Compile error during mfold v3.5 install - can someone help?

23 replies [Last post]
ken.n
Offline
Joined: 05/09/2011

Hello,

I'm still fairly new with linux. I've managed to install mfold-util 4.5, and then I tried to install mfoldv3.5 but got the following compile errors:

****************
....
d write(6,*) 'Reduce 3'' end (free): i,j,e = ',i,j,e
1
Error: Non-numeric character in statement label at (1)
rna.f:1043.1:

d write(6,*) 'Base pair: ',i,j
1
Error: Non-numeric character in statement label at (1)
make[2]: *** [nafold-rna.o] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2
....
****************

I tried as best I could to follow the Install instructions to modify the makefile (e.g. I followed the example at http://www.cs.bgu.ac.il/~sebastia/sebastia/mfold-3.1.2/Makefile) , but to no success. But then again I don' really understand the meaning of the modifications.
I suspect it has something to do with the fortran compiling?

I'm using Linux Ubuntu 10.04.

Would someone be able to help?

Thanks,
Ken

hong_010
Offline
Joined: 04/24/2013
libgd.so error

Hi!
when I modified the c++/algorithm.cpp that contains: #include .When run to make,it errors:
/gcc/x86_64-pc-linux-gnu/4.6.3/../../../../x86_64-pc-linux-gnu/bin/ld: skipping incompatible /usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/../../../libgd.so when searching for -lgd
/gcc/x86_64-pc-linux-gnu/4.6.3/../../../../x86_64-pc-linux-gnu/bin/ld: skipping incompatible /usr/lib/libgd.so when searching for -lgd
/gcc/x86_64-pc-linux-gnu/4.6.3/../../../../x86_64-pc-linux-gnu/bin/ld: cannot find -lgd
collect2: ld returned 1 exit status
make[2]: *** [efn2] Error 1
make[2]: Leaving directory `mfold-3.5/c++'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/mfold-3.5'
make: *** [all] Error 2

Would you give me some help?

Thanks!

hong_010
Offline
Joined: 04/24/2013
make error

Hi,I installed mfold-3.5 for gentoo.It shows:
make[2]: *** [efn2.o] Error 1
make[2]: Leaving directory `mfold-3.5/c++'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `mfold-3.5'
make: *** [all] Error 2

The files efn2.cpp and algorithm.cpp are supplied with header files with extension .h i.e "#include string.h",
but it still has the error.

Can you give me some help?

Thanks!

mildldie
Offline
Joined: 11/20/2012
I also have same problem with

I also have same problem with Ubuntu 10.04. any solution?

jtcafe
Offline
Joined: 03/26/2012
problems

hi,
i can problem to install mfold-util, i have OS 64 bit, can you help me, i don't know how to rislve this issue?thx,

BrianSmiley
Offline
Joined: 03/17/2012
Working

Just got it to work for Ubuntu 10.10 :D!

zukerm
Offline
Joined: 11/12/2010
Please post exactly what you

Please post exactly what you did to compile the package for Ubuntu.

Why are you (and all the others who have posted comments) spending time installing the obsolete mfold software when the superior UNAFold package is available for download? I would like to know if there are issues with respect to installing UNAFold for Ubuntu

David Dawson
Offline
Joined: 03/02/2012
Same

Having real issues , personally I think the issue lies in Ubuntu, hopefully someone can come up with a fix/work around

tahia
Offline
Joined: 02/29/2012
Hi ken I'm using ubuntu 10.10

Hi ken

I'm using ubuntu 10.10 and have changed the algorithm.cpp

#include cstring

and it is working fine now.

Regards

Tahia

zukerm
Offline
Joined: 11/12/2010
Bug in mfold

I am very sorry that this bug has caused a problem. It is for a very stupid reason. In Fortran, any line beginning with 'c' is a comment line and is ignored when the code is compiled. A line beginning with 'd' is also treated as a comment, except that it can become a valid part of the code if a special compilation is performed. Unfortunately, some Fortran compilers are very stupid and do not recognize this convention. I know this and these lines should never have been left in the code. They are for debugging only.

To eliminate the problem, please edit the source code and remove all lines beginning with d. I will correct the download version of mfold. Thank you for reporting the problem.

ken.n
Offline
Joined: 05/09/2011
From memory, I checked and

From memory, I checked and edited all fortran related files (.f) . That got rid of that particular error.

ken.n
Offline
Joined: 05/09/2011
Unfortunately I also have

Unfortunately I also have been unable to install it in the end as I had other errors (see my other post), so I have given up on local mfold and been using the webserver instead and also the Vienna package.
I'm not sure if it's the mfold code itself or something with Ubuntu (I'm also using Ubuntu), but I don't have these sorts of problems with compiling with other software.
good luck!

