Re: st: Capturing the directory of a .do file

[Phil Schumm <pschumm@uchicago.edu>]

On Sep 7, 2007, at 4:39 AM, Sergiy Radyakin wrote:
> 2Austin: Why not just use relative paths, and have folks run the do- 
> file in the relevant starting directory? Because when a .do file is  
> run, current directory is not necessarily the directory where  
> the .do file is located.

True, but you can check for that, at least partially.  In particular,  
you can use -confirm file- to verify that the various files on which  
your do-file depend(s) are in the expected locations relative to the  
current working directory.

Given the original question, I think Austin's suggestion was spot on  
(including the bit about the path separators, which should always be  
specified using -c(dirsep)- to make sure that your code works cross- 
platform).  IOW, write all your code using paths relative to the  
project's root directory, and then specify in the README that users  
should -cd- to the project directory before running it.  As long as  
you distribute (or move) the entire project directory, you'll be fine.

Note that, for those with Unix experience, this is quite natural.  If  
I download a program and want to compile it, I would typically need  
to -cd- to the root of the distribution before compiling, and I would  
expect the Makefile to refer to all files distributed with the source  
by their relative paths (so that I am free to compile wherever I want  
on my system).

Of course, if you want users to be able to double-click or use menu's  
to execute do-files, then things become a bit more complicated, and  
the ability for a do-file to "know it's location" becomes important.   
However, we're talking about Stata here -- what do you need the GUI  
for? ;)


-- Phil

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/