Show ReadMe.txt syntax highlighted
========================================================================
STATIC LIBRARY : vsAUTO Project Overview
========================================================================
Changes made to original AUTO 2000 source:
1) added berkeley getopt.c (found somewhere on the net, modified to work
for windows).
2) updated timing function to work for both windows and *nix.
3) disabled PTHREADS and MPI for WIN32 builds.
4) fixed fseek/ftell problems for windows. (windows has that stupid,
automatic +2 at a \n character that hoses ftell. (for now, i only did
this for the fp3 guy... i anticipate replacing the file based
mode of operation, so i don't want to waste time right now).
5) added namespace encapsulation for C++ builds and modified a few
demos to deal properly with the void* vs. doublereal* in pvls(...)
6) added auto_api.{h,cpp}, which contain everyting for the API.
this is part of the new library interface. ContinuationConstants is a
object that knows about the constants file; it can write itself to the
constants file (for preparation to run the AUTO_main proc) and can read
itself as well. SolutionSet is the .8 file IO, PlotSet is the .7 IO etc.
7) made fp2, fp3, fp7, fp8, fp9, fp12 all open from fort_name[13][512] so
that we can set the names of the files for different runs.
changes in v0.7+
8) changed all static locals in eispack (and some other places) to have
a reentrant library. the user may like to do several continuations at
once... still need to thoroughly check this.
9) 0.7.4: changed NPARX, NPARX2 etc to a dynamically settable feature. all memory
is now allocated for vars needing this.
10) 0.7.4: change par[9], par[10],... , par[14] to
par[sysoff], par[sysoff+1],... par[sysoff+5] so that we can specify pdim,
use all those for the user pars, and then add any systems pars above that.
i still need to check the par[(k<<1)+18] blah blah to move them above these,
but we're close. i also need to add a new AUTO_main for the library version
to set num_model_pars, num_total_pars, NSYSPAR etc and to decide whether or not we want to
keep a C version of the code intact, or should I just stick with C++ (in the
first case, I should go back and change the new/delete syntax back to a
malloc/free version.
11) 1.9.3: several small changes have happened int the last many months.
one of note is the nit/nit1 error in swprc (int autlib1.cpp). this was
and inifite loop.
12) disabled fort.9 for Release mode now since it's cause an ugly disk space
issue for the bifurcation matching (which runs many thousands of
BDs).
See more files for this project here