RE: Re: Variable Length Records FSwrite problem
"Rocky" <[email protected]>
| Newsgroups | gmane.comp.emulators.hercules390.vm |
|---|---|
| Message-ID | <[email protected]> |
YUP AND YUP Actually I think I've got it resolved. But thanks for the offer - if that turns out not to be the case I will take you up. As usual - after I tried everything I could think of for two days - thats when I put the memo out, 20 minutes later I found a bug that mey be the reason. Anyway you are going to get all the code sometime during this week It is really working and I have being writing the editor in the editor. takes a bit of getting used to but so did X-edit and ISPF edit (and they were not written in a month,) Roc _____ From: [email protected] [mailto:[email protected]] On Behalf Of Robert O'Hara Sent: Sunday, August 15, 2010 01:52 To: [email protected] Subject: [H390-VM] Re: Variable Length Records FSwrite problem Roc, I assume you are trying to write only one record at a time (NOREC=1), and that you are writing sequential records (RECNO=0). When writing V format files both of these must be true or unpredictable results occur. Can you post a snippet of your code where you fill in the FSCB and then do the FSWRITE? Bob --- In [email protected] <mailto:H390-VM%40yahoogroups.com> , "Rocky" <rocsystems@...> wrote: > > I am doing fine tuning of some of the features for the full screen Editor > Comng along very well and I am no longer working on it full time > Probaly will have the final Beta shortly. Actually using it to write the application and its not bad. > > PROBLEM: > > FSWRITE for variable length records > > I am writing the variable length records with the BSIZE parameter on FSWRITE > > Of course I am reusng the same FSCB's. > > Every time I read a new file I reinitialize it completely. > For some reason the system is igonoring that.Perhaps by design. > > I do the reads according to the FLIST parameter which shows the buffer size. > > But if I read or write a variable file that has buffer size of say 200. > > I can't write smaller files any more. Bigger ones it writes > I can read a variable length file and add 20 bytes to the longest record and that works fine. See it the flist with another 20 bytes. > > I close the FSCB Open it again for another fie that is smaller > and do the FS write with the proper BSIZES for the second file . > I put a trace on it and can see that I am using the proper sizes > > But for some reason when I do the flist again it show that the new file has the buffer size of 200 from the previous file. > I know I am writing the proper because if I rread the file the short records are short. But LISTF-FLIST show the max buffer size of the previous file > > Fixed files work fine - no problems > Even in betweeen variable files. > > But once I write a large variable file I can't write smaller ones after that. Well I can get them to show up propery in LISTF > > So the nex time I read tha file I would be using a bigger buffer > but the actual data is shorter. > > I did an erase of the file before the write using FSERASE and then wrote it. > > I erased the file externally but it seems that I cant get rid of the the info in the handle unti I close the application. > > I even reaccessed the "A" disk before the write > > If I close the application and start from scratch - no problem > > but reusing or reopening the an FSCB does not seem to scratch everything out. > > Evertyhing sems to work except the BSIZE which is probably stuck somewhere in memory > > Anybody have any ideas? > > Roc >