Chuck H.

I don't want to create a fork of GNUCOBOL, so I have removed the links to download any of the multiple window builds and also the RM/COBOL WCB builds. I am going to step away from GNUCOBOL for a while as I'm a bit frustrated now. It will be better to revisit this after a while as hopefully my level of frustration will diminish with some time for reflection. Chuck Haatvedt

This is meant to start a discussion of GNUCOBOL’s development goals. Should the development focus be limited to providing support for other COBOL compiler vendors OR should we also allow for the inclusion of callable GNUCOBOL exclusive functions to extend the usefulness of GNUCOBOL. In the case of the CBL_GC_WINDOW ABI, the infrastructure can be used to clone the windowing implementation of other compiler vendors. NOTE... the following are my opinions only and are provided only to start a discussion...

As it pertains to this topic, also see my post about GNUCOBOL DEVELOPMENT GOALS. The PopUp-Window function would only be available if compiled with the “-std=rm” compiler option. The ACUCOBOL-GT extensions DISPLAY MESSAGE BOX and DISPLAY WINDOW/ MODIFY WINDOW / DESTROY WINDOW syntax would only be available if compiled with the “-std=acu” compiler option. The PopUp-Windows using the Window-Control-Block (the WCB extension originates from RM/COBOL) is a test version built upon the CBL_GC_WINDOW ABI...

In discussion with Simon, it was decided to implement MULTIPLE WINDOW support in the SVN repository with the CBL_GC_WINDOW API marked as EXPERIMENTAL so that it would not be available unless the configure step included the following --enable-experimental (GnuCOBOL) Enable experimental code (Developers only!) however this would mean that all of the experimental code would be included, not just the MULTIPLE WINDOW API. This means to to create a MULTIPLE WINDOWS compiler for testing the compiler would...

Mickey, I analyzed your programs the flat2index1001.cbl created an index file with a fixed length of 121 bytes. the cust01.cbl program defined the fixed length index file with a record length of 127 bytes... from fileio.c ret = indexed_start_internal (f, COB_EQ, key, bdb_opts, test_lock); if (ret != COB_STATUS_00_SUCCESS) { return ret; } f->record->size = p->data.size; if (f->record->size > f->record_max) { f->record->size = f->record_max; ret = COB_STATUS_43_READ_NOT_DONE; } else { ret = COB_STATUS_00_SUCCESS;...

Mickey, could you post the output of a "cobcrun --info" so that we can see which indexed file access library that you are using in you testing ? Chuck Haatvedt

As per Simon's request I will be modifying the logic to use a single CBL_GC_WINDOW function to provide all of the functionality currently available with the various CBL_GC_WINDOW_XXXX functions. This will require an additional variable in the existing communication parameter area for both of the copy members. I will provide more information about the communication area parameters when I start the work on this. These changes will delay the implementation of these feature as it will require some time...

fixed [bugs:#999] ACCEPT with TIMEOUT issue when looping thru the verb