Re: breaktable checks in dbLexRoutines.c in 3.14.9

[Andrew Johnson <anj@aps.anl.gov>]

Hi Steve,

Steven Hartman wrote:
> Yes, such an approach could be acceptable. However, would it be possible
> to also have a way to over-ride the default behavior at build time? The
> breakpoint table dbd currently gets rejected during the make process by
> dbExpand.

I did mention that this would be the case in an earlier message, when I 
recommended that you move your breakpoint tables into a separate DBD 
file which you load independently at runtime.  If you want a different 
default value for the dbBptNotMonotonic you can always alter your copy 
of base however you like, but I don't want the official version of base 
to skip this check unless explicitly told to do so by someone like 
yourself who understands the issues, and unfortunately there's no easy 
way to make that happen conditionally when running dbExpand.

Breakpoint tables actually hold a different kind of content than any of 
the other DBD entries (which are all metadata), and as a result I could 
very easily see them being moved into a different kind of file at some 
point in the future.  If we do that we'd definitely make them reloadable 
at runtime, which may work right now but is not thread-safe IIRC.  As a 
result splitting the tables into their own DBD file(s) now might just 
save you from having to do it later.

Actually it should be possible to include breakpoint table data in a .db 
file today â VDCT won't handle them properly, but this might be a better 
place for them conceptually and the msi tool that does template 
substitutions at build-time won't reject a non-monotonic breaktable like 
dbExpand does.

HTH,

- Andrew
--
When a distinguished but elderly scientist states that something is
possible, he is almost certainly right.  When he states that something
is impossible, he is very probably wrong.  -- Arthur C. Clarke