Mark J. Nelson wrote: > Jordan Brown wrote: >> I was just surprised when an "hg stat" didn't tell me about the files >> I'd created but hadn't "hg add"ed, so I looked around and found that >> .hgignore is set to ignore everything. >> >> Is that the long-term plan? > > Yes and no. It would be great to get a detailed, appropriate .hgignore > file for ON. But as it currently stands, such a file overwhelms the > Mercurial implementation of hgignore. Fixing either the ON build (to > make the file more sane) or the hg ignore internals (to deal with an > insane file) are not even on the radar. > > So "no," we're not explicitly planning to stay in this state forever. > But "yes," realistically, I don't see this ever happening. ENORESOURCES.
All understood. What I wonder is whether it's possible to approach the problem incrementally, much as has been done for lint, with particular source branches handled "correctly" and the rest ignored. One would then hope that over time the build structure could be reorganized to better separate sources from build artifacts, the .hgignore file fleshed out, or both. A check for changes in the "hg stat" noise could be added to nightly builds, again analogous to lint. I'm not sure this is actually practical, but I'll think about it a bit and maybe experiment.
