Neil Hodgson wrote:
Its mostly a better idea to post changes to the mailing list or bug
tracker. That way the proposed changes are seen by those more
interested in them (I don't use sessions or Haskell) who can comment
before the code gets committed.
Ok. The list it is.
you can load a session using "-loadsession:<name>" in the command
line, but then when you exit SciTE, the session is autosaved in the
home directory, instead of updating the loaded session file. This
makes it awkward to keep multiple session files for different
directories.
The fix is simple: in SciTEBase::SaveIfUnsureAll(), SciTEIO.cxx line
710, replace the SaveSession("") with
SaveSession(props.Get("SessionPath").c_str()).
Due to settings lifetime issues, this may not be reliable. The
properties system is for retrieving settings from settings files and
communicating values to scripts. Values that are set and used by SciTE
code should be instance variables instead.
I'm not sure I understand. Either:
- Your comment is about code organization and conventions you don't want
broken. In that case, I don't agree but you're the boss.
- You're afraid the fix won't work because the properties will be
recomputed at some point, and this particular property will be
invalidated (it is only initialized when a session is loaded or saved).
In that case, it will also disappear for any scripts so IMHO it is a bug.
As far as I can see, this fix works just as advertised. I've been
expanding it a little (added a SessionDirectory property, the path
without the filename, which is also useful for session-aware scripts).
JCAB
_______________________________________________
Scintilla-interest mailing list
[email protected]
http://mailman.lyra.org/mailman/listinfo/scintilla-interest
