 |
Collaborative Computational Project No. 4 |
 |
|---|
|
|
Collaborative Computational Project No. 4 |
|
|---|
| Home | About CCP4 | CCP4 Projects | Downloads | Documentation | Courses | Developers | CCP4 people | WG1/WG2 | Privacy |
|---|
These are the known problems with release 4.1 of the CCP4 suite, together with fixes where appropriate. Code changes are given as patches, which can be applied with the UNIX patch utility. New files are put in the prerelease directory, which can be found at ftp://ccp4a.dl.ac.uk/ccp4/prerelease. And there is some more info on what the files are.
Hints and advice about applying the fixes can be found here.
Warning for Linux users: if you are running CCP4 compiled with gcc 2.96 may be unreliable.. This has been seen with RedHat 7.0 and Mandrake 8.1.
A list of general problems with CCP4i can be found here.
Under IRIX64 mosflm may take a considerable amount of time to compile, 600+ minutes. This is due to the setting of Olimit=0 (some cases only) which optimises all routines, regardless of length.
The solution is to set the Olimit values in the configure script to a sensible value (1500 is used in most cases).
diff -C1 -c -r1.126 -r1.127
*** configure 2001/01/31 14:41:52 1.126
--- configure 2001/03/15 10:01:27 1.127
***************
*** 745,747 ****
IRIX64\ 6.5*|IRIX\ 6.5*)
! XFFLAGS=${XFFLAGS-"-64 -mips3 -nocpp -OPT:Olimit=0 -w"}
XCFLAGS=${XCFLAGS-"-64 -DPROTOTYPE -w"}
--- 745,747 ----
IRIX64\ 6.5*|IRIX\ 6.5*)
! XFFLAGS=${XFFLAGS-"-64 -mips3 -nocpp -OPT:Olimit=1500 -w"}
XCFLAGS=${XCFLAGS-"-64 -DPROTOTYPE -w"}
***************
*** 755,757 ****
*)
! XFFLAGS=${XFFLAGS-"-64 -mips3 -nocpp -OPT:Olimit=0 -w"}
XCFLAGS=${XCFLAGS-"-64 -DPROTOTYPE -w"}
--- 755,757 ----
*)
! XFFLAGS=${XFFLAGS-"-64 -mips3 -nocpp -OPT:Olimit=1500 -w"}
XCFLAGS=${XCFLAGS-"-64 -DPROTOTYPE -w"}
***************
*** 781,783 ****
IRIX64*)
! XFFLAGS=${XFFLAGS-"-OPT:Olimit=1500 1500 -nocpp"}
# XLDFLAGS=${XLDFLAGS-"-lfpe"} # to enable floating point traps
--- 781,783 ----
IRIX64*)
! XFFLAGS=${XFFLAGS-"-OPT:Olimit=1500 -nocpp"}
# XLDFLAGS=${XLDFLAGS-"-lfpe"} # to enable floating point traps
Alternatively config.status may be similarly edited.
(Thanks to Robert Campbell for spotting this)
Under certain circumstances the configure script in xdl_view will fail with:
configure: error: newly created file is older than distributed files! Check your system clock
This occurs when the filesystem with ccp4 is mounted on a remote machine, and the remote machine's clock is slower than the local machine's. A solution is, for the configure script in xdl_view:
diff -c -C1 configure configure_new *** configure Thu Jan 18 09:33:26 2001 --- configure_new Fri Mar 16 11:40:23 2001 *************** *** 971,972 **** --- 971,973 ---- echo timestamp > conftestfile + touch conftestfile # Do `set' in a subshell so we don't clobber the current shell's
(Thanks to Randy Read for the suggestion)
There is a bug in Amore which shows itself only when the search is in a monoclinic space group where b is shorter than c. Thanks all those who pointed the error out (and Eleanor for fixing it!).
Please download and
install a fixed amore.f from the
prerelease area.
Problem symptom is a program exit of the following nature:
> S/R PAC4 < MAXIMAL INDICES EXCEEDED: Indices are: 6 -52 0 Limits are: 83 48 106 Adding: 168 100 216 Number of symmetry equivalent: 2 Limit is: 5 Generated Packed number ( should be > 0..) 890108 AMORE: > s/r pac4 < maximal indices exceeded:
(Thanks to Eleanor Dodson for the changes.)
An error in the refmac5 output means that some graphs fail to display in (x)loggraph.
Please download and install fixed versions of the subag_scale.f and refmac_newdic.f files from the prerelease area. NB: these two files should be put into the $CPROG/refmac5_ directory.
If you are using REFMAC5 and you have residues like NAD, NDP, FAD etc then the program may stop with the message:
No match for full atom ID - subtract one characterIt is related to the fact that REFMAC5 cannot extract element names for these atoms. The solution is to add element names for these residues.
Garib Murshudov has written a small jiffy program called add_element to do this - a .tar.gz file with the source code and a README file (explaining how to compile and run the jiffy) can be downloaded from the prerelease area.
Under some strict compilers (e.g. Portland Group), REFMAC5 may fail trying to open the same file on two different unit numbers simultaneously.
Apply the following patch to $CPROG/refmac5_/lsq_newdic_rest.f and remake refmac5:
diff -c -r1.4 lsq_newdic_rest.f
*** lsq_newdic_rest.f 2001/02/28 18:12:55 1.4
--- lsq_newdic_rest.f 2001/03/22 09:35:55
***************
*** 3644,3649 ****
--- 3644,3651 ----
ENDIF
GOTO 10
999 CONTINUE
+ CLOSE(ISCRB)
+ ISCRB = 0
RETURN
END
C
(Thanks to Ian Tickle)
NB: a "broken" version of this fix was originally posted here on 20/03/2001 - under certain systems it generated corrupted values for the mean and rms densities in the map header.
If you have already applied the earlier version of the patch then the second patch needed to fix the problem is:
diff -c -r1.36.2.1 -r1.36.2.2
*** maplib.f 2001/03/22 17:17:23 1.36.2.1
--- maplib.f 2001/06/01 08:55:47 1.36.2.2
***************
*** 374,380 ****
ARMS = 0.0
SUMRHO = 0.0
SUMRH2 = 0.0
! OFFSTR = 2.0*QOFFSET
DO 30 I = 1,20
DO 20 J = 1,10
READ (BLANK,FMT=6002) LABELS(I,J)
--- 374,380 ----
ARMS = 0.0
SUMRHO = 0.0
SUMRH2 = 0.0
! OFFSTR = 2.0*QOFFST
DO 30 I = 1,20
DO 20 J = 1,10
READ (BLANK,FMT=6002) LABELS(I,J)
Otherwise, the full corrected patch is below.
Many apologies for any inconvenience this has caused, and to Ian Tickle (who's original fix was mangled) - pjb
If a program creates several maps in the same run then the second (and subsequent) maps may have random values for the mean and RMS densities - this is because the sums are not being re-initialised in maplib.
Apply the following patch to $CLIBS/maplib.f to fix this:
diff -c -r1.36 -r1.36.2.2
*** maplib.f 2000/02/25 13:24:47 1.36
--- maplib.f 2001/06/01 08:55:47 1.36.2.2
***************
*** 330,342 ****
COMMON /MOHDR/NC,NR,NS,MODE,NC1,NR1,NS1,NXYZ(3),CEL(6),MAPCRS(3),
+ AMIN,AMAX,AMEAN,ISPG,NSYMBT,LSKFLG,SKWMAT(3,3),SKWTRN(3),
+ JUNK(17),ARMS,NLAB,LABELS(20,10),NCHITM,ITMHDR,ITMSC1
COMMON /MSTRM/LSTRM(12)
C ..
C .. Equivalences ..
EQUIVALENCE (NC,HEADER(1))
C ..
C .. Save statement ..
! SAVE /MSTRM/,/MOHDR/,FILE
C ..
C .. Data statements ..
C
--- 330,345 ----
COMMON /MOHDR/NC,NR,NS,MODE,NC1,NR1,NS1,NXYZ(3),CEL(6),MAPCRS(3),
+ AMIN,AMAX,AMEAN,ISPG,NSYMBT,LSKFLG,SKWMAT(3,3),SKWTRN(3),
+ JUNK(17),ARMS,NLAB,LABELS(20,10),NCHITM,ITMHDR,ITMSC1
+
+ COMMON /MOHSUM/ SUMRHO, SUMRH2, OFFSTR
+ DOUBLE PRECISION SUMRHO, SUMRH2, OFFSTR
COMMON /MSTRM/LSTRM(12)
C ..
C .. Equivalences ..
EQUIVALENCE (NC,HEADER(1))
C ..
C .. Save statement ..
! SAVE /MSTRM/,/MOHDR/,/MOHSUM/,FILE
C ..
C .. Data statements ..
C
***************
*** 369,374 ****
--- 372,380 ----
AMAX = -99999999.0
AMEAN = 0.0
ARMS = 0.0
+ SUMRHO = 0.0
+ SUMRH2 = 0.0
+ OFFSTR = 2.0*QOFFST
DO 30 I = 1,20
DO 20 J = 1,10
READ (BLANK,FMT=6002) LABELS(I,J)
(Thanks to Ian Tickle)
Attempts to run the Strategy in Rotgen fail with:
Error message: Error opening symmetry operators file (SYMOP)
The environmental variable SYMOP is not set. Set the environment variable SYMOP to point to $CLIBD/symop.lib
A related problem; a call to CCPFYP is missing from rotgen.f. This will allow the setting of SYMOP from the command line:
diff -c -r1.1 -r1.2
*** rotgen.f 2000/10/24 13:37:16 1.1
--- rotgen.f 2001/03/29 08:59:07 1.2
***************
*** 240,245 ****
--- 240,246 ----
C
CALL INIT_FLAGS
CALL INIT_SETTINGS
+ CALL CCPFYP
CALL MTZINI
C
C====== Initialise the xdl_view routines
(Thanks to Dave Love)
An error in the rwbrook library disables the fractional conversion for pdb files when REMARK 290 SYMTRY is present. This has the knock-on effect that calls to XYZCOORD to return FRACTIONAL coordinates are disabled. As yet this is only known to affect acorn. The following corrects this:
diff -c -r1.62 -r1.61
*** rwbrook.f 2001/04/25 07:53:10 1.62
--- rwbrook.f 2001/04/23 10:45:28 1.61
***************
*** 304,310 ****
PARAMETER (MAXFILESOPEN=90,MAXSYM=96)
C ..
C .. Arguments ..
! INTEGER IFAIL,IUNIT,ICRYST,II,III,JJ,K,ISYM
REAL AM,BM,RCHK1,RCHK2,FAC
CHARACTER*(*) FILTYP,LOGNAM,RWSTAT
C ..
--- 304,310 ----
PARAMETER (MAXFILESOPEN=90,MAXSYM=96)
C ..
