 1, ucount );
setprogresspos( _t("filtering entities..."), ipercent );
// update entity layer and colour
// for now, we just output details to screen
pent = pentiter->entity( oddb::kforwrite );
// this is temporary. we will eventually have a
// lookup map to locate correct layer name based
// on layer / colour pairs
strlayer.format(_t("layer_%s_colour_%d"), pent->layer(), pent->colorindex() );
// select (and create if required) the correct layer
idlayer = createlayer( pdb, strlayer, pent->colorindex() );
pent->setlayer( idlayer );
// pent->setcolorindex( odcmentitycolor::kbylayer );
}
savefile( pdb, m_strfolder );
}
thanks. i have tried to sort it to no avail.
andrew
here is some more information.
if i modify the drawing as shown in code, but save file as dxf, the dxf opens up in autocad with no errors.
it is the dwg saved files that cause the problem.
andrew
hi andrew,
i'm not sure whether this can cause a problem:
but you're trying to save the database with the model space being opened for write. actually there is no need to open it for write in this case (though for your "real" code you need it).
can you try just "closing" the iteration to ensure all the objects are closed:
i.e.
code:
if( ucount > 0 )
{
// get modelspace id
idmodelspace = pdb->getmodelspaceid();
{
// get modelspace entities (for modification)
pblock = idmodelspace.safeopenobject();
// iterate modelspace entities
for( pentiter = pblock->newiterator(), uentity = 0; !pentiter->done(); pentiter->step(), uentity++ )
{
// update progress complete
ipercent = calculatepercent( uentity + 1, ucount );
setprogresspos( _t("filtering entities..."), ipercent );
// update entity layer and colour
// for now, we just output details to screen
pent = pentiter->entity( oddb::kforwrite );
// this is temporary. we will eventually have a
// lookup map to locate correct layer name based
// on layer / colour pairs
strlayer.format(_t("layer_%s_colour_%d"), pent->layer(), pent->colorindex() );
// select (and create if required) the correct layer
idlayer = createlayer( pdb, strlayer, pent->colorindex() );
pent->setlayer( idlayer );
}
}
savefile( pdb, m_strfolder );
}
regards
chudomir
hi
i have tried removing the kforwrite bit, but that makes no difference. didn't quite follow youon "closing iteration".
andrew
quote:
originally posted by chudo
can you try just "closing" the iteration to ensure all the objects are closed:
this was the answer. by putting a pent.release() inside the for loop at the bottom. if i had declared the oddbentityptr inside the for loop, it would have handled it itself.
the file now processes and saves and loads ok in dwg mode.
thanks for your help.
andrew
i meant to place "{" before opening the model space and "}" before savefile. and yes, to declare the pblock as a local variable in the thus formed code block:
code:
....
{
oddbblocktablerecordptr pblock = idmodelspace.safeopenobject();
for( pentiter = pblock->newiterator(), uentity = 0;...
...
}
savefile(...)
...
this will force the smart pointer pblock to close the opened model space at the closing "}"
regards
chudo
sorry, i noticed the braces but didn't understand their purpose. however, the entity pointer would have still needed closing.
yes, i've forgot the entity pointer - so the idea is just to declare the smart pointer as "local" as possible in order to have the objects closed quickly.
hope this should repair your problem.
regards
chudomir