Long ago, people who worked with the major computer operating system in the universe ("IBM" mainframe) would maintain a library of reference manuals (yes, paper - and a lot of paper). A significant quantity of these manuals would be incredibly large glossaries of system termination codes. Each code would have a special significance, such as attempting to enter a date containing an alphabetic character. Seeing a given code for a specific database entry, a database engineer would look up the code in the reference manual, then patch the record in error accordingly, and the problem would be solved.
The bX- codes have no similar significance. They are simply unique codes, which identify each individual point of abnormal termination. There's no secret glossary assigning the cause for any single code, just a pointer to the individual termination point in the Blogger codebase. When enough bloggers report a given code, a Blogger employee simply examines the termination point, and using the diagnostic information hopefully provided by the bloggers problem reports, makes a diagnosis of the problem cause. Rarely, the code will be added to a small database (which may or may not be active) which lists uniquely significant codes.
While checking to see why I was getting these "bx" codes myself, I found this wonderful tutorial on how things are done differently these days (and not necessarily for the better). I couldn't have said it better, so copied!
No comments:
Post a Comment