C .. Arguments ..
! INTEGER IFAIL,IUNIT,ICRYST,II,JJ,K,ISYM
REAL AM,BM,RCHK1,RCHK2,FAC
CHARACTER*(*) FILTYP,LOGNAM,RWSTAT
C ..
***************
*** 460,466 ****
READ(BROOKA(11:18),'(I8)')NSYMCHK
CALL symfr2 (BROOKA,22,nsymchk,rsymchk)
write(6,'(a,i3,4(/,4f10.3))')' remark 290',nsymchk,
! + ((rsymchk(iii,jj,nsymchk),iii=1,4),jj=1,4)
ENDIF
ENDIF
C
--- 460,466 ----
READ(BROOKA(11:18),'(I8)')NSYMCHK
CALL symfr2 (BROOKA,22,nsymchk,rsymchk)
write(6,'(a,i3,4(/,4f10.3))')' remark 290',nsymchk,
! + ((rsymchk(ii,jj,nsymchk),ii=1,4),jj=1,4)
ENDIF
ENDIF
C
(Thanks to Eleanor Dodson)
A minor bug in detwin produces confusing output in the summary, where the twinning fraction used for the output mtz file is alway reported as 0.48. This does not affect the running of the program. This also occurs for CCP4 version 4.0.x
diff -c -r1.7.2.1 -r1.7
*** detwin.f 2001/04/25 15:23:37 1.7.2.1
--- detwin.f 2000/06/15 09:23:45 1.7
***************
*** 863,869 ****
if(twfac.gt.0.001)call lwclos(mtzout, iprint)
c
call ccp4h_summary_beg()
! write(6,1020) twfac,nout,nequ,nmiss
1020 format(/,
+ 'Output file written for twinning fraction',f6.3,/,
+ 'Number of reflections written to output file',I7,/,
--- 863,869 ----
if(twfac.gt.0.001)call lwclos(mtzout, iprint)
c
call ccp4h_summary_beg()
! write(6,1020) alpha,nout,nequ,nmiss
1020 format(/,
+ 'Output file written for twinning fraction',f6.3,/,
+ 'Number of reflections written to output file',I7,/,
(Thanks to Dr Nicholas Keep for spotting this)
AREAIMOL cannot treat ANISOU cards correctly in input PDB files - you will get additional phantom atoms which will distort the accessible surface area values and the analyses.
Please download and install a fixed versions of AREAIMOL from the areaimol-110501.f file on the prerelease area.
(Thanks to John Jenkins for reporting this)
There is a problem with the INPUT USER option of COMBAT - it doesn't work.
A fixed version of the combat files can be downloaded as the file combat-080801.tar.gz from the prerelease area.
Note that this updated version has some other changes to the version
distributed with 4.1.1 - most noticeably: the MTZF option now accepts F(+) and
F(-) as an alternative to mean F, and the MISBATCH option has been removed: the
program now only outputs batches which are present in the input file.
It also fixes a more recently discovered bug, see below.
Ian Tickle reports that changes made between version 4.0 and 4.1* have broken the program. There is also another minor fix where an array is uninitialsed and causes problems on some OSs.
A fix is available on the prerelease area.
act fails to recognise HOH as solvent, confusing it with HIS, when it has non-zero occupancy. This is corrected by moving the test for solvent before that for the main chain atoms in BSTATS.
A fix is available on the prerelease area.
COMBAT sets the lattice type wrongly for centred spacegroups in MTZ files - it is always set to "P".
A fixed version (which also includes the fixes for earlier reported problems) can be downloaded as the file combat-080801.tar.gz from the prerelease area.
(Thanks to Phil Evans for fixing this)
On some systems LIBCHECK may crash when trying to load certain library descriptions e.g. MO6. This problem occurs whether the program is run from the command line or from the Monomer Library Sketcher interface in CCP4i.
The following patch must be applied to $CPROG/refmac5_/libcheck.f:
*** libcheck.f 2001/01/22 14:37:44 1.6
--- libcheck.f 2001/09/11 13:34:39 1.7
***************
*** 3141,3147 ****
CALL NB_VMOD(V2,AV2)
CALL NB_VPROD(V1,V2,S)
COSA1= S/(AV1*AV2)
! ANG1 = ACOS(COSA1)
ANG1 = PI-ANG1
C --
C ----Tabl. I4 AV THETA --------------------------------
--- 3141,3147 ----
CALL NB_VMOD(V2,AV2)
CALL NB_VPROD(V1,V2,S)
COSA1= S/(AV1*AV2)
! ANG1 = ACOS(AMAX1(AMIN1(COSA1,1.0),-1.0))
ANG1 = PI-ANG1
C --
C ----Tabl. I4 AV THETA --------------------------------
***************
*** 5555,5561 ****
CALL NB_VMOD(V2,AV2)
CALL NB_VPROD(V1,V2,S)
COSA1= S/(AV1*AV2)
! ANG1 = ACOS(COSA1)
ANG1 = PI-ANG1
C --
C ----Tabl. I4 --------------------------------
--- 5555,5561 ----
CALL NB_VMOD(V2,AV2)
CALL NB_VPROD(V1,V2,S)
COSA1= S/(AV1*AV2)
! ANG1 = ACOS(AMAX1(AMIN1(COSA1,1.0),-1.0))
ANG1 = PI-ANG1
C --
C ----Tabl. I4 --------------------------------
(Thanks to Garib Murshudov)
There are two problems with coordconv:
coordconv xyzin orig.pdb xyzout tmp.frac <<eof-1 INPUT PDB ORTH 5 OUTPUT FRAC CELL 9.488 16.853 31.684 95.77 98.07 99.4500 eof-1 coordconv xyzin tmp.frac xyzout reorth.pdb <<eof-2 INPUT FRAC OUTPUT PDB ORTH 1 CELL 9.488 16.853 31.684 95.77 98.07 99.4500 eof-2Note that PDBSET has a REORTHOGONALIZE keyword which can also be used to switch between different orthogonalisations.
