On Monday 23 May 2005 18:12, Claes Wikstrom wrote:
> Ahh, never thought of that. It would make sence to do that .. maybe.
> Then again, it may not always be that easy. The could very well be running
> processes that would then .. possibly get killed.
> Frankly, I can't make up my mind what is the right thing (TM) to do
> here. Does the list have an opinion?
My current development cycle is:
1. Make change to erl
3. Copy across beam files
4. Restart yaws
Hupping yaws is quicker than a full restart, so it would shave a little time
off my development cycle. It is documented that hupping will "flush internal
content caches", I haven't investigated this, but could it mean yaws could
find updated .yaws files but not updated .beam files leading to .yaws files
trying to import code from a .beam file older than expected?
The workaround is quite easy and the fix was not obvious to me, so this is not
a big issue.
> Besides, did you get get the restart problem from earlier posts
> sorted out. It looked buggy.
I haven't seen it recently, which suggests I had something misconfigured. I
would like to know what though, so if I see it again I will try and debug it.
BTW, for any Debian users out there, there is now a yaws package for unstable: