|
From: Joachim E. <joa...@gm...> - 2005-03-22 18:06:57
|
On Monday 21 March 2005 22:41, Bob Eby wrote: > I just downloaded and started using KDiff3 a couple days ago. It seems > to work great for letting me diff large volumes of CVS files, avoiding > differences in the comments, and differences in "lint" that I don't > really care about. However, using cygwin's version of sed and the line > by line preprocessor is rather slow and invasive because 2 new command > windows are opened for each pair of diff'ed files. I'm just wondering > if there's a better/faster way to do regular expression matching using > the pre-processor on Windows (2000 and 2003 server in this case). > Perhaps a native version of sed or a different tool entirely? (It's > taking me 1.5 hours to process 2500 files... pretty dismal) For the moment you might try the sed that comes with the package from unxutils.sourceforge.net. This doesn't depend on cygwin and might be faster. But I admit that for so huge comparisons it would be best to do a built-in regexp-analysis, because this would avoid the creation of temporary files alltogether. Perhaps you could create a RAM-disk for your temporary files and sed.exe. > I'm also wishing I could do a directory merge that lets me have file by > file control without forcing me to go in alphabetical order, but I'm > still learning what is and isn't possible here, so I'll ask more about > that later. Try menu "Directory"->"Run Operation for Current Item" (F6). Cheers, Joachim |