(Thanks to Nick Keep)
It has long been known that there are problems with some ccp4 probrams compiled under Linux RedHat 7.0 and 7.1. If your compilers are uptodate (gcc version 2.94-85) or are the RedHat 7.1 native compiler replacing the $CLIBS/symlib.f solves the majority of these problems. However, caution is still recomended. A fixed symlib.f (symlib270921.f) is available on the prerelease area.
(Thanks to Garib Murshudov)
The limits of the real-space asymmetric unit for I212121 listed in s/r SETLIM are wrong. A fixed symlib.f (symlib.f) is available on the prerelease area.
Apparently scalepack will sometimes (probably more recent version) subtract the batch number by 1 - giving you a batch number of zero. Zero is not a legitimate batch number in a multibatch mtz file so addbatch must be set to at least 1. This has shown to be a problem when outputting with scalepack option "nomerge original indices" with a very recent version of the HKL package.
This problem can be spotted as combat (and rotaprep) give the obscure message:
*** Error From LWBSETID : no batches present to write to! combat: *** Program Terminated Times: User: 0.8s System: 0.1s Elapsed: 0:01
The EXTEND COPY mode is badly broken. A new version of the $CPROG/mapmask_/mapmask.f file can be found as mapmask-251001.f on the prerelease area. Recompiling mapmask with the updated file should fix the problem.
(With thanks to Kevin Cowtan)
Clemens Vonrhein vonrhein@globalphasing.com: there is a bug in the EXPAND and REDUCE routine when a phase shift is necessary and the input file contains a mixture of phases (type 'P') and Hendrickson-Lattman coefficients (type 'A'). Re-using the pshft variable for the C and D HLs means that columns of types 'A' or 'P' _following_ the first set of Hls are treated wrong.
diff:
3311c3311 < real pshft,phares,packmat(9),gt_real,phashft,value,inrange --- > real pshft,pshft2,phares,packmat(9),gt_real,phashft,value,inrange 3423,3425c3423,3425 < pshft=pshft*2 < ref(kpos+2)=value*cos(pshft)-ref(kpos+3)*sin(pshft) < ref(kpos+3)=value*sin(pshft)+ref(kpos+3)*cos(pshft) --- > pshft2=pshft*2 > ref(kpos+2)=value*cos(pshft2)-ref(kpos+3)*sin(pshft2) > ref(kpos+3)=value*sin(pshft2)+ref(kpos+3)*cos(pshft2) 9245c9245 < real pshft,phares,value,inrange --- > real pshft,pshft2,phares,value,inrange 9353,9355c9353,9357 < pshft=pshft*2 < ref(noff+k+2)=value*cos(pshft)-ref(noff+k+3)*sin(pshft) < ref(noff+k+3)=value*sin(pshft)+ref(noff+k+3)*cos(pshft) --- > pshft2=pshft*2 > ref(noff+k+2)=value*cos(pshft2)- > . ref(noff+k+3)*sin(pshft2) > ref(noff+k+3)=value*sin(pshft2)+ > . ref(noff+k+3)*cos(pshft2)Line numbers may vary slightly.
In the 'Prepare Data for HA Search' task the input 'SAD data as F+F-' to prepare data for Rantan does not output the correct MTX column labels to the command file.
Download a fixed prephadata.script from the prerelease area.
An error in $CCP4I_TOP/bin/loggraph.tcl means that if loggraph is run before ccp4i the configure.def file will not be created correctly. The stack trace:
Running auto-configure for configure
Creating new configure parameters file
/prog/ccp4/4.1.1/ccp4i/etc/unix/configure.def
Error in startup script: invalid command name "SaveArray"
while executing
"SaveArray $taskname $filename $arrayname -save_types"
(procedure "InitialisePreferences" line 45)
invoked from within
"InitialisePreferences configure configure"
(file "/prog/ccp4/4.1.1/ccp4i/bin/loggraph.tcl" line 52)
invoked from within
"source [file join $env(CCP4I_TOP) bin loggraph.tcl]"
(file "/prog/ccp4/4.1.1/ccp4i/bin/loggraph" line 5)
The following patch corrects this:
diff -C1 -c -r1.1 -r1.1.2.1 *** loggraph.tcl 2000/11/14 17:50:44 1.1 --- loggraph.tcl 2001/03/16 08:48:10 1.1.2.1 *************** *** 43,44 **** --- 43,45 ---- source [file join $env(CCP4I_TOP) src system.tcl] + source [file join $env(CCP4I_TOP) src utils.tcl]
Alternatively, ccp4i should be run first.
The Install New Task option has some problems:
Download a fixed version of $CCP4I_TOP/src/database.tcl from the prerelease area.
Please note that this an update on the last fix to database.tcl, which was dated 27/3/2001.
This is an extremely undesirable feature of the interface which may manifest itself under certain special circumstances - if a user accidentally assigns the output file name for a task to be an existing directory (for example, by clicking on OK in the file browser without selecting a file), and then tries to run the task, then CCP4i will give them the option of deleting the existing ``file''. Clicking on OK at this point will result in the deletion of both the directory and its contents.
