Access changing lettercase - redux
This has probably been beaten to death, but I'm fuming, so here goes. I want Microsoft to DO SOMETHING to help. Access has the highly annoying habit of changing lettercase on identifiers, seemingly at random. This is a real pain when using add-ins like MSAccessVCS, which makes it possible to use git for source code control. Here' what I've deduced is happening: There exists a dictionary/symbol-table containing every identifier used in a project. The symbol table contains ALL identifiers, including those defined by MSAccess, referenced libraries, and your code. Searching that table is case-insensitive, but it stores the canonical version of the identifier with case preserved. Normally, when you type an identifier that already exists in the table, the VBA editor "corrects" the lettercase of your entry to match the table's entry. The extremely annoying part is that sometimes that last step works in reverse. You type an existing identifier, but with a different lettercase. Instead of correcting what you typed to match the table, it instead updates the table with the version you typed, and decides that's the way future and existing identifiers should be spelled. Here's a recent example. I tried to create a class module constructor (Class_Initialize) but typed it "class_initialize" implicitly expecting that the VBA editor would "fix" it. Instead, it updated the symbol table and set the canonical lettercase to the all-lowercase version... and then proceeded to change the lettercase on the Class_Initialize() method in all my class modules. It has now decided the canonical lettercase is "class_initialize" and that's what it "corrects" to from now on. Of course, when I did the next MSAccessVCS export, git showed a bunch of unwanted (but cosmetic) changes. This pollutes the changeset and makes source control more difficult. THIS. IS. A. BUG. My workaround is, when starting a checkin, I go through the list of changes and group all such changes into one commit called "VBA Artifacts". But come on, this shouldn't be necessary. At least let us know WHY this happens, and give us some control of the process. A way to specify/correct the canonical casing would help.