Ian- Thanks for the pointers (and the mighty quick response!). For my own projects I'd just post the source, unfortunately I'm not at liberty to post the code from my employer without reducing or significantly obfuscating it. Delta looks like a very neat tool, I'll give that a whirl and see if it'll pare the file down for me. Sorry for all the hassle due to source issues, but I have one more question -- in case Delta still leaves a large file, is there a good tool to randomize the names of the variables and types in a C++ file? Thanks again for all your help. --- Ian Lance Taylor <iant@xxxxxxxxxx> wrote: > Aliesha Finkel <agfinkel@xxxxxxxxx> writes: > > > I'm a developer on a large C++ application that's > > using gcc as our compiler. We have one file that > > takes *much* longer to compile than all the > others, > > thus making distributed compiles a lot less > efficient. > > > > My question is, is there any easy (automated?) way > to > > isolate the problem so that I could feed back a > test > > case to gcc developers? Or am I stuck just > hacking > > out pieces and iterating. It's a very large file > with > > lots of template definitions in include files, so > > trial and error seems like an impossible task. > > A smaller test case is nice but not essential. If > you are permitted > to upload the source code to a public site, then > don't spend a lot of > time reducing it. Just follow the instructions at > http://gcc.gnu.org/bugzilla/ and attach the > preprocessed file. > > If you can't do that, so you need to try to pin this > down some other > way, then take a look at http://delta.tigris.org/. > You could probably > make the file quite a bit smaller by using a delta > script which timed > the compile. > > Ian > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