Note that the directory is only deleted if the user selects OK - if ``Abort'' is chosen instead then nothing happens.
A patched version of the file $CCP4I_TOP/src/runjob.tcl can be obtained from the prerelease area. Using this version CCP4i should warn when the output ``file'' is actually a directory, and will not give the option to delete it but simply abort the run.
(Thanks to Nick Keep and David Knight for reporting this.)
loggraph cannot open a file selected from the ``Open File'' option in the file menu, and the default directory alias is always blank.
Download a fixed $CCP4I_TOP/loggraph/loggraph.tcl from the prerelease area.
translation function with known solutions will misorder the models. The following corrects this fault in $CCP4I_TOP/utils/amore_utils.tcl:
diff -c -r1.5 amore_utils.tcl
*** amore_utils.tcl 2001/03/13 16:20:05 1.5
--- amore_utils.tcl 2001/04/30 08:31:12
***************
*** 353,359 ****
set sol ""
foreach im $imodel_list {
incr nl
! append sol "SOLUTION $fix $im [lrange [lindex $input_list $nl] 2 end]\n"
}
if { $solution0 != "" } {
foreach s0 $solution0 {
--- 353,360 ----
set sol ""
foreach im $imodel_list {
incr nl
! set mc [ lindex [lindex $input_list $nl] 1 ]
! append sol "SOLUTION $fix [lindex $imodel_list [expr $mc -1] ] [lrange [lindex $input_list $nl] 2 end]\n"
}
if { $solution0 != "" } {
foreach s0 $solution0 {
typo in $CCP4I_TOP/scripts/solomon.script effects the sigmaa component.
diff -r1.3 -r1.2 solomon.script 140c140 < set FP $FO --- > set FP $F0(thanks to Yen-Chywan Liaw)
Several problems with the contact interface. The interchain option without
ranges, and the intermolecular option both fail. The contact.tcl and
contact.com in the
prerelease
area fix these.
(thanks to Dr. Nick Keep)
When running refinement with ARP/wARP under CCP4i, the job may fail with a message from ARP such as:
The program run with command: arp_warp XYZIN /disk3/cirilli/PROJECTS/ABAD/CCP4/6nov.pdb MAPIN1 /tmp/abad_test_1_5_mapin_1.tmp MAPIN2 /tmp/abad_test_1_6_mapin_2.tmp XYZOUT /disk3/cirilli/PROJECTS/ABAD/CCP4/6nov_arp1.pdb has failed with error message child process exited abnormally
You need to check which version of ARP/wARP you are running - the latest version (5.1) from the EMBL is not compatible with CCP4i, which requires an older version (5) which is distributed with CCP4 4.1.*.
If the interface is using the `wrong' version of ARP then it has to be configured to pick up the correct version as follows:
When running lsqkab through ccp4i (superpose) the requested rms table will not display under loggraph. This is due to a double processing of the output table, which results in two headers. A quick fix is to remove the second header. A version of superpose.script which fixes this problem is on the prerelease area.
Some of the names in the menu of library fragments (in fact, the last three in the menu: "theoretical-10-helix", "theoretical-5-strand" and "theoretical-10-strand") are mixed up, and if selected will use a different fragment when the program actually runs.
The fix is to download a new version of the fffear.tcl file from the prerelease area to replace the existing version in $CCP4/ccp4i/tasks/.
Errors in two files cause the child script to crash. This manifests itself in the refmac job keeping the running status even after finishing. The map output file will also have spurious lines in it.
The fix is to download new versions of refmac_maps.script (replaces existing version in $CCP4/ccp4i/scripts) and map_utils.tcl (replaces version in $CCP4/ccp4i/utils). Both are available from the prerelease area. Or, to apply the following:
*** refmac_maps.script 2000/11/14 18:25:48 1.2
--- refmac_maps.script 2002/02/19 13:52:03 1.4
***************
*** 38,42 ****
foreach line $hkl_in {
append hkl_out [format "%5i%5i%5i%12.1f%12.1f%12.1f" \
[lindex $line 0 ] [lindex $line 1 ] [lindex $line 2 ] \
! [lindex $line 3 ] 1.0 [lindex $line 4]] \n
}
--- 38,43 ----
foreach line $hkl_in {
+ catch {
append hkl_out [format "%5i%5i%5i%12.1f%12.1f%12.1f" \
[lindex $line 0 ] [lindex $line 1 ] [lindex $line 2 ] \
! [lindex $line 3 ] 1.0 [lindex $line 4]] \n }
}
*** map_utils.tcl 2001/03/13 16:20:06 1.6
--- map_utils.tcl 2002/02/19 13:34:15 1.7
***************
*** 165,167 ****
set n -1; foreach lab $xtal_labin { incr n
! append text " [subst $lab]=[lindex $labin $n]"
}
--- 165,168 ----
set n -1; foreach lab $xtal_labin { incr n
! if { $n < [llength $labin] } {
! append text " [subst $lab]=[lindex $labin $n]" }
}
at stage 2.3 we're told to de-select the radio button:
keep the input intensities in the output MTZ file
but I found that I also needed to de-select the button:
Ensure unique data & add FreeR column for 0.05 fraction of data.
(if ive understood it correctly if you do choose to select this then you can
skip step 2b! But I never said that!)
