Hi, On Mittwoch, 9. Juli 2008, Martin Preuss wrote: > On Mittwoch, 9. Juli 2008, Jerry Jaskierny wrote: [...] > Since GnuCash is directly looking for an AccountInfo object for the > expected account number (in this example "xxxxx1234") it doesn't succeed. [...]
Being finally able to compile GnuCash again (thanks for the hints!!) I had a look at that particular code and I see a general problem. The code only looks for a particular account in the data returned (namely the one it requested statements for). However, this doesn't always work as one would expect. Some banking protocols do not allow to request statements for a specific account but rather for all accounts a given user is authorized for (e.g. to some extend EBICS, but certainly YellowNet). That was the reason to begin with for AqBanking not to return the requested data with the job structure but rather let the protocol backend gather all received data in those ImExporterContexts. This isn't just a problem of not reading superflous data: Some protocols (especially YellowNet, but EBICS as well) allow the server to physically remove data reported to the client, so data you loose in gnc-ab-gettrans.c will most certainly be lost for good. This most likely isn't what the caller intented when he requested statements... So what I propose is this: You must already have a generic import/export interface which is able to import data from an AB_IMEXPORTER_CONTEXT, since you have an importer for SWIFT MT94x, so why not use this interface to import the data returned? For HBCI it makes no difference, but the other protocols can then also be safely used with GnuCash. Regards Martin -- "Things are only impossible until they're not" Martin Preuss - http://www.aquamaniac.de/ AqBanking - http://www.aqbanking.de/ LibChipcard - http://www.libchipcard.de/ _______________________________________________ gnucash-devel mailing list [email protected] https://lists.gnucash.org/mailman/listinfo/gnucash-devel
