Re: [Doxygen-users] generic variable for output directory

[Martin J. <mar...@is...>]

On 02.07.2008 08:56, Peter Schöll wrote:
> as far as doxygen supports to evaluate environment variables within its config file, I'd recommend to write
>     OUTPUT_DIRECTORY="$(ENV1)/$(ENV2)"
> in Doxyfile itself.
>   
I agree that this is the best, most straight-forward way of changing 
Doxygen parameters if you can use the same environment variable names 
for passing the information to Doxygen in every run.
>> My objective is to pass path info (which may be changed during the course
>> of
>> project) through command line.
>>     
Does this mean that you want to set a different OUTPUT_DIRECTORY for 
every Doxygen run? If yes, then you should set an environment variable 
to the desired directory and assign this environment variable to 
OUTPUT_DIRECTORY in the Doxyfile as Peter Schöll suggested.
>> I tired to pass env variable using the process mentioned in the #17 but it
>> gave an error as illegal variable.
>>     
Where did the error message come from? The shell? Doxygen?
>> ( cat Doxyfile ; echo "OUTPUT_DIRECTORY = $(ENV1)/$ENV2/)" ) | doxygen
>>     

Did you forget the dash argument to Doxygen?

Besides, this line will not produce the desired effect on modern Unix 
shells because $(ENV1) is a modern variant of `ENV1`, i.e. the shell 
will try to execute a command called ENV1 and replace $(ENV1) with its 
output. This is probably the reason for the error you experienced.

As said above you should simply use ENV1 and ENV2 directly in the 
Doxyfile. If for any reason you have to use different environment 
variable names for different Doyxgen runs you could either use

(cat Doxyfile; echo "OUTPUT_DIRECTORY=\"$ENV1/$ENV2\"") | doxygen -

or

(cat Doxyfile; echo 'OUTPUT_DIRECTORY="$(ENV1)/$(ENV2)"') | doxygen -

In the first case, the shell itself substitutes the values of ENV1 and 
ENV2, whereas in the second case Doxygen does this.

Regards
Martin Jerabek