This summer I am working on the Google Summer of Code project “Revive GNU Prolog for Java”. It now has a project page and a Git repository which resulted in a rather entertaining screen shot.

Yesterday I found out that last year someone did a lot of the work I was intending to do this summer and they gave me a svn dump of the changes that they made. I had previously found two other people that had made some changes over the last 10 years while the project was dormant.

So I was faced with the task of taking the SVN dump of the changes made by Michiel Hendriks (elmuerte) and splicing them onto the old CVS history of the code he took the source .zip and then converting it into Git (which is what we are now using as our VCS). I was kind of hoping that with luck the two development histories would then share a common root which could be used to help with merging the two development histories back together that hope was in vain (though I have another idea I might try later).

Anyway this whole splicing thing was non-trivial so I thought I would document how I did it (partly so that I can find the instructions later).

So there were problems with the SVN dumpfile I was given: it couldn’t be applied to a bare repository as all the Node-Paths had an additional extension ‘trese/ample/trunk’ on the beginning. No one fix I tried was to add

Node-path: trese
Node-kind: dir
Node-action: add

Node-path: trese/ample
Node-kind: dir
Node-action: add

Node-path: trese/ample/trunk
Node-kind: dir
Node-action: add

Into the the first commit using vim: now this worked but it meant that I still had the extra ‘trese/ample/trunk/’ which I didn’t want.
So I removed that using vim and “:%s/trese\/ample\/trunk\///g”, unfortunately there were a couple of instances where trese/ample/trunk was refereed to directly when files were being added to svn:ignore unfortunately I didn’t find out how to refer to the top level directory in an svn dump file so I just edited those bits out (there were only 2 commits which were effected). So now I had a working svn dumpfile. To do the splicing I used svndumptool.py to remove the first commit resulting in a dumpfile I called gnuprolog-mod.mod2-86.svn.dump (see below for instructions on how to do this).

I got a copy of the CVS repository, so that the current working directory contained CVSROOT and also a folder called gnuprolog which contained the VCed code.

# This makes a dumpfile 'svndump' of the code in the gnuprolog module I only care about the trunk.
cvs2svn --trunk-only --dumpfile=svndump gnuprolog
# Then we use svndumptool to remove the first commit as cvs2svn adds one to the beginning where it makes various directories.
svndumptool.py split svndump 2 8 cvs-1.svn.dump
# Then we use vim to edit the dump file and do :%s/trunk\///g to strip of the leading 'trunk/' from Node-Paths
vim cvs-1.svn.dump
# Create a SVN repository to import into
svnadmin create gnuprolog-mod.plaster.svn
# Import the CVS history
svnadmin load gnuprolog-mod.plaster.svn < cvs-1.svn.dump
# Import the SVN history
svnadmin load gnuprolog-mod.plaster.svn < gnuprolog-mod.mod2-86.svn.dump
# Make a git repository from the SVN repository
git svn clone file:///home/daniel/dev/gnuprolog/gnuprolog-mod.plaster.svn gnuprolog-mod.plaster.git

Things that I found which are useful

The SVN dump file format
svn-to-git got me the closest to being able to import the svn dumpfile into git.
These instructions on how to fix svn dumpfiles.

Sorry this post is rather ramblely stuck half way between a howto and an anecdote but hopefully someone will find it useful.

Today I received a call at about 10:05 to my home landline. I rapidly realised it was some kind of computer based scam and decided to have some fun seeing what they would try and do.
I had great fun doing this but I think that someone who does not understand computer could have easily been taken in.

As in many such scams they claimed to be a company working for Microsoft and offering this free service of finding out what is wrong with my computer as they detected that it was downloading lots of junk files from the internet which were slowing it down. Now our old Windows XP desktop is indeed old and slow and this is quite possibly due to junk. However it was obvious that they were making all this up. So they wanted me to turn my computer on – now obviously I wasn’t going to risk following any instructions on the real computer so I booted my XP VM on my laptop instead (which I will subsequently need to wipe).

Having booted the XP VM and possibly being passed onto a different call centre person. I was given a series of instructions the purpose of which was to prove that the computer had a problem. This involved going to the event viewer in computer administration (Start -> right click on “My Computer” -> Manage -> Event viewer and then to both Application and System. With a little sorting for effect we get a screen something like the following:
I suppose many people might find that quite scary but I have previously looked at such screens and it was what I expected to see.

Having ‘proved’ that there was something wrong with my computer they then proceeded to try and get me to provide greater access to them. This was done by getting me to visit www.logmein123.com and use the code 807932 (which they really didn’t want me to reveal to anyone).