The html fix will be in the next release. Alun Ashton 30/04/2001
The problem can be avoided by updating the timestamps on a number of files before running configure, as follows:
> cd $CCP4/x-windows/xdl_view/src > touch configure.in && sleep 1 > touch Makefile.am && sleep 1 > touch aclocal.m4 && sleep 1 > touch configure && sleep 1 > touch Makefile.in && sleep 1Unfortunately if the installation fails and the files are corrupted then the only way to restore the correct versions is to download and unpack the x-windows files again first - for example the pub/ccp4/ccp4/ccp4-4.1/x-windows.tar.gz file on the ccp4 ftp server - and then applying the above work-around.
(Thanks to Clemens Vonrhein.)
Apply the following patch to $CCP4/x-windows/Makefile.in to obtain the correct behaviour:
diff -c -r1.23 -r1.26
*** Makefile.in 2000/12/14 14:04:07 1.23
--- Makefile.in 2001/02/02 09:48:42 1.26
***************
*** 160,166 ****
#
# install - not yet installed!
#
! install_obj = ipdisp/ipdisp ipdisp.exe hklview xdldataman xdlmapman \
rotgen mosflm
install: all
$(INSTALL_PROGRAM) $(srcdir)/ipdisp/ipdisp $(bindir)
--- 160,166 ----
#
# install - not yet installed!
#
! install_obj = ipdisp.exe hklview xdldataman xdlmapman \
rotgen mosflm
install: all
$(INSTALL_PROGRAM) $(srcdir)/ipdisp/ipdisp $(bindir)
***************
*** 187,205 ****
cd $(libjwc_f_dir) ; $(MAKE) clean ;\
cd $(mosflm_dir) ; $(MAKE) realclean
distclean: realclean
! -rm -rf Makefile config.statusempty-targets:
! for i in $(install_obj) ; do \
! test -s $$i && rm -f $$i && touch $$i || true ;\
! done
uninstall:
for i in $(install_obj); do \
rm -f $(bindir)/$$i ;\
done
cd $(xdlview_dir) ; $(MAKE) uninstall ;\
cd $(libjwc_c_dir) ; $(MAKE) uninstall ;\
cd $(libjwc_f_dir) ; $(MAKE) uninstall
test -f $(libdir)/font.bin && rm -f $(libdir)/font.bin ; \
test -f $(libdir)/nice_font.bin && rm -f $(libdir)/nice_font.bin
FORCE : ;
--- 187,217 ----
cd $(libjwc_f_dir) ; $(MAKE) clean ;\
cd $(mosflm_dir) ; $(MAKE) realclean
distclean: realclean
! -rm -rf Makefile config.status ;\
! cd $(xdlview_dir) ; $(MAKE) distclean ;\
! cd $(rotgen_dir) ; $(MAKE) distclean ;\
! cd $(libjwc_c_dir) ; $(MAKE) distclean ;\
! cd $(libjwc_f_dir) ; $(MAKE) distclean
uninstall:
for i in $(install_obj); do \
rm -f $(bindir)/$$i ;\
done
+ rm -f $(bindir)/ipdisp ;\
cd $(xdlview_dir) ; $(MAKE) uninstall ;\
cd $(libjwc_c_dir) ; $(MAKE) uninstall ;\
cd $(libjwc_f_dir) ; $(MAKE) uninstall
test -f $(libdir)/font.bin && rm -f $(libdir)/font.bin ; \
test -f $(libdir)/nice_font.bin && rm -f $(libdir)/nice_font.bin
+
+ # this leaves empty, non executable files in place of the targets which have
+ # been built, saving space in the build directory, but enabling re-building
+ # contingent on the empty files being out-of-date with respect to their
+ # dependencies.
+
+ empty-targets :
+ -for i in $(install_obj); do \
+ test -s $$i && rm -f $$i && touch $$i || true ;\
+ done
FORCE : ;
(Thanks to Clemens Vonrhein, Katherine McAuley and Nobuhisa Watanabe.)
Apply the following patch to $CCP4/configure to obtain the correct behaviour.
cvs diff -r1.125 configure < # $Id: configure,v 1.125 2001/01/26 15:30:46 pjx Exp $ --- > # $Id: configure,v 1.126 2001/01/31 14:41:52 pjx Exp $ 771c771 < filec_CFLAGS='$XFFLAGS $FOPTIM -Dsgi' \ --- > filec_CFLAGS='$XCFLAGS $FOPTIM -Dsgi' \ 1279c1279 < filec_CFLAGS='$XFFLAGS $FOPTIM -Dlinux' "} --- > filec_CFLAGS='$XCFLAGS $FOPTIM -Dlinux' "} 1312c1312 < filec_CFLAGS='$XFFLAGS $FOPTIM -Dlinux' \ --- > filec_CFLAGS='$XCFLAGS $FOPTIM -Dlinux' \
(Thanks to Pete Dunten)
(Thanks to Ethan Merritt)
The fix to $CPROG/ccp4mapwish_/c_scr/ccp4mapwish.c is:
diff -c -r1.2 ccp4mapwish.c
*** ccp4mapwish.c 2001/02/19 08:01:24 1.2
--- ccp4mapwish.c 2001/02/20 08:47:53
***************
*** 30,35 ****
--- 30,37 ----
Must be version 8.0 or better */
#include <tcl.h>
+ #include <tk.h>
+
/* Application-specific definitions and prototypes */
#include /* CCP4 machine dep. include file */
#include "ccp4mapwish.h"
Note that further fixes need to be applied for run-time bugs (see
below).
If both BLAS and LAPACK libraries are needed then the BLAS library must be specified after LAPACK, since the loader only reads symbols once. To fix this, make the following change to $CCP4/configure:
***************
*** 1812,1818 ****
cd ../..
#
lapackdir=lib/lapack
! XLAPACK_LIB="-L$BLASDIR -l$BLASLIB -L$srcdir/lib/lapack -llapack"
fi
# Clean up
if test -f testlink.f || test -f testlink ; then
--- 1812,1818 ----
cd ../..
#
lapackdir=lib/lapack
! XLAPACK_LIB="-L$srcdir/lib/lapack -llapack -L$BLASDIR -l$BLASLIB"
fi
# Clean up
if test -f testlink.f || test -f testlink ; then
Apply the following patch to correct the call
diff -c sfch_graph_2.f sfch_graph_2.f_old
*** sfch_graph_2.f Mon Feb 5 17:18:17 2001
--- sfch_graph_2.f_old Mon Jan 29 15:20:12 2001
***************
*** 396,402 ****
CALL LSCALE (XR,YH6,-BMH,-TLEN,3,0.0,MAXDENSID,.FALSE.)
CALL LSCALE (XR,YL7,BMH,-TLEN,2,0.0,MAXBFACT,.FALSE.)
CALL LSCALE (XR,YH8,-BMH,-TLEN,2,0.0,MAXBFACT,.FALSE.)
! CALL LSCALE (XR,YL9,BMH,-TLEN,2,0.0,1.0,.FALSE.)
IF ((XCHSTRT .GT. 0).AND.(XCHSTRT.NE.XR)) THEN
CALL CHAINBOX (XCHSTRT,XR,YL,CHAINID)
ENDIF
--- 396,402 ----
CALL LSCALE (XR,YH6,-BMH,-TLEN,3,0.0,MAXDENSID,.FALSE.)
CALL LSCALE (XR,YL7,BMH,-TLEN,2,0.0,MAXBFACT,.FALSE.)
CALL LSCALE (XR,YH8,-BMH,-TLEN,2,0.0,MAXBFACT,.FALSE.)
! CALL LSCALE (XR,YL9,BMH,-TLEN,2,1.0,.FALSE.)
IF ((XCHSTRT .GT. 0).AND.(XCHSTRT.NE.XR)) THEN
CALL CHAINBOX (XCHSTRT,XR,YL,CHAINID)
ENDIF
inwrite: off end of record *** Execution terminated (110)
The problem is in the file $CPROG/molrep_/main_molrep_mtz.f - apply the following patch to fix:
diff -r1.5 main_molrep_mtz.f 215c215 < WRITE(LINE,*) '--- MTZ file : ',FILE_MTZ(1:L) --- > WRITE(LINE,'(A,A)') '--- MTZ file : ',FILE_MTZ(1:L)
Apply the following patches
diff -c molrep_prog.f ccp4-4.1/src/molrep_/molrep_prog.f
*** molrep_prog.f Mon Feb 19 11:11:16 2001
--- ccp4-4.1/src/molrep_/molrep_prog.f Mon Jan 29 15:18:39 2001
***************
*** 3283,3289 ****
C ----------------------------------------
NAME=NAMED
IF(NSYM.GT.1) THEN
- MTF = 43
NAME='molrp_scratch_rdc'
PATH=' '
EXT='dat'
--- 3283,3288 ----
diff -c molrep.f ccp4-4.1/src/molrep_/molrep.f
*** molrep.f Mon Feb 19 11:18:45 2001
--- ccp4-4.1/src/molrep_/molrep.f Wed Feb 7 16:54:52 2001
***************
*** 5295,5301 ****
C ---
NP = RESULTA(1,12)
IF(NP.LE.0.0) GO TO 900
! IOUT = 11
NAME = 'molrep_srf.tab'
PATH = ' '
EXT = ' '
--- 5295,5301 ----
C ---
NP = RESULTA(1,12)
IF(NP.LE.0.0) GO TO 900
!
NAME = 'molrep_srf.tab'
PATH = ' '
EXT = ' '
(Thanks to Arthur J. Chirino)
A version of AREAIMOL which automatically searches +/-2 lattice vectors is available on the prerelease area, as areaimol-080201.f This updated version should work in these situations, and find all the buried area correctly.
(Thanks to David Buckler)
diff -r1.6 oasis.f 508c508 < IFLAG(I) = 1 --- > IFLAG(J) = 1
diff -c -r1.3 solvent_contr_mask.f
*** solvent_contr_mask.f 2001/01/22 14:32:42 1.3
--- solvent_contr_mask.f 2001/02/15 17:11:20
***************
*** 412,418 ****
NSEC = N3
IUN_MAP = 0
TITLE = 'Mask from refmac'
! CALL MWRHDL(IUN_MAP,'MSKOUT',,TITLE,NSEC,IUVW,MXYZ,
& NW1,NU1,NU2,NV1,NV2,CELL_LOCAL,LSPGRP,LMODE)
C
MU = N1
--- 412,418 ----
NSEC = N3
IUN_MAP = 0
TITLE = 'Mask from refmac'
! CALL MWRHDL(IUN_MAP,'MSKOUT',TITLE,NSEC,IUVW,MXYZ,
& NW1,NU1,NU2,NV1,NV2,CELL_LOCAL,LSPGRP,LMODE)
C
MU = N1
diff -c -r1.2 tls_newdic.f
*** tls_newdic.f 2001/01/22 14:38:22 1.2
--- tls_newdic.f 2001/02/15 17:13:47
***************
*** 371,376 ****
--- 371,377 ----
INTEGER NPARS
PARAMETER (NPARS = 200)
+ CHARACTER OLINE*60
c---Things for parser
CHARACTER LINE*600,KEY*4,CVALUE(NPARS)*4
This requires a fix to $CPROG/ccp4mapwish_/c_src/ccp4mapwish.c:
diff -r1.1 -r1.2 141c141 < int i,j,k, lenname, ifail=0, iunit, nsize; --- > int i,j,k, isec, lenname, ifail=0, iunit, nsize; 270a271 > /* Section numbering may not start at zero... */ 272a274 > /* Set ifail = 1 to switch on debug mode */ 274c276,280 < getsection_(&iunit, &i, plane = data, &nsize, &ifail); --- > /* Section numbering in the map file can start from > an arbitrary value > Start section is stored in nw1 */ > isec = i + nw1; > getsection_(&iunit, &isec, plane = data, &nsize, &ifail); 691c697 < section->ulimits[1] = start + map->ext[iu] - 1; --- > section->ulimits[1] = map->start[iu] + map->ext[iu] - 1; 693c699 < section->vlimits[1] = start + map->ext[iv] - 1; --- > section->vlimits[1] = map->start[iv] + map->ext[iv] - 1; 696c702,706 < Once again must think about different possible permutations */ --- > this_sec points to the "absolute" section number from the mapfile > This must be reset relative to the section numbers stored in memory */ > this_sec = this_sec - start; > > /* Once again must think about different possible permutations */ 1782c1792,1793 < if ((xlim1 > 1 && xlim2 > 1) && (ylim1 > 1) && (ylim2 > 1)) --- > if ((fabs(xlim1) > 1.0 && fabs(xlim2) > 1.0) > && (fabs(ylim1) > 1.0) && (fabs(ylim2) > 1.0))
The following is a simple fix.
diff -c distang.f ccp4-4.1/src/distang.f
*** distang.f Mon Feb 19 16:23:20 2001
--- ccp4-4.1/src/distang.f Mon Jan 29 15:17:07 2001
***************
*** 229,237 ****
C **********************
CALL RES3TO1(RESNAM,RESNM1)
C **********************
! C add extra check to recognise water labelled HOH and WAT
! IF( RESNAM(1:3) .EQ. 'HOH' .OR.
! + RESNAM(1:3) .EQ. 'WAT' ) RESNM1 ='O'
C
WRITE (NAME(NAT),FMT='(a1,i4,a4,a1)') RESNM1,IRESCD,ATNAM,CHNNAM
C
--- 229,236 ----
C **********************
CALL RES3TO1(RESNAM,RESNM1)
C **********************
! C add extra check to recognise water labelled HOH
! IF( RESNAM(1:3) .EQ. 'HOH') RESNM1 ='O'
C
WRITE (NAME(NAT),FMT='(a1,i4,a4,a1)') RESNM1,IRESCD,ATNAM,CHNNAM
C
The solution is to use formats that reflect the size of the real.
diff -c revise.f ccp4-4.1/src/revise.f
*** revise.f Thu Feb 22 18:59:46 2001
--- ccp4-4.1/src/revise.f Thu Feb 22 18:56:54 2001
***************
*** 1436,1448 ****
+ ' :Ratio at Whole Resolutions of data ',I2,' over ',I2,
+ ':N:1,12:',/,' $$ Ratio ')
DO 1210 NSS = 1, NRMAX
! IF(RES(NSS).GE.100.0) THEN
! WRITE(LUNOUT,'(A,F6.2)') ' Resolution_at_range_',RES(NSS)
! ELSE IF(RES(NSS).GE.10.0) THEN
! WRITE(LUNOUT,'(A,F5.2)') ' Resolution_at_range_',RES(NSS)
! ELSE
! WRITE(LUNOUT,'(A,F4.2)') ' Resolution_at_range_',RES(NSS)
! END IF
1210 CONTINUE
WRITE(LUNOUT,FMT='(A)') ' Ratio_at_Whole_Resolution $$'
WRITE(LUNOUT,1220) (NSS,NSS=1,NRMAX)
--- 1436,1443 ----
+ ' :Ratio at Whole Resolutions of data ',I2,' over ',I2,
+ ':N:1,12:',/,' $$ Ratio ')
DO 1210 NSS = 1, NRMAX
! WRITE(LUNOUT,1205) RES(NSS)
! 1205 FORMAT(' Resolution_at_range_',F5.2)
1210 CONTINUE
WRITE(LUNOUT,FMT='(A)') ' Ratio_at_Whole_Resolution $$'
WRITE(LUNOUT,1220) (NSS,NSS=1,NRMAX)
(Thanks to Eleanor J. Dodson)
diff -c -r1.97 mlphare.f *** mlphare.f 2001/02/22 11:12:44 1.97 --- mlphare.f 2001/02/23 16:23:09 *************** *** 4679,4685 **** C a = b*c C C ************************** ! CALL MATMUL(BANGST,AMAT,BANGST1) C ************************** C C---- bangst(2) = [amat]*[bangst(1(TRANSPOSE)] --- 4679,4685 ---- C a = b*c C C ************************** ! CALL MATMUL(BANGST1,AMAT,BANGST) C ************************** C C---- bangst(2) = [amat]*[bangst(1(TRANSPOSE)]
There is an additional omission from the $CEXAM/tutorial2000/results directory: the file model_molrep1.pdb is unintentionally missing.
