Re: Xy3 memory

[Robert Holmgren holmgren@xxxxxxxx]

** Reply to note from xywrite@xxxxxxxx Thu, 20 May 1999 17:35:20 -0300

Myron:

VIEWS.PM struck a vague bell, but only when you mentioned XYWPMKIT did I
realize that it was my program (frame "twoscr" in XYWWWEB.U2, much more
bullet-proof). Here's my suggestion, for a simple workaround. Before you
start using VIEWS on a given document, simply go to EOF, then to TOF, then
back to your current cursor position. Then run VIEWS. It should be more
reliable. The issue here is not "releasing" memory, but simply reading the
entire document into XyWrite, i.e. loading it all into either memory or into
the swap file. That, anyway, would work in Xy4 and probably works in Xy3
too. If it doesn't work, then I'll look into the code and insert a
simple check procedure to make sure that JuMPs are accurate. Gosh, VIEWS.PM
must date from the 80s. Time flies. Where did you get XYWPMKIT.ZIP? From
CCAT FTP? I don't have a copy, at least not with me (traveling)...




> The XPL program "views.pm" (which is part of XYWPMKIT.ZIP on the web
> archive) allows to you switch between two (or up to 9) different "views" of
> the *same copy* of the same file, so any changes in one view are
> automatically reflected in the other views (since they are "views" of the
> same working copy). "Views" are not XyWrite windows, as two or more "views"
> occupies a single XyWrite window. I assume it does this by jumping around
> in the file, "remembering" cursor locations in various SGT's. Once I have
> done an ordinary SEarch that covers more than about 64K, Views.pm won't
> "remember" (or perhaps find) locations that are 64+K apart in the file. It
> will still jump between locations closer than that (50K, for example), and
> until I have done an ordinary SEarch over a distance of 64K+, it will work
> fine over distances like 100K. It is as though using the SEarch function
> permanently claims memory that Views.pm needs in order to jump over large
> distances. I haven't found a way to "release" the memory so Views.pm can
> use it --- at first I thought shorter SEarches sort of "reset" the SEarch
> memory to a smaller size, but I can't get it to work predictably.



---------
Robert Holmgren
holmgren@xxxxxxxx
---------