I'm comparing two implementations of "redo":

First, jdebp/redo 1.4 [1]
Second, apenwarr/redo 0.41 [2]

[1] http://jdebp.eu/Softwares/redo/
[2] https://github.com/apenwarr/redo

Given a project "toplevel" with a subdirectory "subdir" where one's current dir 
is "toplevel":

In jdebp implementation, running "redo subdir/target" invokes target.do with 
working dir "toplevel"
In jdebp implementation, running "cd subdir && redo target" creates a new .redo 
database in subdir

In apenwarr implementation, running "redo subdir/target" invokes target.do with 
working dir "toplevel/subdir"
In apenwarr implementation, running "cd subdir && redo target" reuses 
"toplevel/.redo" (assuming it already exists)

The apenwarr implementation seems more useful for constructing a build system
for a project with many nested subprojects, but I may be missing something (I
don't have much experience with any redo implementation). Anyone have comments
on this, or example nested subprojects using "redo"?

-- 
Patrick

Reply via email to