Molpro version 2002.6 release notes from patch history of preceding version

5 Nov 2002 newmyrinetformat4.2002.5
When fixing bug 390 I noticed a similiar problem: MPIGM_FORMAT
is ill-defined when switching from a sequential to a parallel
environment. This glitch is cleaned up together with another
myrinet related omission in ./configure.

5 Nov 2002 prop.2002.5
Fixes integer overflow in formgm
Replaces real*4 constants by double precision

5 Nov 2002 forces_perm2.2002.5
Removes Molcas permutation in various places and do this
just before calling alaska (and in t2tran_drv). Necessary
preparation for ecp forces.

5 Nov 2002 dec_build3.2002.5
Recent versions of gcc (e.g., 3.2) will not silently ignore -i8 flag.
This isn't needed anyway.

5 Nov 2002 ecpmp2grad.2002.5
Evaluates mp2 gradients with ecp using the correct one-particle density matrix

6 Nov 2002 fix_history.2002.5
Patch history in src/forces/krafta.f corrupted

7 Nov 2002 aosort.2002.5
Compute mxbuck dynamically

8 Nov 2002 freqsym.2002.5
Disable use of symmetry unique displacements if wavefunction is
symmetry broken

8 Nov 2002 gio1.2002.5
Add GIO library calls to iom and iow.

11 Nov 2002 gio2.2002.5
Patch gio1 broke a link in ./develop that is necessary for
building the Programmers' Manual. This patch also fixes
two LaTeX bugs in iom.F.

12 Nov 2002 gio3.2002.5
Some compilers seem to have problems with an empty globio.F
in cases where gio is not set. I rearranged the molpro
precompiler switches: Now in non gio cases one will get
a dummy function instead of an empty source file.

12 Nov 2002 lib_dup.2002.5
Bug(s) fixed: 394
Prevent duplicate entries being written via LIB command.

13 Nov 2002 bucky3ext_5.2002.5
Implements bucketsort for D3ext transformations in LCCSD/LMP@GRAD and alleviates
the I/O problem hampering these algorithms so far...

13 Nov 2002 xtraopt1.2002.5
This patch adds another option to molpro.c: -c <option>.
<option> specified this way will be used as a command line
option for the parallel job wrapper (e.g. mpirun). Multiple
-c <option> can be specified; up to NXTRAOPTS=5.

14 Nov 2002 table_plot2.2002.5
Bug(s) fixed: 399
Correct spurious omission of last point in spline fit in table;plot.
Also make more robust the algorithm for detecting duplicate data points.

14 Nov 2002 ahlrichs_cbasen.2002.5
Puts all current Ahlrichs cbasen into LIBMOL. This is done with
program read_ahlrichs_basen.f, saved in utilities
prel basis sets are excluded for the moment; I am not sure what they are.

15 Nov 2002 zmat_auto.2002.5
Initialize auto=.false. This elimninates error when using ZSYMEL variables

15 Nov 2002 ifc_libs.2002.5
ifc requires some shared libraries which are in non-standard
locations to be linked at runtime. If LD_LIBRARY_PATH is not set
then this causes an error. Added arguments to the linker which
explicitly adds these directories to the runtime search path.

16 Nov 2002 rs2grad.2002.5
Rs2 gradients with core and state-avered reference functions.

18 Nov 2002 dk_bases.2002.5
Bug(s) fixed: 392
Added cc-pVXZ-DK basis sets (X = D, T, Q, 5). If DKROLL = 1, look for
foo-DK basis when basis foo is specified.

19 Nov 2002 bucky4ext_2.2002.5
Implements bucketsort for D4EXT transformations in LCCSD and alleviates
the I/O problem hampering these algorithms so far...

19 Nov 2002 dualfock.2002.5
Better implementation of dual basis sets. Only increment of fock
operator computed. New orbitals and fock matrix stored in dump record.