They then got remote access to my computer and went and installed a fake scanner from http://majorgeeks.com/Advanced_WindowsCare_v2_Personal_d4991.html This proceeded to produce some fake results:

They then wanted to see if my “software warranty” had expired as this would be why my computer was “downloading junk files which can’t be removed by anti-virus”.
This was done by opening cmd and doing

cd \
tree

and while tree was running typing “expired.” so that it would appear at the bottom.

At this point they went in for the kill and opened up a form and claimed that “it is a timed http form so we can’t look at it” and that it would “automatically go in 8 minutes so you need to fill it in quickly”.
Obviously I wasn’t going to fill this form in so at that point I revealed that I knew that they were scammers. They denied this and got progressively more angry and incoherent and when I asked to be put through to their supervisor they hung up.

Follow up

Now obviously it is my duty to try and prevent this kind of thing from happening again.
So my first step was to try to find out the number which was used to call me using 1471 but unfortunately this did not work. I then tried the local police but they could not be of any help and they advised me to contact BT unfortunately BT could not help either as it was an international call with no number given.
I then reported relevant URLs to google and the exe to Stop Badware.

I contacted the company behind logmein123.com which seems to be a legitimate company telling them that their services are being abused and requesting comment from them about this. I received a very positive response: “Thanks for the heads up on this. We take this stuff very seriously and will investigate immediately. Any misuse of the product or trials for the purpose you describe is a violation of our terms and immediate grounds for termination of the service. Thank you for sending the PIN as it helps us not only track this down to end their service, it also gives us information we need should we decide to press legal action. …”

Now looking to see whether anyone else has discovered this gogreenpc.net scam I found that they have. So gogreenpc.net is a big scam site. Now I need to work out how to take them out. :-D

The people on #cl on irc.srcf.ucam.org were helpful in providing advice on follow up.

It is Ada Lovelace day today.

I would like to thank Frances Zhang for putting up with me in supervisions this year and for being really hard working and understanding things I don’t (e.g. probability) and not boasting about it as I tend to (sorry).

Dina Bester is generally awesome, has supplied pizza for many hungry CompScis and was willing to take on project Bravo which I thought looked pretty impossible and got a good project out at the end.

Sarah Gordon for doing a Women@CL talk which showed how testing can be exciting and be just as complicated and technical as writing the code that runs the project while requiring a greater understanding of how the whole project fits together.

I also think that Ada Lovelace day is a good idea and I am glad that my operating system (Ubuntu) is supporting it (or at least the Ubuntu Planet is). Making the internet a better place, one step at a time :-).

The first programmer was a women (Ada), the first programmers (ENIAC) were women, it is sad that the proportion of women in Computer Science now is as low as it is. Especially since they get proportionally more 1sts in CompSci.

I thought I had found a concurrency bug in PostgreSQL because a ResultSet that was local to a method was being closed by activity occurring in a different thread.

Then I read: “Once you make another query with the Statement used to create a ResultSet, the currently open ResultSet instance is closed automatically.” on http://jdbc.postgresql.org/documentation/84/resultset.html

Which means that I will have to make my code slightly less concurrent to work around that :-(.

I suspect it prevents bad coders from causing resource leaks though. So it is probably a good design decision.

This is why your unit tests should check for concurrency flaws :-).

I was surprised to find that the co-operative bank’s policy is not as evil as “Security Engineering” would suggest bank’s policies are. Specifically:

“We will repay you any money that is taken from your account due to: any error by our staff or our systems, a computer crime which is not found and stopped by our security system.”

Whereas “Security Engineering” suggests that in general UK banks say ‘you are an evil criminal’ if a computer crime against your account succeeds.
Halifax says:

“If a customer of our online service is a victim of online fraud, we guarantee that they won’t lose any money from their account and will always be reimbursed in full.”

but I suppose the “our system is secure and so online fraud is not possible so you are a criminal” trick might work there…
Possibly this means that banks policies are improving as they realise that tackling fraud is their responsibility. (Perhaps they read the book which is very good).

From the point of view of login security the co-operative would give me a chip and pin card reader to verify online transactions which gives better security than Halifax’s username + password + some random fact that would be very easy to find out using something like facebook. (though there are flaws in such a chip and pin system detailed in the book).

Only 5 chapters left to read… :-)

I have obtained the source code for JULES which is one of the components that the Met Office uses to forecast the weather. Compiling it was not straightforward and required the modification of several source files.

The modified files and .patch files can be found here.

First I will outline the route I took to the most working binary.

