On Thu, Jan 31, 2008 at 10:26:25PM -0500, Max Battcher wrote: > Nathan Gray wrote: >> So in my scenario, I have a central development repo that generates HTML >> documentation from source code that is pushed to it. I would like to >> only generate documentation for files that are affected by the push, >> rather than all files in the repo. Somehow my documentation generating >> program needs to be told which files to process via the posthook. > > In one of my repos I've just been using something like: > > find -mtime -1 | my_posthook > > I realize it's not a perfect solution, but if you are desperate for any > solution its a working stopgap.
I fortunately am not desperate for a solution, but I am finding that I am wanting this functionality in more repos. The stopgap solution that I currently have is described in http://bugs.darcs.net/issue524 and has been okay. As I prepare to implement this in more repos, I just wanted to check in and see if anyone had time to put this into the posthook code. > I know that several proposals have been made for what might be done to pass > information to posthooks, but would a simple "pipe-mode" for posthook be > able to make use of existing code, and thus maybe be something that could > be finished sooner rather than later? > > That is give me an option to get the equivalent of ``darcs changes -s > --xml-output`` for just the newly applied patches piped to my posthook... > I would assume you could just use the printing code already used by darcs > changes and just redirect standard output to a pipe to a posthook. I have been generating a list of file names and a list of hash ids, but I would be okay with an XML summary. I would rather have the summary in an environment variable, rather than from a pipe, in case there are multiple programs I want to run in the posthook, all needing this information. -kolibrie _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