20 Nov 2002 ifc_lib2.2002.5
fixing error in linux.conf brought about by patch ifc_libs

21 Nov 2002 muopt.2002.5
Gives an error message if wrong orbital record is specified in mcscf
geometry optimizations

22 Nov 2002 necsx_tr.2002.5
Drive NEC flavour of /usr/bin/tr in bin/molver

22 Nov 2002 cidoc.2002.5
Adds a reference to the multi save section in section select

22 Nov 2002 ecpgradtest.2002.5
Provides new testjobs for ecp gradients

25 Nov 2002 matrop_pot.2002.5
Bug(s) fixed: 405
Fixed uninitialized variable leading to errors.

26 Nov 2002 ciinit_formats.2002.5
Modifies formats in ciinit.f to allow print of more than 999 external
orbitals and more than 999 pairs.

26 Nov 2002 lmp2_extdom8.2002.5
Generalize domain extension. Distant pair domains now independent of strong pair domains.
Domains for asymmetric distant pairs now addressed by ioff_distdom+iorb.
Speed up lmp2 residual:
New option lmp2algo=2:
As lmp2algo=0, but scales quadratically. 2 times faster in small and medium size cases.
New option lmp2algo=3:
Similar to lmp2algo=3, but using united lmp2 domains: Should scale linearly for very large
molecules, but possibly slower for medium size molecules.
New option thrvrp: Threshold for skipping converged pairs in lmp2.
Pairs are skipped if their last update norm is smaller than the largest update
norm of the previous iteration times thrvrp (from 3rd iteration on).
Clean up some print (integer and format overflows).

26 Nov 2002 symluccsd4.2002.5
implements symmetry for open-shell lccsd

27 Nov 2002 symdluccsd.2002.5
implements symmetrie-option for direct luccsd
files
dgmat1.f src/lccsd/
dtraf_lmp2.f src/dtraf/
uccsd_loc.f src/lccsd/

20 Nov 2002 defbas_lmin.2002.5
Bug(s) fixed: 393
Specify minimum and maximum orbital angular momentum values in defbas.

29 Nov 2002 scf_formats.2002.5
Increase format width for 2-el repulsion.

29 Nov 2002 hp_comments.2002.5
The ibm compilers do not like c$HP$
Added preproccesor flags to comment out these lines on ibm machines

29 Nov 2002 ecpgrad_mpp.2002.5
Bug(s) fixed: 409
This patch fixes a bug in the ecp gradient calculation in mpp cases:
the global_sum was missing.

30 Nov 2002 dfmp2_variables.2002.5
Set additional variables CPUINV_CFIT and CPUFIT_CFIT in dfmp2.f

30 Nov 2002 fastrob2.2002.5
Faster algorithm for robust third term
Declare logical first in rib_lcc.f (hjw)

2 Dec 2002 dflmp2_fitdoms5.2002.5
Reduce memory requirement in dftrans_loc. Modify basis_gpt_info to
allow reduced max memory

3 Dec 2002 patcher_debug.2002.5
Fixed compilation of patcher when conditional debugging code is
included.

4 Dec 2002 fastrobfix3.2002.5
Faster algorithm for robust third term, bug fix

6 Dec 2002 luccsddiscon.2002.5
implements option discon=0,1 for luccsd
files
ccv_ucc_loc4.f src/lccsd/
uccsd_loc.f src/lccsd/

9 Dec 2002 makegalib_myrinet.2002.5
This patch removes libtstdio, libt, libtrillium, and libargs
from the list of libraries to look for in case of myrinet version.

9 Dec 2002 vscf_setup2.2002.5
Initial setup for vscf of Gerber et al.
New keyword vscf, and vscf_drv subroutine.
New routines get_frequencies and get_normalmodes in molpro_interface.f

10 Dec 2002 ecpmanual.2002.5
Update of the manual concerning ECP-calculations

11 Dec 2002 pagew3.2002.5
Implements paging for w in pairdom3 and shrinkfile

11 Dec 2002 integer4a.2002.5
Bug(s) fixed: 407 416
Deal with integer*4 parts of the code that will fail on
unix-nec/unix-i8, due to integer*8 being provided when integer*4 is
requested.

12 Dec 2002 patcher_files.2002.5
Bug(s) fixed: 418
Make patcher avoid non-directory files in the patch directory; give
more information when an error is flagged in installed_patch_stat.

12 Dec 2002 opdom.2002.5
Fixes a bug in opdom_loc (wrong lenj, lenk)

12 Dec 2002 clear.2002.5
Allows to clear more than one variable on one clear command

12 Dec 2002 patcher_bootstrap.2002.5
Bug(s) fixed: 413
Allow patcher to be built before successful configure. It is now
possible to do "make -C utilities patcher.bootstrap" from the top
directory.

13 Dec 2002 doc_install_patcher.2002.5
Bug(s) fixed: 412
Describe how to build patcher before making the program or before
configuring the program.

14 Dec 2002 matrop_export.2002.5
New option in matrop to import/export dump records. Useful for
transfering orbitals etc from 32 to 64 bit system or vice versa.

17 Dec 2002 matrop_export2.2002.5
Bug(s) fixed: 423
Added a missing endif which should have been present
in src/scf/matrop.f with mpp

19 Dec 2002 dfllmp2k.2002.5
Use fit domain info in transformation.

19 Dec 2002 defbas_poisson.2002.5
Sets minl=1 for MP2 aux basis sets

20 Dec 2002 sdiag2_init.2002.5
Initialize upper triangle of matrix constructed in sdiag2. Some vendor
LAPACK libraries fail if there is garbage in this.

20 Dec 2002 dec_conf2.2002.5
Identify problematic combinations of GNU Make and DIGITAL UNIX.

20 Dec 2002 cifile.2002.5

21 Dec 2002 jkbasen.2002.5
Adds Weigend's jk fitbasen to libmol and defbas

21 Dec 2002 dfexch.2002.5
New arguments fitcoul, fitexch, orb for subroutine hfma. Orbitals
passed to hfma for exchange fitting. Fix scf with exchange fitting.

2 Jan 2003 propa.2002.5
Fixes input for origin of operators.
Now the following possibilities should work again:
pot,0,x,y,z
pot,,x,y,z (same as previous)
pot(0),,x,y,z (same as previous)
pot,1 (for center 1, refers to z-matrix row)
pot(1) (same as previous)
pot(sym) (for center of atom sym)

2 Jan 2003 argos_labels2.2002.5
Change 11 bit to 12 bit encoding of labels. Symmetry labels now 3 bit only.

6 Jan 2003 sav_den.2002.5
Save CCSD relaxed and unrelaxed density matrices using save_dens

6 Jan 2003 direom.2002.5
Allow to use dkint for EOM-CCSD if iseom=1. Allow iflag>0 for isymt>1 in cckext.

7 Jan 2003 logcpu.2002.5
Increments ncall in logcpu also if init=1. Alows to call logcpu
with init=1 just to set cpu times (used in dfexch).

7 Jan 2003 Molcas_maxbfn.2002.5
Replaces fixed dimension for maxbfn by parameter value in Molcas/input.f

8 Jan 2003 leom3.2002.5
1) CIS with K/J op. in PAO basis, use of reduced jop list possible.
2) small changes in local EOM: vector-following from 0th iteration (not from 1st);
improve first approximation for local; clean some subroutines.
3) correct printing format

10 Jan 2003 ccsd_status.2002.5
ccsd, convergence, status
Set status correctly if convergence in ccsd
Always do triples, independent of convergence
New testjob for status check

10 Jan 2003 leom3cor.2002.5
correct error in leom3 in setting defaults in the case of negative iopdom

12 Jan 2003 dfmp2doc.2002.5
Update documentation for df-mp2 and df-lmp2

12 Jan 2003 blasflops.2002.5
Avoid integer overflop in calculation flop count in matrix multiplications

12 Jan 2003 timing2.2002.5
Calls citime from add_mpptim if nprocs=1
Replaces (estimated) I/O time by wall clock time in timing_molpro
Modifies print of CPU times at job end.

13 Jan 2003 dfexchb.2002.5
Fixes bugs in dfexch.f.
New thresholds thrcoul/throv in cfit_fock.f.
Disable change of scf screening threshold in fitting case.
Implements Schwarz screening and local coulomb
Large records in dftrans_loc and dfexch

14 Jan 2003 necsx_updates.2002.5
Updates to the NEC SX port; contributed by Dr. Janet L Fredin
(jfredin@atcc.necsys.com).

15 Jan 2003 newbatchs5.2002.5
Removes memory bottleneck from batchs.
An additional option, gpt_available, can be used to indicate whether GPT data has been
computed or not. If gpt_available is false all GPT data is computed on the
fly, and input arrays igpt_ao, rgpt_ao and tmatao are ignored.
HJW: Routines which call basis_3idx_batchs modified.
New CFIT parameter GPTCALC (default 0) added. If 1, GPT data are precomputed.

16 Jan 2003 l3opd1.2002.5
Local=3 with reduced operator list JOP (full domains). There is still small difference
in comparison with local=4,opdom=1 (>1.e-7), so for the time being it is
called by command eom,testop=1.

16 Jan 2003 accnt3.2002.5
Fixes incorrect usage of accnt in some cases.
Calls to accnt eliminated from ccsd_loc, uccsd, uccsd_loc.
Final CPU summary only printed if iprint(39).ge.0

17 Jan 2003 linux_ia64.2002.5
Updates from Janet Fredin for linux-ia64
New linux.conf not included
cckext.f changed: iad1, iad2 are initialized to zero,
and therefore the original check for .gt.0 is now .ne.0
(rather than .gt.ibase)

18 Jan 2003 trsm.2002.5
Bug(s) fixed: 433
A bug in transformation of sample CIS vectors from VMOs to PAOs was fixed (nty and ntx were mixed).
It caused that a wrong starting vector was found in one case in loc_eom3.test (but it finally converged
to the correct vector).

20 Jan 2003 newbatchs7.2002.5
Places correct memory management for GPT data caching in batchs initialization routine. And removes old error statement so that the batcher works with both precomputed and on-the-fly GPT data.

21 Jan 2003 ccv_loc3.2002.5
Fix OpenMP syntax error (illegal stop) in ccv_loc3.f

22 Jan 2003 cosmo2.2002.5
Basis cosmo implementation for rhf without symmetry.
The keyword cosmo allows default cosmo calculation.
Not tested for parallel version.
Changes by HJW:
Replaced binary fortran I/O by Molpro I/O
Replaced fortran formatted open/close by Molpro equivalent; get_tape.f eliminated.
Output and other formatted data files written to current directory.
The names are like output file name with .out replaced by .cosmo_...
Default values defined in cosmo.registry, and Molpro input utility used.
cosmo_init.f changed accordingly. read_cosmo_inp.f eliminated
calc_pot.f and traop_cosmo simplified (still very slow). Wrong
input stuff eliminated from calc_pot.f (screwed up input processing).
Print parameter added. Dielectric energy printed in each iteration
only if iprint_cosmo.ge.0
New testjobs and "make cosmotest" added.

22 Jan 2003 undef_cis.2002.5
correct undef in cis.f

22 Jan 2003 linux_ia64b.2002.5
Updates for linux-ia64

23 Jan 2003 compress.2002.5
Bug(s) fixed: 425
write last record in compress.f even if empty.
Check in readcmpr for zero length of buffer.

23 Jan 2003 ccsdp2.2002.5
add a more effective solver for Lambda, clean up files.

23 Jan 2003 linux_ia64_makegalib.2002.5
Bug(s) fixed: 438
Remove debugging sh -vx from bin/makegalib

24 Jan 2003 integrals.2002.5
Adds two new integral shell routines for future gaussian geminals code.

25 Jan 2003 cosmo_consts2.2002.5
Fixes wrong format in consts (compilation error on dec)
Replaces iout by 6 for error messages in consts.f

26 Jan 2003 config_lpar.2002.5
lsattr doesn't work properly on IBM LPAR systems, changing to lscfg

27 Jan 2003 mo_squeeze.2002.5
Updates for mo_squeeze. Write no-lmos to dump record.
Atoms added by domain extensions or with charges .lt.0.5 set negativ in idoma list.
orbdom.f mo_squeeze.f modified accordingly.
Subroutine mrgwl_dom moved from opdom.f to orbdom.f
Statistics for lmos

28 Jan 2003 locef2.2002.5
1) Correct error in offsets for eom domains. Results for small molecules
are not affected.
2) Correct error in finish_ffapr, resulting in chosing nonoptimal start vector
in some cases.
3) local=3 with reduced list of K/J.

29 Jan 2003 proc_set.2002.5
Bug(s) fixed: 428
Use new syntax to set system variables; avoid trying to set a read
only variable (ENERGY). Examples and documentation updated.

29 Jan 2003 ios_direct.2002.5
Remove ios_direct functions.

29 Jan 2003 nsdavid.2002.5
Bug(s) fixed: 440
avoid calculating ws and wp for the CIS final print

30 Jan 2003 dleom.2002.5
Add domain offsets for t1 in dkint to allow to use it in local EOM-CCSD.

30 Jan 2003 dfexchp.2002.5
Optimizes memory usage in dfexch.
Implements paging of J-matrix (option PAGEJ)
Implements inversion of precomputed local J-matrices (option INVERT)
Fixes call to df_writebl in dfmp2.f

31 Jan 2003 table_error.2002.5
Bug(s) fixed: 441
Avoid plotting a table when an error has been detected.

1 Feb 2003 seward_potint5.2002.5
New routine to compute potential on a grid.
Preparation for direct computation of one-electron
operators in read_op and seward.
Get symmeytry working in drv_pot
Genberalization of drv_pot to electric field case

2 Feb 2003 maxdim2.2002.5
Increases the upper limit for maxbfn from 2000 to 4095 and maxval to 512
Remove object files if dimensions change.

4 Feb 2003 molpro_interface.2002.5
Adds a routine get_dipole to molpro_interface.f

4 Feb 2003 aosort_bucksize3.2002.5
Bug(s) fixed: 442
Increases max bucket size and segment size. Much better performance on HP

4 Feb 2003 iobench_mpp.2002.5
Fix iobench: set variables on all processors, and avoid division by zero

4 Feb 2003 dkexts_mpp.2002.5
Adds missing sync in sortopk

4 Feb 2003 test_cosmo.2002.5
Increases memory in c2h4_cosmo.test

4 Feb 2003 ccsd_loc.2002.5
Remove undefined variable rext from ccsd_loc.f

5 Feb 2003 powell_return_flag2.2002.5
Fix missing integer declaration in min2d.f

5 Feb 2003 max_buffer.2002.5
Bug(s) fixed: 417
Limit the size of buffers written/read by writem/readm. Buffers that
are too large can cause failure on some OSs.

5 Feb 2003 gio4.2002.5
Use improved broadcast routine from Karsten's libgatools.a where applicable.

6 Feb 2003 ifc7.2002.5
Intel Fortran compiler version 7: changed module handling leads to revised makedepend options
and the possibility for parallel make in a directory.

6 Feb 2003 leom4.2002.5
Fix bug (itran3) in cceom_loc.f

6 Feb 2003 drvprop.2002.5
Fixes problem in drvprop, causing crash on sun

6 Feb 2003 sunconf.2002.5
Lower optimization for util/gparam_init.f on sun
This avoids crash of mpp version.

6 Feb 2003 ccmp3_loc3.2002.5
Fixes crash in ccmp2_loc3 due to division by zero if a processor has zero work

6 Feb 2003 cosmo_mpp.2002.5
Writes cosmo output files only on processor 0. Avoid crash on hp/mpp

7 Feb 2003 taskprint.2002.5
Add global print option TASKS to print tasks lists
Move parameters nprt, nprc to common/cprint
Clean up a number of files, containint either nprc or nprt

7 Feb 2003 linux-ia64d.2002.5
updates from Kirk Peterson for linux-ia64 (hp)

7 Feb 2003 mpptune.2002.5
Quick tuning job to determine mpp tuning parameters

8 Feb 2003 getmachine_linux.2002.5
Correct wrong handling of ifc/efc in getmachine. ifc was discared if efc not found

8 Feb 2003 hpconf.2002.5
Adds ccv_loc4.f to the f_opt1 list for hp-ia64
Otherwise crash with mpp=2

8 Feb 2003 multi_numforce.2002.5
save,grad=-1 option to avoid error message if no cpmcscf
is done in a numerical sa-mcscf geometry optimization

8 Feb 2003 get_inpl.2002.5
Adds routines to handle logicals in getinp.f
get_inpl_g, get_inp, set_inpl_g, set_inpl etc.
Logical can be set in registry files, using 'l' or 'L' as
type, and either 1 or .true. for true and 0 .or. .false. for
false. Internally, false is stored as 0.d0 and true as 1.d0

8 Feb 2003 doc_variables.2002.5
Add some stuff about system variables to control.tex.
Minor additions for basis set specification (basis block types) in
integral.tex. Documentation for dfit still missing!!

9 Feb 2003 checkinp.2002.5
Fixes checkinp to recognize array variables

11 Feb 2003 cfit.2002.5
Bug(s) fixed: 376
Fixes bug causing crash on alpha linux hardware in CFIT testjobs.

11 Feb 2003 configure_gio3.2002.5
Modifies configure to touch common files which depend on the gio preprocessor
option. All files depending on these commons must be recompiled. It is assumed
that this is done automatically.
Fixes a bug in iow.f (only with gprint,io=1 and gio active)

12 Feb 2003 string_utils.2002.5
Collects various string utilitities in a new file string_utils.
Adds documentation for this to progman

12 Feb 2003 dfexch_mpp9.2002.5
Fully parallelizes dfexch.f and basis_vec_contract_A
Implements cfit new option dfmpp:
mpp=0: use global_sum in sort step (parallel read)
mpp=1: use brdcst in sort step
Adds support for gio to dfexch.f
Let mpptim print timings if nprocs=1 and iprint.ge.1
New cfit options dfimpl, minfit for mpp tuning

12 Feb 2003 mpham.2002.5
Fix a few bugs in mpham

13 Feb 2003 ucckext_mpp2.2002.5
Fix mpp case in ucckext. Variable len was undefined when incore
was determined.

2003-02-13 405 2003 solve4.2002.5
SOLVE command for solution of simple non-linear equations using Secant, Regula-Falsi or Ritter
methods. Basically working, but robustness and method still to be done.

13 Feb 2003 parse_registry.2002.5
Removes double slash in full_filename

13 Feb 2003 bmat_mpp4.2002.5
Modifies bmat.f and subroutines to do all fortran I/O only on processor 0
New testjob caffeine_opt.test
Symmetry clean resulting coordinates in bmat.f

2003-02-13 251 2003 ibm_no_omp.2002.5
Bug(s) fixed: 435
The -qsmp=omp directives were removed from the fortran options
as they caused the fci testjob to fail with i4.

2003-02-13 310 2003 makedistrib.2002.5
Bug(s) fixed: 371 422
Added extra filters when creating distribution to remove
molpros_* molprop_* parallel molpro.rc
report report.*.tar.gz *.out *.errout
Rcsdiff commit msplit

14 Feb 2003 vscf.2002.5
vscf implementation for molpro

14 Feb 2003 vscf2.2002.5
Adds new input options ncoup and iexc to vscf_drv.f

14 Feb 2003 binput.2002.5
Bug(s) fixed: 394
Fixes for binput:
1. disallow zmat or geometry as a subcommand (following input) of int
2. minimum ratio for even-tempered basis now can be specified through
variable even_min_ratio
3. increased limits for large primitive sets (also in basis.f)
4. Stop some memory leaks in basis.f

14 Feb 2003 zmat_d2.2002.5
Bug(s) fixed: 250
Internally change specified symmetry generators yz,xz to xz,yz and
yz,xy to xy,yz so that argos and seward stand a chance of working.

16 Feb 2003 doc2.2002.5
documentation updates: wf card, system variables, numforces, frequencies

16 Feb 2003 optandfreq.2002.5
Fixes auto symmetry option for mp2/ccsd/rccsd.
Fix bmat optimization with numerical gradients.
Clean up COORD input options for forces, numerical and optg.
Fixes ovl2bas to treat correctly orbital conversion between different zymmetries.
Definition of symmetry-unique atoms is not stored in dump record (2002.6 only).
Reorder symmetry operations provided by variable ZSYMEL in aoint.f and zmatrix.f.
Fix start option in scf.

16 Feb 2003 ovl2bas_checkversion.2002.5
Check version in ovl2bas and revert to old code if necessart

16 Feb 2003 fujitsu_testjobs.2002.5
Increas memory in cl2dft.test and cr2_ecp_forces.test to run on fujitsu vpp

16 Feb 2003 loc_eom.2002.5
Activate read of eps in do_resid.f for all cases. In some cases the array eps was
undefined, leading to crash on fujitsu. Now some calculations in loc_eom.test
converge MUCH faster! But slight changes of the results (ca 1.d-7),
therefore testjob modified.

16 Feb 2003 undefs.2002.5
Fixes undefined variables (harmless)

16 Feb 2003 undefs2.2002.5
Further undefined variables fixed

18 Feb 2003 tradf2_clean.2002.5
Call clean in tradf2 to set undefined matrix elements to zero

18 Feb 2003 fujitsu.2002.5
Updates from Ross Nobes <r.nobes@fle.fujitsu.com>.

18 Feb 2003 cxml.2002.5
Check for cxml library.

18 Feb 2003 examples4.2002.5
Updated files from examples/ to use correct syntax for setting
variables; added memory allocation statements to some testjobs; use a
Makefile.

2003-02-18 171 2003 patchlevel_molver.2002.5
Bug(s) fixed: 468
Molver side of patchlevel

2003-02-18 171 2003 ldip_densav.2002.5
Bug(s) fixed: 470
Fix uninitialized densav_ori in src/eom/ldip.f

19 Feb 2003 cilscc2.2002.5
Put dimensioned arrays in cilscc into dynamic memory, avoiding
crash if mxval is large. Also, I believe the dimensions were wrong
anyway (nsoc can be larger than nval, and the arrays jcaxa/b are written
with i=0 but the dimensions started with 1). Peter, please double check
this!!!

19 Feb 2003 intdone-after-delete2.2002.5
Bug(s) fixed: 472 475
Ensure that INTDONE gets unset on DELETE,1, otherwise for sorted integral
file format 'eaf' the program thinks the integrals are available.
Also fix checking of INTDONE.

2003-02-19 228 2003 sgi_conf2.2002.5
Bug(s) fixed: 470
lower optimisation to -O2 src/multi/muaug.f to
fix wrong numbers in testjobs

19 Feb 2003 pseudo_libmol.2002.5
Newest version of pseudo.libmol

19 Feb 2003 propgrid5.2002.5
New command propgrid to compute properties on a grid
All Molcas operators implemented. POT, EF, FG efficiently,
others brute force.

19 Feb 2003 makedepend2.2002.5
Bug(s) fixed: 457 474
Get makedepend to recursively check for included FORTRAN files; added
verbose option (-v).

19 Feb 2003 vscf4.2002.5
Adds new input options ncoup and iexc to vscf_drv.f
Get restart option working, add a testjob for restart

2003-02-20 218 2003 t-aug-cc.2002.5
d-aug and t-aug basis sets by even-tempered extrapolation as specified
by van Mourik et al, Mol Phys 96 (1999) 529.

2003-02-20 228 2003 install-html.2002.5
Minor change to doc/Makefile to avoid warning message from chmod on
make install-html

20 Feb 2003 batchs_sphtr.2002.5
Improve contraction and spherical transformation steps in batchs

20 Feb 2003 makegalib_ia64.2002.5
Remove non-general configuration.

21 Feb 2003 slater_startdens2.2002.5
Prepares for slater start density

2003-02-21 173 2003 power4_i8.2002.5
changed getmachine to make integer8 default on power4

21 Feb 2003 vscf5.2002.5
Get restart option working, add a testjob for restart