* On 22 May 2014, jpac...@redhat.com wrote: > I'm not sure if it's a good idea as I don't know mutt that thoroughly to > be able to think of all cases where it could cause harm. And so I'm > asking you if you could provide me with some examples of unexpected > behavior due to DEBUG being allowed (e.g. security, performance, UI > friendliness, macros etc.) without explicitly enabling the debug mode in > runtime using -d N.
There is a marginal performance penalty to enabling debug, but I don't expect it to be significant. A profiler would tell, since dprint() is macroed out under -UDEBUG. Generally I think this would be useful, but if you're evaluating this from the perspective of a packager -- I don't know who packages mutt for distros, currently -- why not just include two builds? /usr/bin/mutt for regular use, /usr/bin/mutt.debug -d2 for debugging runs. On my system a stripped mutt exec with most options enabled is under 900K. -- David Champion • d...@bikeshed.us