zukerm
Offline
Joined: 11/12/2010
installing mfold 3.5 with Ubuntu

As I wrote when discussing problems encountered with the mfold_util package, I'm not personally using Ubuntu so all I can do is suggest solutions. I'm hoping that users will share their experiences on this forum. First and foremost, mfold is an obsolete package that is still supported for a variety of reasons: 1) intellectual property reasons, 2) ease of creating specialized versions and 3) backwards compitibility, allowing users to compare current foldings with old ones computed by mfold. I strongly urge everyone to download and install the far superior UNAFold package that supercedes and replaces mfold. You can still predict suboptimal foldings in the mfold style using UNAFold. Run hybrid-ss-min with the "--mfold" flag. The UNAFold.pl Perl script produces mfold-like output by default. You should find UNAFold much easier to install. If there are still problems with Ubuntu, then we will make an effort to solve them.

zukerm
Offline
Joined: 11/12/2010
This message is for anyone

This message is for anyone who uses UNAFold (or mfold for that matter). It is no longer necessary to install mfold_util before installing UNAFold. If the required programs from mfold_util (sir_graph and boxplot_ng are installed, the UNAFold scripts will detect and use them. Otherwise, plot files will not be created.

ken.n
Offline
Joined: 05/09/2011
new error during mfold compilation

Dear Prof. Zuker,
Thank you for the reply.

I found that rna.f and rna-quik.f both had lines beginning with 'd', which I deleted. I couldn't find this in the other .f codes.
Unfortunately when I tried compiling again, I ran into other errors which seems to be in the C++ code (see some lines below).
My steps for install were:
./configure
make

when I ran into the errors. Not sure if the undeclared functions is due to my system now.

If you like I can send you an output of the attempt via email.

Best regards
Ken

*****************************************************************
e.g. :
...
Warning: Deleted feature: Loop variable at (1) must be integer
misc-quik.f:323.9:

do a = 1,5
1
Warning: Extension: REAL array index at (1)
misc-quik.f:401.15:

do d = 1,5
1
...
algorithm.cpp:2326: error: ‘strcmp’ was not declared in this scope
...
algorithm.cpp:2617: error: ‘strcpy’ was not declared in this scope
...
algorithm.cpp:3656: warning: ignoring return value of ‘char* fgets(char*, int, FILE*)’, declared with attribute warn_unused_result
...
algorithm.cpp:4007: warning: deprecated conversion from string constant to ‘char*’
...
algorithm.cpp: In function ‘void sortstructures(structure*)’:
algorithm.cpp:4030: error: ‘strcpy’ was not declared in this scope
algorithm.cpp: In function ‘void opensav(char*, structure*, arrayclass*, arrayclass*, arrayclass*, int*, int*, int*, datatable*)’:
algorithm.cpp:4314: error: ‘strcpy’ was not declared in this scope
efn2.cpp: In function ‘void getdat(char*, char*, char*, char*, char*, char*, char*, char*, char*, char*, char*, char*, char*, char*, char*, char*, char*)’:
efn2.cpp:14: error: ‘strcpy’ was not declared in this scope
efn2.cpp:15: error: ‘strcat’ was not declared in this scope
efn2.cpp: In function ‘int main(int, char**)’:
efn2.cpp:67: error: ‘strcpy’ was not declared in this scope
efn2.cpp:70: error: ‘strcpy’ was not declared in this scope
efn2.cpp:100: error: ‘strlen’ was not declared in this scope
efn2.cpp:102: error: ‘strcmp’ was not declared in this scope
efn2.cpp:104: error: ‘strcat’ was not declared in this scope
make[2]: *** [efn2.o] Error 1
make[2]: Leaving directory `/usr/local/bin/mfold-3.5/c++'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/usr/local/bin/mfold-3.5'
make: *** [all] Error 2

*********************************************************

telukir
Offline
Joined: 08/05/2011
It worked if the files

It worked if the files efn2.cpp and algorithm.cpp are supplied with header files with extension .h i.e "#include string.h"

zukerm
Offline
Joined: 11/12/2010
Compile errors in mfold 3.5

As of today, the source code for the mfold package (version 3.5) has been updated. All comment lines beginning with "d" have been replaced by equivalent lines beginning with a "c".

I think I know what the problem is with the efn2 compilation. Give me a bit of time.

telukir
Offline
Joined: 08/05/2011
Compile errors in mfold 3.5

These are errors seems to be still there. But also header file error is also there i.e in efn2.h and algorithm.h if you include string.h header file then it got installed but however C error still coming.

jonoave
Offline
Joined: 05/11/2011
I'm using Ubuntu 10.10 and

I'm using Ubuntu 10.10 and having the same problem as well.

flamelily
Offline
Joined: 03/02/2012
Ubuntu 10.10 has been causing

Ubuntu 10.10 has been causing me similar issues. can someone advise?

zukerm
Offline
Joined: 11/12/2010
See above.

See above.