jptrainor
21 Aug 2012, 2:02 AM
Hi.... this may be a question, or it may be bug report, or maybe a feature request. I'm not sure how to classify it.
I have my Architect 2.1 project under revision control using subversion. I committed the whole thing (root files, entire app directory, and entire metadata directory).
When I rename a class (using the "userClassName" property in Architect) and save a project I find that Architect entirely removes and recreates the directory in which the file resided. This behavior is the root of my question.
For example, start with one store named X that is under revision control. My "store" directory will contain the X.js file and my .svn directory (.svn is the subversion directory). Rename X to Y and save the project in Architect. Now I find the .svn is gone (not expected), X is gone (expected) and there is a Y.js file (expecteD).
This makes for a very awkward workflow because in order to commit my changes I have to fix the problem with the store directory. That means running an "svn update" which brings back what was there, getting the Architect changes back in place, manually removing the old X.js file, and finally committing. So the work flow is: rename a class, save, flip over to subversion to sort out the mess, get back to work.
Is this a bug of some kind, or am I simply incorrect, to begin with, by putting everything in revision control?
The problem would not exist if the entire directory was not removed when the class name is changed. It's common, for example with Eclipse, to rename a class. When you do that you see a new file appear that is not under revision control (expected) and you see the old fine disappear (expected) and it's all very predictable. Architect... not so much.
I have my Architect 2.1 project under revision control using subversion. I committed the whole thing (root files, entire app directory, and entire metadata directory).
When I rename a class (using the "userClassName" property in Architect) and save a project I find that Architect entirely removes and recreates the directory in which the file resided. This behavior is the root of my question.
For example, start with one store named X that is under revision control. My "store" directory will contain the X.js file and my .svn directory (.svn is the subversion directory). Rename X to Y and save the project in Architect. Now I find the .svn is gone (not expected), X is gone (expected) and there is a Y.js file (expecteD).
This makes for a very awkward workflow because in order to commit my changes I have to fix the problem with the store directory. That means running an "svn update" which brings back what was there, getting the Architect changes back in place, manually removing the old X.js file, and finally committing. So the work flow is: rename a class, save, flip over to subversion to sort out the mess, get back to work.
Is this a bug of some kind, or am I simply incorrect, to begin with, by putting everything in revision control?
The problem would not exist if the entire directory was not removed when the class name is changed. It's common, for example with Eclipse, to rename a class. When you do that you see a new file appear that is not under revision control (expected) and you see the old fine disappear (expected) and it's all very predictable. Architect... not so much.