CVS log for ccp4/src/mtzutils.f
	Help

Request diff between arbitrary revisions / Display revisions graphically

merge to branch-30_07_11

[#4551] and [#4555]. Increased MAXBAT and similar parameters in various places, plus KPRI in /home/fcx32934/ccp4_test/ccp4/lib/src/ccplib.f. Removed parameter_bigdet.fh and temp.fh from /home/fcx32934/ccp4_test/ccp4/src/scala_

Increasing the value of LINE as requested by ejd

update licence banners

change action of CELL keyword so that it updates _all_ dataset cells
call to LWCELL is obsolete and has been removed.

EJD: request that line for keyworded input be extended to 1500 characters,
to cope with longer MTZ labels from CCP4i.

add new licence banners

changes from Clemens

update assignment of batches to datasets, for crystal level

Correct usage of LRHIST is to supply the dimension the history array on
input.

Increase size of character string H1, which is ultimately sent to GETENV.
Change is to fix problem under Intel compilers: if this string is too small
to hold the result then it is left blank, rather than being truncated.

first stab at handling crystal level

Added a warning if SYMMETRY is selected with multirecord MTZ file input.

maxcol -> mcols -> 500 (CV)

CV - maxcol -> MCOL -> 500

Changes to carry dcell and dwavel from input to output

raised MAXBAT to 5000

Changes to deal with dataset headers, and dataset pointers for
columns and batches.

reset_magic altered.

UNIQUE option data from file 2 was being output instead of file 1. Associated
columns properly scaled with SCALE option.

Trouble with PRUNIQ. H, K and L work not assigned properly when using
INCL/EXCL.

LWCLOS was bypassed in some cases. Fixed.

Problems with AXIS/ZONE output. Now OK.

Missing number stuff built from r1.25  i.e. Dave's attempt was overwritten.
Seems to be OK for all options, no functionality was removed.

Previous missing data change was insufficient.

Use NaNs (not 0.0) for missing data in output on joins etc.

192 symops

The UNIQUE option zeroed columns from hklin2 for reflexions not in hklin1.

use new qnan

EJD: the UNIQUE option failed when there
were more reflections in hklin2 than in hklin1
This works now, and will not have DESTROYED anything for other options.
However there may still be errors I have not encountered.
DL: avoid -999.0 marker (use qnan)

Now defaults to ONEF if HKLIN2 not assigned and also checks for HKLIN;
per old documentation but wasn't implemented previously.

Declared LSTLSQ as real and external, fix VMS problem.

Error message output when using 'INCL/EXCL all' if two output columns are the
same.

EJD: didnt find HHL reflections correctly:
  Not sure how lots of it could have worked?

comments HEADER arg

Tidy up input in line with parser changes.
LOGZON wasn't defined in ONEDSN.

Remove bare `x' format items; some tidying.

Add/change history info writing for a bunch of programs.  Mostly call the
new LWHSTL.  Several previously called incorrectly LWHIST with scalar
instead of character array.

DUNTEN@XRAY.BMC.UU.SE:
Problem:   with 2 input files, options >include 1 all
                                       >include 2 PHIS12 FOMS12
                                       >reso 100 3.5
   ==> the hklout .mtz file doesn't obey the reso restriction.

Cause:     mtzutils read a reflex from hklin1, then hklin2, and
           if they are not the same hkl, and if hkl1 comes before
           hkl2 in the sort order, it should output the
           reflex from hklin1 if it obeys the reso restriction.
           However, the reso of the reflex from hklin2 is used to
           do the reso check, so it may pass the test while the
           correct check of the reso of the reflex from hklin1
           would have failed.  I suppose it might also fail the
           reflex when it shouldn't.

Fortran:   lines 4386-4388 read the reflex for these mtzutils options.

104   IF(NINT(RECIN_1(NHC1)-RECIN_2(NHC2)))  101,102,103
102   IF(NINT(RECIN_1(NKC1)-RECIN_2(NKC2)))  101,105,103
105   IF(NINT(RECIN_1(NLC1)-RECIN_2(NLC2)))  101,106,103

           and the problem reflex will goto 101, after which the

      CALL CHKRES(RESOL,OK)

           call uses RESOL from hklin2 rather than hklin1.

SCALE option from Howard

version pulled straight from york

SAVE changes

! comment

Fixed long-standing nasty bug in mtzutils, Merge option. As it
was, once it had hit the end of file 2, it ignored all the
rest of file 1.   (Phil)

portability

from PB: change CLEAN arg of SETOUT in PRMERG (not sure what this fixes)

had missed old Hamburg fix:  Program wasn't applying resolution cutoffs to
the output file.

brought in line with lmb/york

change ccpvrs to ccprcs

EJD: It lists ALL reflections in a zone - boring for 6000

portability

prior to proper attempt at version control

Roughly correposnds to first `official' release of MTZ version for un*x
and VMS around 14--15/5/92.

Initial revision