As a followup to this (I seem to be answering my own questions??), I
have tweaked the process a bit, which seems to have eliminated the
issue. I have made the [reactor]/project directory writable by the
coldfusion server user. The [reactordata] directory is writable by my
ftp user (and not coldfusion). So I can upload the reactordata files as
I modify them, and Reactor can update the files in [reactor]/project as
needed. Everytime I make a table change though, I need to recompile the
[reactordata] files locally (I guess maybe just the dictionary files?),
merge in any changes (again, only the dictionary?) then republish them
to my testing server. Still a pain, but it seemed to eliminate Reactor
wanting to regenerate my [reactordata] files.
So I guess that brings me to another point (maybe I should've started a
new thread): Is there any way for a set of main dictionary files to be
generated in the [reactor] project directory and then allow them to be
overridden by a local copy--just like all the cfcs work? So I could put
my custom dictionary items and any dictionary items I want to change in
a [reactordata] dictionary xml file, but the default set of dictionary
items, including any that change as the table structure changes, would
be in the "untouchable" [reactor]/project directory. I'm not sure how
that would happen (merging of xml files), but at the moment it seems
like a good thing because right now the generation/overriding of the
dictionary files don't follow the same rules as all the other generated
files.
Thanks,
Mark
Mark Mazelin wrote:
I have Reactor turned to production mode in the config xml file. Yet
when I make a change to one of the generated Record cfc files, Reactor
wants to attempt to regenerate the Record file in the Reactor project
directory. Is that necessary? My current process is this:
1. Use Jared's cool compiler program (thanks Jared!) to generate all
the reactor objects on my localhost (windows)
2. Upload those files to my testing server (linux), which has *limited
write rights *for security purposes
3. Modify and uploading the generated files to my testing server as needed
Because of the complexity of my setup, I don't have the luxury of
running everything on my Windows machine and just uploading everything
when "no more changes" will be made, so I have to do something like
this. The problem comes in because ColdFusion running on my linux box
runs as a different user than the user I upload files as. So if I
generate the reactor files on the linux box, then I have to download
them, edit them, then change permissions on the linux box to re-upload
them. It becomes a real pain.
So, back to my question for this e-mail: Why is Reactor doing this,
and is it necessary? Am I making this too hard, e.g. is there a better
way?
Thanks,
Mark
-- Reactor for ColdFusion Mailing List -- [email protected] --
Archives at http://www.mail-archive.com/reactor%40doughughes.net/
-- Reactor for ColdFusion Mailing List -- [email protected]
-- Archives at http://www.mail-archive.com/reactor%40doughughes.net/
