From brandon.barker at gmail.com Wed Aug 1 01:02:09 2007 From: brandon.barker at gmail.com (Brandon Barker) Date: Wed, 1 Aug 2007 01:02:09 -0400 Subject: [emboss-dev] External Libraries, C++ App Message-ID: <3e078a5e0707312202ua3af58dg4ad75657b9a81c91@mail.gmail.com> Hello, I'm currently developing a software package that I would like to try to add to EMBOSS or EMBASSY. There are two potential problems: the package I'm working on is in C++, and most likely won't be changed to C. Also, it makes heavy use (in innermost loops, unfortunately, but this can't be avoided) of arbitrary precision rationals (or integers) so I need to use an external library (GMP). All the licensing should be fine, since all of the software and libraries I'm using are under the LGPL or the GPL. The application area is parametric and inverse parametric sequence alignment (or regular alignment with arbitrary precision parameters, but I don't think too many people are interested in that aspect). My question is: should I shoot for acceptance in to EMBOSS, EMBASSY, or neither? I would also like to know if someone is already working on a parametric alignment application for EMBOSS =). Thanks, -- Brandon Barker From pmr at ebi.ac.uk Wed Aug 1 04:10:59 2007 From: pmr at ebi.ac.uk (Peter Rice) Date: Wed, 01 Aug 2007 09:10:59 +0100 Subject: [emboss-dev] External Libraries, C++ App In-Reply-To: <3e078a5e0707312202ua3af58dg4ad75657b9a81c91@mail.gmail.com> References: <3e078a5e0707312202ua3af58dg4ad75657b9a81c91@mail.gmail.com> Message-ID: <46B04013.5050009@ebi.ac.uk> Brandon Barker wrote: > Hello, I'm currently developing a software package that I would like to try > to add to EMBOSS or EMBASSY. There are two potential problems: the package > I'm working on is in C++, and most likely won't be changed to C. Also, it > makes heavy use (in innermost loops, unfortunately, but this can't be > avoided) of arbitrary precision rationals (or integers) so I need to use an > external library (GMP). All the licensing should be fine, since all of the > software and libraries I'm using are under the LGPL or the GPL. The > application area is parametric and inverse parametric sequence alignment (or > regular alignment with arbitrary precision parameters, but I don't think too > many people are interested in that aspect). C++ should not be a problem. Licensing is also not a problem. But because you do need to compile with C++ and add an external library I suggest you make this an EMBASSY package. The reason is that you can tailor the EMBASSY package's configure.in and Makefile.am files for your build. The EMBOSS CVS server has an EMBASSY directory "myemboss" that can be used as a template for new EMBASSY packages. You have a choice of requiring users to install GMP or providing a copy in the EMBASSY package distribution. Please mail emboss-bug at emboss.open-bio.org for any help in configuring (or emboss-dev if it is of enough general interest) regards, Peter Rice From brandon.barker at gmail.com Thu Aug 2 00:05:23 2007 From: brandon.barker at gmail.com (Brandon Barker) Date: Thu, 2 Aug 2007 00:05:23 -0400 Subject: [emboss-dev] External Libraries, C++ App In-Reply-To: <46B04013.5050009@ebi.ac.uk> References: <3e078a5e0707312202ua3af58dg4ad75657b9a81c91@mail.gmail.com> <46B04013.5050009@ebi.ac.uk> Message-ID: <3e078a5e0708012105k4b708807je4835d63fd0ffa86@mail.gmail.com> Thanks for the tips and info, sounds good. Another question is regarding the interoperability between AJAX and GMP. Would it be acceptable for me to add some GMP oriented functions that are analogous to existing functions in AJAX and add an optional check in configure.in for GMP? Since building EMBOSS with GMP would still be optional, my parametric alignment package could still be part of EMBASSY, but the AJAX GMP functions would be part of EMBOSS (but only built if the user has GMP, and maybe only if the user explicitly tells configure they want it). If not, I can try to make my own AJAX-like library for GMP types, but this doesn't seem ideal. On 8/1/07, Peter Rice wrote: > Brandon Barker wrote: > > Hello, I'm currently developing a software package that I would like to try > > to add to EMBOSS or EMBASSY. There are two potential problems: the package > > I'm working on is in C++, and most likely won't be changed to C. Also, it > > makes heavy use (in innermost loops, unfortunately, but this can't be > > avoided) of arbitrary precision rationals (or integers) so I need to use an > > external library (GMP). All the licensing should be fine, since all of the > > software and libraries I'm using are under the LGPL or the GPL. The > > application area is parametric and inverse