An almost working way

1) in Makefile:

1.1) Setting the location of the netcdf libaries (you will probably need to install these)

JULESDIR=$(PWD)

CDF_LIB_PATH=/usr/lib/

CDF_MOD_PATH=/usr/include/

1.2) and to enable the use of the gfortran compiler

ifeq ($(COMPILER),g95)

include $(JULESDIR)/Makefile.comp.g95

endif

ifeq ($(COMPILER),gfortran)

include $(JULESDIR)/Makefile.comp.gfortran

endif

ifeq ($(COMPILER),nag)

include $(JULESDIR)/Makefile.comp.nag

endif

rather than

ifeq ($(COMPILER),g95)

include $(JULESDIR)/Makefile.comp.g95

endif

ifeq ($(COMPILER),nag)

include $(JULESDIR)/Makefile.comp.nag

endif

2) in Makefile.common.mk:

GRMPATH=$(which grm)

ifneq ($(strip $(GRMPATH)),)
    RM=grm -f
endif

RMPATH=$(which rm)

ifneq ($(strip $(RMPATH)),)
    RM=rm -f
endif

ifndef RM
    $(error No rm equivalent command found)
endif

rather than

RM=grm -f

grm means GNU rm, on a linux system rm is GNU rm, so rm should be used rather than grm the conditionals mean that if the met office run the code on their setup it ‘should work’ and if someone runs it on a normal linux distribution it should still work.

cp Makefile.comp.g95 Makefile.comp.gfortran

3) in Makefile.comp.gfortran:

In order to select the gfortran compiler rather than g95, as gfortran is available for Ubutnu/debian while g95 is not.

FC=gfortran -fno-underscoring

rather than:

FC=g95

and

MOD_PUT=-J

rather than:

MOD_PUT=-fmod=

4) Then to compile

make BUILD=run COMPILER=gfortran CDFDUMMY=true

#BUILD could also be debug or fast.
#CDFDUMMY is needed even though you can install netcdf from the repositories – because if you set it false you get:

/home/daniel/dev/JULES/jules-v2-0/libjules.a(RWERR_MOD.o): In function `__rwerr_mod_MOD_rwerr':
RWERR_MOD.f90:(.text+0xaae): undefined reference to `__netcdf_MOD_nf90_inquire_dimension'
RWERR_MOD.f90:(.text+0xed7): undefined reference to `__netcdf_MOD_nf90_strerror'
/home/daniel/dev/JULES/jules-v2-0/libjules.a(JULES_NETCDF.o): In function `__jules_netcdf_MOD_closecdf':
JULES_NETCDF.f90:(.text+0xd): undefined reference to `__netcdf_MOD_nf90_close'
/home/daniel/dev/JULES/jules-v2-0/libjules.a(JULES_NETCDF.o): In function `__jules_netcdf_MOD_check_nc_dims':
JULES_NETCDF.f90:(.text+0x2e1): undefined reference to `__netcdf_MOD_nf90_inq_dimid'
/home/daniel/dev/JULES/jules-v2-0/libjules.a(JULES_NETCDF.o): In function `__jules_netcdf_MOD_readvar2dreal_ncvector_gswp2':
JULES_NETCDF.f90:(.text+0x566): undefined reference to `__netcdf_MOD_nf90_inq_varid'
JULES_NETCDF.f90:(.text+0x8b1): undefined reference to `__netcdf_MOD_nf90_get_var_1d_fourbytereal'
JULES_NETCDF.f90:(.text+0xbb9): undefined reference to `__netcdf_MOD_nf90_get_var_1d_fourbytereal'
JULES_NETCDF.f90:(.text+0xeed): undefined reference to `__netcdf_MOD_nf90_get_var_1d_fourbytereal'
JULES_NETCDF.f90:(.text+0x1160): undefined reference to `__netcdf_MOD_nf90_get_var_1d_fourbytereal'
JULES_NETCDF.f90:(.text+0x1477): undefined reference to `__netcdf_MOD_nf90_get_var_1d_fourbytereal'
/home/daniel/dev/JULES/jules-v2-0/libjules.a(JULES_NETCDF.o):JULES_NETCDF.f90:(.text+0x178e): more undefined references to `__netcdf_MOD_nf90_get_var_1d_fourbytereal' follow
/home/daniel/dev/JULES/jules-v2-0/libjules.a(JULES_NETCDF.o): In function `__jules_netcdf_MOD_opencdf':
JULES_NETCDF.f90:(.text+0x1c24): undefined reference to `__netcdf_MOD_nf90_open'
collect2: ld returned 1 exit status
make: *** [jules.exe] Error 1

Successful compilation should result in:

gfortran -fno-underscoring -o jules.exe /home/daniel/dev/JULES/jules-v2-0/JULES.o \
	    /home/daniel/dev/JULES/jules-v2-0/UTILS/netcdf_dummy/JULES_NETCDF_DUMMY.o \
	    -L/home/daniel/dev/JULES/jules-v2-0  -L/home/daniel/dev/JULES/jules-v2-0/UTILS/netcdf_dummy  \
	    -J/home/daniel/dev/JULES/jules-v2-0/MODS -I/home/daniel/dev/JULES/jules-v2-0/MODS -I/home/daniel/dev/JULES/jules-v2-0/UTILS/netcdf_dummy  \
	    -ljules

Alternatively try to use g95 as recommended

1) Install g95

From here: http://www.gfd-dennou.org/library/cc-env/g95/index.htm.en#label-5
g95 is not available in either Ubuntu or Debian repositories – for unknown reasons.