parametric sequence alignment (or > > regular alignment with arbitrary precision parameters, but I don't think too > > many people are interested in that aspect). > > C++ should not be a problem. Licensing is also not a problem. > > But because you do need to compile with C++ and add an external library I > suggest you make this an EMBASSY package. The reason is that you can tailor the > EMBASSY package's configure.in and Makefile.am files for your build. > > The EMBOSS CVS server has an EMBASSY directory "myemboss" that can be used as a > template for new EMBASSY packages. > > You have a choice of requiring users to install GMP or providing a copy in the > EMBASSY package distribution. > > Please mail emboss-bug at emboss.open-bio.org for any help in configuring (or > emboss-dev if it is of enough general interest) > > regards, > > Peter Rice > -- Brandon Barker Phone: (859) 948-5335 From ajb at ebi.ac.uk Thu Aug 2 10:46:06 2007 From: ajb at ebi.ac.uk (ajb at ebi.ac.uk) Date: Thu, 2 Aug 2007 15:46:06 +0100 (BST) Subject: [emboss-dev] External Libraries, C++ App In-Reply-To: <3e078a5e0708012105k4b708807je4835d63fd0ffa86@mail.gmail.com> References: <3e078a5e0707312202ua3af58dg4ad75657b9a81c91@mail.gmail.com> <46B04013.5050009@ebi.ac.uk> <3e078a5e0708012105k4b708807je4835d63fd0ffa86@mail.gmail.com> Message-ID: <47223.81.98.241.17.1186065966.squirrel@webmail.ebi.ac.uk> Hi Brandon, The fewer configuration options there are the better. If the library routines are of general use then they could go into AJAX, however anything there should usually have the AJAX coding style/conventions (PCRE is a bad precedent in that respect). In the case of libgmp the best options are probably either: a) Make it a condition of installation of the EMBASSY package that users install libgmp. or, easier for the user (and support requests), b) Make a sublibrary of the libgmp functions you need within the EMBASSY package. An example of how to do that is the Vienna EMBASSY package in CVS. In the src subdirectory the Makefile.am specifies that some of the source files are used to create a libvienna and then the other source files link against that and the standard EMBOSS libraries. ATB Alan > Thanks for the tips and info, sounds good. Another question is > regarding the interoperability between AJAX and GMP. Would it be > acceptable for me to add some GMP oriented functions that are > analogous to existing functions in AJAX and add an optional check in > configure.in for GMP? Since building EMBOSS with GMP would still be > optional, my parametric alignment package could still be part of > EMBASSY, but the AJAX GMP functions would be part of EMBOSS (but only > built if the user has GMP, and maybe only if the user explicitly tells > configure they want it). > > If not, I can try to make my own AJAX-like library for GMP types, but > this doesn't seem ideal. > > On 8/1/07, Peter Rice wrote: >> Brandon Barker wrote: >> > Hello, I'm currently developing a software package that I would like >> to try >> > to add to EMBOSS or EMBASSY. There are two potential problems: the >> package >> > I'm working on is in C++, and most likely won't be changed to C. >> Also, it >> > makes heavy use (in innermost loops, unfortunately, but this can't be >> > avoided) of arbitrary precision rationals (or integers) so I need to >> use an >> > external library (GMP). All the licensing should be fine, since all of >> the >> > software and libraries I'm using are under the LGPL or the GPL. The >> > application area is parametric and inverse parametric sequence >> alignment (or >> > regular alignment with arbitrary precision parameters, but I don't >> think too >> > many people are interested in that aspect). >> >> C++ should not be a problem. Licensing is also not a problem. >> >> But because you do need to compile with C++ and add an external library >> I >> suggest you make this an EMBASSY package. The reason is that you can >> tailor the >> EMBASSY package's configure.in and Makefile.am files for your build. >> >> The EMBOSS CVS server has an EMBASSY directory "myemboss" that can be >> used as a >> template for new EMBASSY packages. >> >> You have a choice of requiring users to install GMP or providing a copy >> in the >> EMBASSY package distribution. >> >> Please mail emboss-bug at emboss.open-bio.org for any help in configuring >> (or >> emboss-dev if it is of enough general interest) >> >> regards, >> >> Peter Rice >> > > > -- > Brandon Barker > Phone: (859) 948-5335 > _______________________________________________ > emboss-dev mailing list > emboss-dev at lists.open-bio.org > http://lists.open-bio.org/mailman/listinfo/emboss-dev > From brandon.barker at gmail.com Wed Aug 1 05:02:09 2007 From: brandon.barker at gmail.com (Brandon Barker) Date: Wed, 1 Aug 2007 01:02:09 -0400 Subject: [emboss-dev] External Libraries, C++ App Message-ID: <3e078a5e0707312202ua3af58dg4ad75657b9a81c91@mail.gmail.com> Hello, I'm currently developing a software package that I would like to try to add to EMBOSS or EMBASSY. There are two potential problems: the package I'm working on is in C++, and most likely won't be changed to C. Also, it makes heavy use (in innermost loops, unfortunately, but this can't be avoided) of arbitrary precision rationals (or integers) so I need to use an external library (GMP). All the licensing should be fine, since all of the software and libraries I'm using are under the LGPL or the GPL. The application area is parametric and inverse parametric sequence alignment (or regular alignment with arbitrary precision parameters, but I don't think too many people are interested in that aspect). My question is: should I shoot for acceptance in to EMBOSS, EMBASSY, or neither? I would also like to know if someone is already working on a parametric alignment application for EMBOSS =). Thanks, -- Brandon Barker From pmr at ebi.ac.uk Wed Aug 1 08:10:59 2007 From: pmr at ebi.ac.uk (Peter Rice) Date: Wed, 01 Aug 2007 09:10:59 +0100 Subject: [emboss-dev] External Libraries, C++ App In-Reply-To: <3e078a5e0707312202ua3af58dg4ad75657b9a81c91@mail.gmail.com> References: <3e078a5e0707312202ua3af58dg4ad75657b9a81c91@mail.gmail.com> Message-ID: <46B04013.5050009@ebi.ac.uk> Brandon Barker wrote: > Hello, I'm currently developing a software package that I would like to try > to add to EMBOSS or EMBASSY. There are two potential problems: the package > I'm working on is in C++, and most likely won't be changed to C. Also, it > makes heavy use (in innermost loops, unfortunately, but this can't be > avoided) of arbitrary precision rationals (or integers) so I need to use an > external library (GMP). All the licensing should be fine, since all of the > software and libraries I'm using are under the LGPL or the GPL. The > application area is parametric and inverse parametric sequence alignment (or > regular alignment with arbitrary precision parameters, but I don't think too > many people are interested in that aspect). C++ should not be a problem. Licensing is also not a problem. But because you do need to compile with C++ and add an external library I suggest you make this an EMBASSY package. The reason is that you can tailor the EMBASSY package's configure.in and Makefile.am files for your build. The EMBOSS CVS server has an EMBASSY directory "myemboss" that can be used as a template for new EMBASSY packages. You have a choice of requiring users to install GMP or providing a copy in the EMBASSY package distribution. Please mail emboss-bug at emboss.open-bio.org for any help in configuring (or emboss-dev if it is of enough general interest) regards, Peter Rice From brandon.barker at gmail.com Thu Aug 2 04:05:23 2007 From: brandon.barker at gmail.com (Brandon Barker) Date: Thu, 2 Aug 2007 00:05:23 -0400 Subject: [emboss-dev] External Libraries, C++ App In-Reply-To: <46B04013.5050009@ebi.ac.uk> References: <3e078a5e0707312202ua3af58dg4ad75657b9a81c91@mail.gmail.com> <46B04013.5050009@ebi.ac.uk> Message-ID: <3e078a5e0708012105k4b708807je4835d63fd0ffa86@mail.gmail.com> Thanks for the tips and info, sounds good. Another question is regarding the interoperability between AJAX and GMP. Would it be acceptable for me to add some GMP oriented functions that are analogous to existing functions in AJAX and add an optional check in configure.in for GMP? Since building EMBOSS with GMP would still be optional, my parametric alignment package could still be part of EMBASSY, but the AJAX GMP functions would be part of EMBOSS (but only built if the user has GMP, and maybe only if the user explicitly tells configure they want it). If not, I can try to make my own AJAX-like library for GMP types, but this doesn't seem ideal. On 8/1/07, Peter Rice wrote: > Brandon Barker wrote: > > Hello, I'm currently developing a software package that I would like to try > > to add to EMBOSS or EMBASSY. There are two potential problems: the package > > I'm working on is in C++, and most likely won't be changed to C. Also, it > > makes heavy use (in innermost loops, unfortunately, but this can't be > > avoided) of arbitrary precision rationals (or integers) so I need to use an > > external library (GMP). All the licensing should be fine, since all of the > > software and libraries I'm using are under the LGPL or the GPL. The > > application area is parametric and inverse parametric sequence alignment (or > > regular alignment with arbitrary precision parameters, but I don't think too > > many people are interested in that aspect). > > C++ should not be a problem. Licensing is also not a problem. > > But because you do need to compile with C++ and add an external library I > suggest you make this an EMBASSY package. The reason is that you can tailor the > EMBASSY package's configure.in and Makefile.am files for your build. > > The EMBOSS CVS server has an EMBASSY directory "myemboss" that can be used as a > template for new EMBASSY packages. > > You have a choice of requiring users to install GMP or providing a copy in the > EMBASSY package distribution. > > Please mail emboss-bug at emboss.open-bio.org for any help in configuring (or > emboss-dev if it is of enough general interest) > > regards, > > Peter Rice > -- Brandon Barker Phone: (859) 948-5335 From ajb at ebi.ac.uk Thu Aug 2 14:46:06 2007 From: ajb at ebi.ac.uk (ajb at ebi.ac.uk) Date: Thu, 2 Aug 2007 15:46:06 +0100 (BST) Subject: [emboss-dev] External Libraries, C++ App In-Reply-To: <3e078a5e0708012105k4b708807je4835d63fd0ffa86@mail.gmail.com> References: <3e078a5e0707312202ua3af58dg4ad75657b9a81c91@mail.gmail.com> <46B04013.5050009@ebi.ac.uk> <3e078a5e0708012105k4b708807je4835d63fd0ffa86@mail.gmail.com> Message-ID: <47223.81.98.241.17.1186065966.squirrel@webmail.ebi.ac.uk> Hi Brandon, The fewer configuration options there are the better. If the library routines are of general use then they could go into AJAX, however anything there should usually have the AJAX coding style/conventions (PCRE is a bad precedent in that respect). In the case of libgmp the best options are probably either: a) Make it a condition of installation of the EMBASSY package that users install libgmp. or, easier for the user (and support requests), b) Make a sublibrary of the libgmp functions you need within the EMBASSY package. An example of how to do that is the Vienna EMBASSY package in CVS. In the src subdirectory the Makefile.am specifies that some of the source files are used to create a libvienna and then the other source files link against that and the standard EMBOSS libraries. ATB Alan > Thanks for the tips and info, sounds good. Another question is > regarding the interoperability between AJAX and GMP. Would it be > acceptable for me to add some GMP oriented functions that are > analogous to existing functions in AJAX and add an optional check in > configure.in for GMP? Since building EMBOSS with GMP would still be > optional, my parametric alignment package could still be part of > EMBASSY, but the AJAX GMP functions would be part of EMBOSS (but only > built if the user has GMP, and maybe only if the user explicitly tells > configure they want it). > > If not, I can try to make my own AJAX-like library for GMP types, but > this doesn't seem ideal. > > On 8/1/07, Peter Rice wrote: >> Brandon Barker wrote: >> > Hello, I'm currently developing a software package that I would like >> to try >> > to add to EMBOSS or EMBASSY. There are two potential problems: the >> package >> > I'm working on is in C++, and most likely won't be changed to C. >> Also, it >> > makes heavy use (in innermost loops, unfortunately, but this can't be >> > avoided) of arbitrary precision rationals (or integers) so I need to >> use an >> > external library (GMP). All the licensing should be fine, since all of >> the >> > software and libraries I'm using are under the LGPL or the GPL. The >> > application area is parametric and inverse parametric sequence >> alignment (or >> > regular alignment with arbitrary precision parameters, but I don't >> think too >> > many people are interested in that aspect). >> >> C++ should not be a problem. Licensing is also not a problem. >> >> But because you do need to compile with C++ and add an external library >> I >> suggest you make this an EMBASSY package. The reason is that you can >> tailor the >> EMBASSY package's configure.in and Makefile.am files for your build. >> >> The EMBOSS CVS server has an EMBASSY directory "myemboss" that can be >> used as a >> template for new EMBASSY packages. >> >> You have a choice of requiring users to install GMP or providing a copy >> in the >> EMBASSY package distribution. >> >> Please mail emboss-bug at emboss.open-bio.org for any help in configuring >> (or >> emboss-dev if it is of enough general interest) >> >> regards, >> >> Peter Rice >> > > > -- > Brandon Barker > Phone: (859) 948-5335 > _______________________________________________ > emboss-dev mailing list > emboss-dev at lists.open-bio.org > http://lists.open-bio.org/mailman/listinfo/emboss-dev >