2) follow steps 1.1 and 2 from the instructions for gfortran.

3) try to compile

make BUILD=run COMPILER=g95 CDFDUMMY=true

If you do this there is no possibility of using netcdf prebuilt binaries according to: http://www.unidata.ucar.edu/support/help/MailArchives/netcdf/msg04125.html
This is because they were built for gfortran and so will only work with gfortran.

unfortunately this results in:

g95 -o jules.exe /home/daniel/dev/JULES/jules-v2-0/JULES.o \
	    /home/daniel/dev/JULES/jules-v2-0/UTILS/netcdf_dummy/JULES_NETCDF_DUMMY.o \
	    -L/home/daniel/dev/JULES/jules-v2-0  -L/home/daniel/dev/JULES/jules-v2-0/UTILS/netcdf_dummy  \
	    -fmod=/home/daniel/dev/JULES/jules-v2-0/MODS -I/home/daniel/dev/JULES/jules-v2-0/MODS -I/home/daniel/dev/JULES/jules-v2-0/UTILS/netcdf_dummy  \
	    -ljules
/home/daniel/dev/JULES/jules-v2-0/libjules.a(SOILHY7A.o):(.data+0x0): undefined reference to `darcy_'
/home/daniel/dev/JULES/jules-v2-0/libjules.a(SOILHY7A.o):(.data+0x4): undefined reference to `hyd_con__'
make: *** [jules.exe] Error 1

Running Jules

To run jules you will need to

mkdir OUTPUT

after successfully compiling

and then to run a test run using one of the shipped examples:

./jules.exe < point_loobos_example.jin

unfortunately doing this I got:

At line 152 of file SFSNOW7A.f

Fortran runtime error: Array reference out of bounds for array 'snowcanpft', upper bound of dimension 1 exceeded (6 > 5)

Also trying:

./jules.exe < grid_gswp2_example.jin

fails because:

Opening /users/global/rjel/jules/gswp2_files/lsmask_vector.nc

 fileFormat=nc

 ERROR: attempting to use a netCDF procedure, but netCDF library not available.

 This is a 'dummy' procedure.

 To use netCDF, remake model with the 'proper' netCDF library.

as we compiled with CDFDUMMY=true …

Conclusion

It is possible to compile JULES on Ubuntu Linux with a bit of work. However compiling a working version of JULES on Ubuntu Linux may be beyond my current ability (I don’t know any FORTRAN).

Hopefully this will help someone.

Thanks as always go to the denizens of #srcf on irc.srcf.ucam.org

This was harder than expected and I did not find any instructions on the internet – the manual from the CASIO website was useless.

The battery required was a CR 2016 which I got from maplin.

Getting the battery out and putting a new one in was not too hard – taking the back off required a small philips screwdriver and then the central section was levered out and the inner metal cover removed by unclipping it. Putting the new battery in was also similarly easy.

However this did not make the watch work – it was still blank and needed reseting – though the light worked. I could find no way of fixing it and so in the end I held down the light button and poked the copper contacts inside with a pin connecting them to the outer metal case in the hope of flipping some switch somewhere and so making it work – to my surprise this worked.

Note: that there is a little spring inside the case and so be very careful when turning the inside section upside down to keep your finger over the spring to stop it falling out.

I realise that this is definitely a namespace clash but hey.
After a few days work and significant help from #cl on the srcf irc my fractals program is ready for 1.0 release – at least in my eyes there are no known bugs – save that there may be more efficient algorithms to use especially from the point of view of quality of display – see here for an example of tc’s better version of carpet.

The ML code is released under GPL version 2 and can be found here

To compile it you will need a version of Moscow ML with its libraries – e.g. NOT the one shipped by Ubuntu which does not have the libraries packaged – hopefully this will be fixed in Jaunty – I used the version on the linux pwf machines at Cambridge University. All other instructions on compilation are included as a comment in the file.

Examples of output can be found here

This is an extension of ML Tick6* Foundations Of Computer Science at Cambridge University

And now for the pretty pictures: