Re: Please push your changes to python-xarray
FWIW, it looks like gbp import-dscs --debsnap should lose less historical detail than a single mass commit. -- Aaron Alastair McKinstry writes: > Hi > > Apologies I am swamped on this. Please go ahead and apply. > > Thanks > Alastair > > On 13/11/2023 12:51, Andreas Tille wrote: >> Ping? >> >> I'd love to apply the patch from the bug report and push everything properly. >> If I do not hear from you I might consider mass-commiting all the releases >> you made without pushing to the repository in one single commit and add what >> I'd like to commit. >> >> Kind regards >> Andreas. >> >> Am Tue, Nov 07, 2023 at 09:26:13AM +0100 schrieb Andreas Tille: >>> Hi Alastair, >>> >>> I realised that the Git repository on salsa[1] is lagging a couple of >>> uploads behind the package pool. Please be so kind to push your changes >>> to Salsa. >>> >>> Thanks a lot for your work on this package >>> Andreas. >>> >>> [1] https://salsa.debian.org/science-team/python-xarray >>> >>> -- http://fam-tille.de
Re: How to build compatible packages that use Eigen?
Dima Kogan writes: > Thanks for replying No problem. > Sorry, I'm not familiar-enough with linker scripts. I would pass this to > the linker when building libg2o.so? Or the end-user would need to use > this when build-time linking their application? The run-time dynamic > linker doesn't need this, right? AIUI, you'd supply it when building libg2o.so itself, via the --version-script linker option (or as an implict linker script, but I'd favor being more explicit here). -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: How to build compatible packages that use Eigen?
Dima Kogan writes: > Now let's say you have a user. They're writing a program that uses both > libg2o and Eigen. They're writing their own application, not intended to > go into Debian. So they build with -msse -mavx and all the other fun > stuff. THEIR weak copies of "aligned_malloc" and "aligned_free" are > different and incompatible with the copies in libg2o. And the > application is then likely to crash because at least something somewhere > will be allocated with one copy and deallocated with another. Oy vey. It might help to make Eigen's symbols local to your shared library; to that end, I think it would work to feed the linker a version script reading something along the lines of { local: extern "C++" { Eigen::*; }; global: *; }; (or add such a local block to an existing version script). -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: Joblib 1.2.0 available on salsa
Nilesh Patra writes: > I ask this because uploading a "1.2.0-2+doc" to experimtnal means > it'd be a problem if we meanwhile want to upload a new revision "1.2.0-3" to > unstable, but > "1.2.0+doc-1" would be ok since we can keep uploading 1.2.0-$rev to unstable > meanwhile. With any luck, the next upload to unstable will be able to include the -doc package. If not, the idea would be to pair each subsequent docless unstable 1.2.0-$rev with a corresponding experimental 1.2.0-$rev+doc; I presume the merges would be straightforward. Thanks! -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: Joblib 1.2.0 available on salsa
Nilesh Patra writes: > I could push the package to new again, with a new revision > but I fear another CVE being discovered meanwhile and we run into circles. You can temporarily have a separate branch nominally targeting experimental, with docs included and version numbers along the lines of 1.2.0-1+doc or 1.2.0-1+exp1; subsequent uploads to unstable could either merge it or proceed with time-sensitive fixes, as appropriate. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: New version of nfft breaks its autopkgtest (Was: nfft breaks pynfft autopkgtest on i386: Segmentation fault
Andreas Tille writes: > [1] https://salsa.debian.org/science-team/nfft/-/jobs/3041717 | + gcc -Wall -DNFFT_PRECISION_SINGLE -lnfft3f -lfftw3f -o simple_test simple_test.c | /usr/bin/ld: /tmp/ccaUJ49M.o: in function `simple_test_nfft_1d': | simple_test.c:(.text+0x2c): undefined reference to `nfftf_init_1d' [...] Please try listing simple_test.c ahead of the libraries, which the linker otherwise discards as apparently unneeded. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: pandas 1.3.5+dfsg-2 autopkgtest failures
Drew Parsons writes: > Thanks Rebecca. Looks like the mdtraj error is transient, passes > eventually. Makes it hard to debug robustly. FWIW, I've found rr very helpful in such cases. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: RFS: reupload of opm-common/2021.10-1
FWIW, reuploads are generally in order following acceptance from NEW -- uploads to NEW still require binary packages for the FTP team's convenience, but migration to testing requires a source-only upload so that all binaries will have been autobuilt. Thanks for your contributions, at any rate! -- Aaron Markus Blatt writes: > Turns out that I was a bit too fast with my request. Seems like there are > build issues with libfmt now, > that I need to resolve. Will do that and test on my system and report back. > > Sorry. > > Markus > > Am Fri, Jan 21, 2022 at 01:23:20PM +0100 schrieb Markus Blatt: >>Hi, >> >>opm-common [0] and opm-material [1] have now been accepted by ftpmaster into >>unstable. Thanks a lot for that. >> >>The upload to ftpmaster was in 11/2021 and as a result the binary-package >>opm-common still has >>a dependency on libfmt7. That library is now replaced by libfmt8. Hence the >>package has an unsatisfied >>dependency on libfmt7. >> >>Would be cool if somebody would reupload the source package opm-common [2] >>again to resolve this. >> >>Due to opm-common not building opm-material has an unsatisfied dependency on >>libopm-common-dev. I >>hope that this gets resolved automatically by just reuploading the opm-common >>source package. >> >>If there is something that I can help with or need to do then please yell. >> >>Thanks a lot. >> >>Regards, >> >>Markus >> >>[0] https://tracker.debian.org/pkg/opm-common >>[1] https://tracker.debian.org/pkg/opm-material >>[2] https://salsa.debian.org/science-team/opm-common >>[3] https://salsa.debian.org/science-team/opm-material >> -- >> >>Dr. Markus Blatt - HPC-Simulation-Software & Services http://www.dr-blatt.de >>Pedettistr. 38, 85072 Eichstätt, Germany, USt-Id: DE279960836 >>Tel.: +49 (0) 160 97590858 -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: can I safely use dpkg from sid on bookworm?
Geert Stappers writes: > On Wed, Jan 19, 2022 at 10:08:16AM +0100, Christophe Trophime wrote: [...] >> It seems that this issue has been fixed with latest dpkg (aka 1.21) >> My question is can I use an update dpkg version from sid on my bookworm >> system? > > Yes, https://www.debian.org/releases/index.en.html says that bookworm is > the next release. ... but https://tracker.debian.org/pkg/dpkg says that the latest dpkg (1.21.1) already migrated to testing (bookworm) a few weeks ago. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: [Help] gcc-11 build issue for fastjet
Andreas Tille writes: >'template class std::auto_ptr' is deprecated: use 'std::unique_ptr' > instead I haven't looked into this specific package, but it's generally safe to go ahead with this formal substitution, and in the rare cases where it isn't safe, you'll generally get a compilation error indicating why not. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: arrayfire update fails in configure step
Andreas Tille writes: > https://salsa.debian.org/science-team/arrayfire/-/jobs/1608881 It looks like you're down to two real errors: CMake Error: File /builds/science-team/arrayfire/debian/output/source_dir/extern/forge/CMakeModules/version.h.in does not exist. CMake Error at CMakeModules/AFconfigure_forge_submodule.cmake:47 (configure_file): configure_file Problem configuring file Please try commenting out the configure_file call at the end of CMakeModules/AFconfigure_forge_submodule.cmake. CMake Error at /usr/share/cmake-3.18/Modules/ExternalProject.cmake:2350 (message): error: could not find git for clone of clFFT-ext Call Stack (most recent call first): /usr/share/cmake-3.18/Modules/ExternalProject.cmake:3206 (_ep_add_download_command) CMakeModules/build_clFFT.cmake:33 (ExternalProject_Add) src/backend/opencl/CMakeLists.txt:15 (include) Please try adding a build dependency on libclfft-dev and replacing src/backend/opencl/CMakeLists.txt's inclusion of build_clFFT with a call to find_package(clFFT) > Thanks a lot for your initial hint No problem. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: arrayfire update fails in configure step
Andreas Tille writes: > /usr/bin/ld: cannot find -lpthreads Thanks for posting a link to the full log! AFAICT, the actual errors appear much earlier, on lines 1573-1593: CMake Error: File /builds/science-team/arrayfire/debian/output/source_dir/extern/forge/CMakeModules/version.h.in does not exist. CMake Error at CMakeModules/AFconfigure_forge_submodule.cmake:47 (configure_file): configure_file Problem configuring file Call Stack (most recent call first): CMakeLists.txt:117 (include) CMake Error at CMakeLists.txt:163 (add_subdirectory): add_subdirectory given source "extern/spdlog" which is not an existing directory. CMake Error at CMakeLists.txt:164 (add_subdirectory): add_subdirectory given source "extern/glad" which is not an existing directory. -- Performing Test has_ignored_attributes_flag -- Performing Test has_ignored_attributes_flag - Success -- Performing Test has_all_warnings_flag -- Performing Test has_all_warnings_flag - Success CMake Error at /usr/share/cmake-3.18/Modules/ExternalProject.cmake:2350 (message): error: could not find git for clone of clFFT-ext Call Stack (most recent call first): /usr/share/cmake-3.18/Modules/ExternalProject.cmake:3206 (_ep_add_download_command) CMakeModules/build_clFFT.cmake:33 (ExternalProject_Add) src/backend/opencl/CMakeLists.txt:15 (include) The subsequent output consists of dumps of CMake's internal logs, which sometimes provide additional clues but need to be taken in context; for instance, the -lpthreads error comes from -- Looking for pthread.h -- Looking for pthread.h - found -- Performing Test CMAKE_HAVE_LIBC_PTHREAD -- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Failed -- Looking for pthread_create in pthreads -- Looking for pthread_create in pthreads - not found -- Looking for pthread_create in pthread -- Looking for pthread_create in pthread - found -- Found Threads: TRUE (ll. 1472-1480). -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: numerical comparisons or not? (platform compatibility during testing)
Stephen Sinclair writes: > It could be the way.. but since this would possibly require some > tuning, is there a way to test on different buildd platforms before > the package is actually deployed? You can request temporary guest access to hosts of various architectures, per https://dsa.debian.org/doc/guest-account/. However, that's somewhat more involved than a sponsored upload. An intermediate option would be to perform a sponsored upload to experimental rigged to accept arbitrary test results and dump them for reference, either to the build log or to a file in some existing binary package (which you could then arrange to download). I suppose you could also test in qemu or the like, either fully or by way of cross-building so that the compiler would at least get to run natively. Thanks for your interest in portability and in efficient use of Debian resources! -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: udev rules for usb devices?
"Getz, Robin" writes: > which that means when someone installs the base library (libiio), they > can't actually use any USB devices (as a non-root user), since the > udev rule isn't installed, and the udev rule is dependent on the > libiio-utils package. I'd suggest putting the udev rule in the -utils package and having the runtime library package (libiio0) recommend that package. (A hard dependency would create a cycle, but downgrading the relationship is sufficient to stay out of trouble.) This approach has three big things going for it: * The rule will be in the same binary package as the tool it needs. * APT will then automatically encourage users of the library to install the -utils package, complete with the rule and the tool. * If and when the shared library's SONAME advances, old and new versions will still be coinstallable as needed to allow for smooth transitions. I suppose you could consider splitting iio_info and the rule out into their own binary package and having the runtime package recommend just that package, but the whole -utils package is small enough that I wouldn't bother. > It's not in libiio-dev either, but that makes sense, since non-root > users want to use and run applications too. Right, -dev doesn't particularly make sense here. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: toulbar2 packaging
Thomas Schiex writes: > it appears I used an improper syntax for "bug closing" in the > debian/changelog. The syntax you used was fine. The problem was that I hadn't registered that you added two changelog entries since the last upload, so my upload accidentally only highlighted the newer entry, which didn't mention the bug; sorry about that! The good news is that no additional upload is needed in this situation; it suffices to send mail to (in this case) 831148-d...@bugs.debian.org with an appropriate Version: pseudo-header, as I'll do in a moment. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: toulbar2 packaging
Thomas Schiex writes: > So I patched the CmakeList file to prevent regeneration. There was > already a quilt patch for it. Easy. Sounds even better, thanks! -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: toulbar2 packaging
Thomas Schiex writes: > I pushed all branches. Great, thanks! When retesting with a tarball available, I discovered that the clean target doesn't work properly, for two reasons: - There is a file named clean in the top-level directory, throwing make off. You can unconfuse it by adding the following lines (at minimum) to debian/rules: .PHONY: clean clean: dh clean - The build process winds up modifying src/MyCPackConf.cmake and src/ToulbarVersion.hpp. You can compensate by copying the files elsewhere before configuring the tree, and moving them back when cleaning them up: override-dh_auto_configure: mkdir debian/orig cp -f src/MyCPackConf.cmake src/ToulbarVersion.hpp debian/orig/ dh_auto_configure override-dh_auto_clean: mv -f debian/orig/* src/ rmdir debian/orig dh_auto_clean NB: I've tested only the first suggestion, not the second, which may have typos. Something along those lines should work, though. Retesting otherwise went well. > I must say I would love to mark this as 'done'. So, if Andreas is not > available, it would be great if you could do it :-) Understood; I'll do it Sunday if Andreas hasn't stepped in by then. > In all cases, independently of what you decide, many thanks for all the > pieces of advice and the guidance. No problem; thanks for your contributions! -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: toulbar2 packaging
Thomas Schiex writes: > I just pushed everything to salsa and would be happy if someone could > give it a look and possibly bless it :-) The master branch looks to be in good shape now, but please don't forget to push the corresponding updates to the pristine-tar and upstream branches. As for uploading, if you have a usual sponsor, I'm happy to defer to him or her; otherwise, I should be able to take care of it over the weekend. Thanks! -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: How to handle a change in "Files-Excluded:" with gbp
Thomas Schiex writes: > Is there a way to know what file in the upstream tarball is generating > the message? Good question; I'm surprised -v doesn't say. Does adding --debug help? Meanwhile, please do double-check your doc-base handling; per dh_installdocs(1), best practice is to give the files standardized names and not attempt to list them in .install files. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: How to handle a change in "Files-Excluded:" with gbp
Thomas Schiex writes: > Thanks Aaron, I should have thought about this :-) No problem; glad to have been of help! > I did and I now have 3 less lintian warnings, and I will push to salsa. Great. > I: toulbar2: arch-dep-package-has-big-usr-share 4967kB 67% > > Is it tolerable? Yes, but it would certainly be better to split out a -doc package. > Is it as simple as editing debian/control to describe two deb files and > have a new debian/install file for the doc package? (I guess not :-/) It's pretty straightforward, but not quite that simple. For one thing, you'll want to rename all the files in debian/ specifically pertaining to the main package to start with "toulbar2." for clarity. Also, please note that destinations in .install files should always be directory names, and that dh_installdocs will take care of doc-base files for you if you name them according to the directions in its man page. If you have any heavyweight build dependencies that are necessary only when building the main binary package, or only when building documentation, it's best practice to split them out into Build-Depends-Arch or Build-Depends-Indep, respectively. Finally, I'd recommend running debc (from the devscripts package) to get an overview of each package's contents and metadata while you're working on this setup. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: How to handle a change in "Files-Excluded:" with gbp
Thomas Schiex writes: > gbp:error: Upstream tag 'upstream/1.0.0+dfsg' already exists > > What would you do? I would add a digit to the suffix, going to 1.0.0+dfsg2. Thanks for checking! -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: Help with libgit2 needed to strip code copy from r-cran-git2r
Andreas Tille <andr...@an3as.eu> writes: > Hi, Hi, Andreas. > gcc -std=gnu99 -I/usr/share/R/include -DNDEBUG -I. -I/usr/include/git2 > -DGIT_ARCH_64 -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64 -DGIT_OPENSSL > -DLIBGIT2_NO_FEATURES_H -DGIT_SHA1_OPENSSL - DGIT_SSH -DGIT_CURL > -DGIT_USE_STAT_MTIM -DGIT_USE_NSEC -DHAVE_FUTIMENS -DHAVE_QSORT_R -fpic > -g -O2 -fdebug-prefix-map=/build/r-base-3.4.3=. -fstack-protector-strong > -Wformat - Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -g -c > git2r.c -o git2r.o Please try leaving off -I/usr/include/git2 to avoid this unwanted shadowing. (git2.h includes the other headers as "git2/*.h", so you shouldn't need that flag.) -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: help with d/rules: is it possible to use with dh_auto_build -- $$arg within a for loop ?
Christophe Trophime <christophe.troph...@lncmi.cnrs.fr> writes: > for D in "Axi 3D"; do \ > > for N in "1 2"; do \ Please try removing the quotes on these two lines; as it is, D and N each take on only a single value. > echo "building hifimagnet$$D_P$$N_N$$N" ; \ > dh_auto_build -- hifimagnet$$D_P$$N_N$$N ; \ Please try using $${D} and $${N} here -- because _ is legal in identifiers, the shell otherwise thinks you're asking about variables named D_P and N_N. Thanks for checking! -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: insighttoolkit4 on i386 (+sse2)
David Bremner <da...@tethera.net> writes: > As dumb as you or I might think it, i386 in Debian needs to support > non-SSE hardware. It sounds like you rather want to drop i386 support > for the package, if upstream's idea of i386 doesn't match > Debian's. That's what I did with darktable (that case was maybe more > clear as upstream also felt 32bits was unsuppportable). Gert, Please try building with -ffloat-store, which eliminates the extra internal precision that can lead to test errors but retains compatibility for older i386 processors. If that *doesn't* work, I'd argue that it's better to require SSE support than to give up on i386 altogether just over this one issue, at least if you make this requirement clear up front. As I recall, there were some plans for a mechanism to help packages with extra CPU requirements, but I don't remember further details, sorry. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: mpich FTBFS on sh4: MPIUI_Thread TLS definition mismatches tbss
Drew Parsons <dpars...@debian.org> writes: > The link error suggests 2 different versions of MPIUI_Thread are used. > But I can only see a definition in src/util/thread/mpiu_thread.c. > src/mpi/comm does not refer to it, and comm_rank.c does not use any > MPIU object. So the error message doesn't make sense to me. comm_rank's reference to MPIUI_Thread is also present on at least amd64, and presumably due to some macro that (directly or indirectly) calls MPIU_THREADPRIV_FIELD. However, it's not clear why multithreaded build settings would be in effect for mpiu_thread but not also comm_rank. (It looks like MPIUI_Thread's thread-locality is conditional on MPICH_IS_THREADED, which mpichconf.h defines centrally.) I don't have time to dig deeper, but hope that brief analysis helps. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: RFS: glm/0.9.8.3-2, gli/0.8.2.0+ds1-2
James Clarke <jrt...@debian.org> writes: > I assume these are generated by Doxygen? If so, fixing it in gli is likely to > be fragile. rdfind could help clean these up without hardcoding any relationships, but might not be worth bothering with. > I see; apparently this is not just jQuery, but also other Doxygen code. Indeed; see https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=736360 and https://sources.debian.net/src/doxygen/1.8.13-4/debian/README.jquery/ . (I've run into the same issue in the course of maintaining fltk1.3.) -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: Compiling with VTK (version 6)
Ole Streicher <oleb...@debian.org> writes: > I am currently packaging a Qt/C++ program (dpuser+qfitsview [1]) that > requires VTK6, and I am curious of how I have to specify the include > path and the libraries. To the best of my knowledge, the only build system VTK supports cleanly is CMake, which I understand to be generally popular in the Qt (or at least KDE) community. Perhaps you could convince/help upstream to migrate? https://sourceforge.net/projects/qmake2cmake/ might provide a decent starting point. > It also looks like a transition is needed for any minor versions of > vtk, right? Yes, but since you've arranged to avoid hardcoding the version, binNMUs should at least suffice; you shouldn't need to be directly involved unless there are incompatible API changes. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: Architecture field based on another package Architecture field
Ghislain Vaillant <ghisv...@gmail.com> writes: > Architecture: ${libfftw3-long3:Architecture} Alas, this won't work -- that information isn't available here, and I don't think substitutions work in the Architecture field anyway. However, debian/rules can conditionally invoke dh with a suitable -N flag. I believe it should work to specify ifeq "" "$(wildcard $(libdir)/libfftw3l.so)" SKIP_NFFTL = -Nlibnfft3-long2 endif %: dh $@ --parallel --with autoreconf \ --dbg-package=libnfft3-dbg $(SKIP_NFFTL) and then conditionalize other commands involving the nfftl tree on ifneq "" "$(SKIP_NFTL)" I am testing these changes now (against a checkout of your experimental branch) and will follow up with a full patch if they work. Good question! -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: Architecture field based on another package Architecture field
Ghislain Vaillant <ghisv...@gmail.com> writes: > So the final solution is to manually list all supported architectures > in d/control (same one as libfftw3-double3) and use your patch for > d/rules? You can leave Architecture: any in d/control; dpkg-genchanges will warn about the discrepancy, but proceed anyway. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: Artistic License 2.0 -> copyright file
Ghislain Vaillant <ghisv...@gmail.com> writes: > Very few license actually have a short statement like the one you > quote for the GPL-2. For the others, including Artistic-2.0, the full > content of the license must be copy-pasted verbatim in d/copyright. In this particular case, you may find parrot's copyright file to be a useful example: http://metadata.ftp-master.debian.org/changelogs//main/p/parrot/parrot_6.6.0-1_copyright -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: RFM: transition of flint
Julien Puydt <julien.pu...@laposte.net> writes: > Both the i386 and kfreebsd-i386 builds were terminated after 150 > minutes, which is precisely 2h30min, so it does look like some round > time. For hurd-i386, it was 180 minutes which is precisely 3h. I > really think it's just a buildd setting which isn't set to the same > value for each build server. And I think the package is ok but tests > are slow and just hit the deadline. Right, the timeout varies by architecture, being longer on those that are generally slower. The good news is that it resets on any output to stdout (or stderr, I think), so adjusting the tests to print progress indicators every so often should let the build proceed. (The timeout is meant to catch outright hangs, which do sometimes happen.) -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Re: r-cran-vcdextra package
Because this upload introduced new package names, it is sitting in the NEW queue waiting for ftpmaster review. https://ftp-master.debian.org/new.html https://ftp-master.debian.org/new/r-cran-gnm_1.0-8-1.html -- Aaron Jonathon Love j...@thon.cc writes: hi andreas, balint reczey submitted the r-cran-gnm package about 48 hours ago. is this enough time for it to appear in the system? i've been waiting for it to appear here (it hasn't yet): https://packages.debian.org/search?keywords=r-cran-gnmsearchon=namessuite=allsection=all is this the right place to be checking (and the package isn't available yet?) or is there a different place i should be checking? with thanks jonathon i've sent a couple of emails to the list regarding the r-cran-vcdextra package, but perhaps you overlooked them. I guess its your turn to respond to my message The Dependency r-cran-gnm is missing as I wrote in my other mail.
Re: Packages in Jessie
Leopold Palomo-Avellaneda l...@alaxarxa.net writes: So, one thing it that the package has important bugs and we are not able to solve it in time, so this package will not be in Jessie. But, are you telling me that the package won't be in Jessie because 6 hours after all the steps that I have done? Not automatically, no. However, if you ask the Release Team nicely, they might consider making an exception. Specifically, you'd want to file an unblock request against the release.debian.org pseudopackage: $ reportbug release.debian.org [...] What sort of request is this? (If none of these things mean anything to you, or you are trying to report a bug in an existing package, please press Enter to exit reportbug.) 1 binnmu binNMU requests 2 britney testing migration script bugs 3 other None of the other options 4 rm Stable/Testing removal requests 5 squeeze-pu squeeze proposed updates requests 6 transition transition tracking 7 unblock unblock requests 8 wheezy-pu wheezy proposed updates requests Choose the request type: 7 Please enter the name of the package: pcl Checking status database... Please enter the version of the package: 1.7.2-1 Enter any additional addresses this report should be sent to; press ENTER after each address. Press ENTER on a blank line to continue. debian-science@lists.debian.org You can disregard the request for a diff against testing, since this package is wholly new. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/udlioj5mug4@contents-vnder-pressvre.mit.edu
Re: Mpich 1/2/3
Torquil Macdonald Sørensen torq...@gmail.com writes: Can I trust dpkg-architecture to always return the correct architecture triplet? Or is there any cross-compilation or such that would force me to do it in other ways? I'm not very familiar with how things are done on the Debian build servers. dpkg-architecture -qDEB_HOST_MULTIARCH should yield the correct triplet at build time, even when cross-compiling (assuming there are no other obstacles to doing so). -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlmwpwsb4z@dr-wily.mit.edu
Re: Rename mpich2 to mpich
Torquil Macdonald Sørensen torq...@gmail.com writes: Can I simply move the repository, thus making the old GIT path invalid? Good question. You can move it and leave a compatibility symlink. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udl4nfpjatv@dr-wily.mit.edu
Re: Bug#694418: ITP: fits -- Java library for the I/O handling of FITS files
Florian Rothmaier froth...@ari.uni-heidelberg.de writes: but then the matching files would all have the name fits_src.jar and uscan would not see if one of them is newer than the local version. Or am I wrong? You can have uscan just look for the directory, as with ncbi-tools6: http://anonscm.debian.org/gitweb/?p=debian-med/ncbi-tools6.git;a=blob;f=debian/watch;h=03093f53f463965cb8febcf1803d6da1ce828e10;hb=HEAD The debian/watch syntax even defines a downloadurlmangle option that should allow uupdate to work, at least for HTTP URLs (as you have). I'm not sure if anyone ever implemented FTP downloadurlmangle support, but don't hugely care because I generally prefer to download archives manually anyway. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlvcc0ipw7@dr-wily.mit.edu
Re: libhe5-hdfeos0 vs. libhdf5-openmpi-7: fails to provide?
Ivan Shmakov oneing...@gmail.com writes: It looks that while libhdf5-openmpi-7 Provides: libhdf5-7, the libhe5-hdfeos0 package somehow insists on installing the “real” libhdf5-7, thereby preventing the cdo package from being installed at the same time. This problem arises because dpkg conservatively refuses to let provided package names satisfy versioned dependencies. The current hdf5 packaging accounts for that by indicating that dependent packages should declare unversioned dependencies, but hdf-eos5 just mixed the fix. The next hdf5-eos upload of any sort should pick up a proper unversioned dependency; a binary-only rebuild may be in order unless its maintainer plans to issue another sourceful upload soon. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlvci28as8@dr-wily.mit.edu
Re: Bug#679990: ITP: clipper -- object oriented development kit for crystallographic computing
Radostan Riedel raybu...@googlemail.com writes: So what are you suggesting about the name? I'd suggest letting it stand, if only to avoid unnecessary differences with Ubuntu. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlbojwyku8@dr-wily.mit.edu
Re: RFS: ngspice_24-1
Gudjon I. Gudjonsson gud...@gudjon.org writes: Finding the minimal latex dependencies is however difficult. I gave up when building sdcc and just sat them to texlive-full, 2GB. Do you have any hint on how to do that? Good question. In general, TeX produces a log file indicating (among other things) all direct input files. However, when using LyX (as in this case), the log normally winds up in a temporary directory; to compensate for that, you can run lyx --export latex manual.lyx and then directly run latex on the result. From there, a shell one-liner will indicate which packages you need: $ sed -ne 's,.*(\(/usr/share/tex\),\1,p' manual.log | xargs dlocate -S --package-only latex2html lmodern texlive-base texlive-generic-recommended texlive-latex-base texlive-latex-extra texlive-latex-recommended However, latex2html depends in turn on most of the texlive-* packages above, so you just need it, lmodern, and texlive-generic-recommended -- plus texlive-lang-greek, which this method didn't catch because TeX doesn't log font usage in such detail. :-/ BTW, I'd also advise moving all such build dependencies to Build-Depends-Indep if possible, as appears to hold here. Thanks! -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlaa0jhrs1@dr-wily.mit.edu
Re: Restricting citation information to single binary packages?
Andreas Tille andr...@an3as.eu writes: Aaron, could you provide some bibliographic information for those ncbi-tools which are not specifically describing blast? This would be a critical point to find a sane decision. I haven't come across any, but could try to dig some up. Regardless, please note that the BLAST engine is a library, and available to other packages; for instance, I think Sequin (from the ncbi-tools-x11 package) may support running BLAST internally. Thanks for asking! -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udl4nt6m2ej@dr-wily.mit.edu
Re: FreeFem++-3.18 upload
Sylvestre Ledru sylves...@debian.org writes: g++ -g -DNDEBUG -O3 -mmmx -msse -msse2 -DDRAWING -DBAMG_LONG_LONG -DNCHECKPTR -fPIC -rdynamic -o FreeFem++-glx xglrgraf.o parallelempi-empty.o ../lglib/liblg.a ../fflib/libff.a -lumfpack -lamd -larpack -llapack -lblas -lGLU -lGL -lXxf86vm -lXext -lX11 -lXpm -lm -ldl /usr/lib/gcc/x86_64-linux-gnu/4.6/libgfortran.so collect2: ld returned 1 exit status Please check that your build area has sufficient disk space. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udld39p2u0q@dr-wily.mit.edu
Re: physics software
Mohammad Hosein m.hosei...@yahoo.com writes: link: http://ab-initio.mit.edu/wiki/index.php/Meep It's already available in Debian, if that's what you're asking: http://packages.debian.org/squeeze/meep -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udl4nvi8po9@dr-wily.mit.edu
Re: New metapackages for Debian Science: science-highenergy-physics
Anton Gladky gladky.an...@gmail.com writes: svn: No repository found in 'svn+ssh://svn.debian.org/blends/projects/science/trunk/debian-science' Please try with a leading /svn: svn+ssh://svn.debian.org/svn/blends/projects/science/trunk/debian-science/ -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udl8vnpqcx7@dr-wily.mit.edu
Re: pdb files license
Picca Frédéric-Emmanuel frederic-emmanuel.pi...@synchrotron-soleil.fr writes: I am a little bit concern by this redistribution restriction. It seems that redistribuing unmodified file is allow but if you modify a file it is not allow to redistribut it with the same name on the FTP server. So DFSGFree or not ? Yes, DFSG #4 grudgingly allows such restrictions, which are annoying but understandable (and shared to a lesser extent by TeX, which requires a name change to accompany any backwards-incompatible modifications). -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlaa9rz7p2@dr-wily.mit.edu
Re: Scientific software packaging
Yngve Inntjore Levinsen yngve.inntjore.levin...@cern.ch writes: (is there a way to do some kind of build --continue so that the packaging does not have to start from scratch?) debuild -nc; you might also find ccache interesting. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlliwwh4yh@dr-wily.mit.edu
Re: updating mpi-defaults (decommissioning lam)
Drew Parsons dpars...@debian.org writes: 1) why is the experimental buildd not using the latest experimental package? Typically because the build dependencies weren't strict enough to force it to, as experimental autobuilders otherwise favor versions from unstable. I haven't checked whether that applies to this particular case, though. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlsjr5xl4n@dr-wily.mit.edu
Re: Installation of GATE 6.1 on Debian Squeeze
Marcelo Santana marcgsant...@yahoo.com.br writes: [1]http://www.opengatecollaboration.org/ That site is very pretty, but contains no obvious link to sources; I wound up examining them via http://www.creatis.insa-lyon.fr/~dsarrut/gate/doxygen/html/. I'm trying and I've got some errors in compilation as you can see attached These mostly appear to stem from unconditional usage of conditionally visible declarations. More specifically: src/GateActions.cc:193: error: ‘GateOutputMgr’ has not been declared GateActions.cc #includes GateOutputMgr.hh only when G4ANALYSIS_USE_GENERAL is defined, but then uses GateOutputMgr unconditionally. One of those policies is clearly wrong. src/GateActions.cc:195: error: ‘GateToRoot’ was not declared in this scope [...] src/GateActions.cc:626: error: ‘ARFDataToRoot’ was not declared in this scope Likewise, GateToRoot.hh and ARFDataToRoot.hh conditionalize their contents on G4ANALYSIS_USE_ROOT, but GateActions.cc uses GateToRoot and ARFDataToRoot unconditionally. src/GateApplicationMgr.cc:700: error: ‘GateUserActions’ has not been declared In this case, I see no relevant #include directive at all; perhaps something would #include GateUserActions.hh indirectly if an appropriate macro were defined, though. src/GateSourceMgr.cc:184: error: expected type-specifier before ‘GateSourcePencilBeam’ [...] src/GateSourceMgr.cc:189: error: expected type-specifier before ‘GateSourceTPSPencilBeam’ As with GateActions.cc, the headers need G4ANALYSIS_USE_ROOT but the source uses their types unconditionally. At a higher level, there's also the problem of why those macros aren't getting defined, as their names suggest they should be; regardless, the code should degrade gracefully in their absence, rather than failing even to compile! I hope this analysis helps. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlr57lfgbd@dr-wily.mit.edu
Re: yade build problem on i386, powerpc and sparc platforms
Anton Gladky gladky.an...@gmail.com writes: Depends: liblog4cxx9-dev which is a virtual package. liblog4cxx10-dev has replaced liblog4cxx9-dev, which is only available in oldstable. Your system might still have it installed, but others will not. Depends: python-sphinx (= 0.99) but it is not going to be installed. I'm not sure what the problem is there, though; the autobuilders would need to install the version from experimental, but if your package is in experimental as well that shouldn't be an issue. Perhaps it depends on something they couldn't install for some reason. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlzkohdehf@dr-wily.mit.edu
Re: fftw: Usage of SSE in 64bit?
Carsten Aulbert carsten.aulb...@aei.mpg.de writes: Is there a specific reason why SSE is enabled for 32bit and not enabled for 64bit as the difference in performance can easily be a factor of 2 (on 64bit). Good question. Perhaps the maintainer thought that those options merely turned on compiler flags which would be redundant on amd64 systems, which all support SSE2 at a minimum. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udl4o6t6heo@dr-wily.mit.edu
Re: Need help/advice with NIST dataplot
Paul E Condon pecon...@mesanetworks.net writes: First, maybe I don't need to compile from source --- Is there a package for this in Debian? What is the package name? I see no sign of any existing package either, but it was a good thought to check first. GD, LIBPLOT, or READLINE I would recommend installing the following packages, all of which I expect should independently improve your experience: build-essential (a metapackage for various generally useful development tools) libgd2-xpm-dev (development files for a library for producing image files) libplot-dev (development files for a graph-drawing library) libreadline-dev (development files for an input-line editing library) You should then be able to say yes for all three. TIA NP. HTH; HAND. ;-) -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udltyib7uve@dr-wily.mit.edu
Re: NASA CDF
Teemu Ikonen tpiko...@gmail.com writes: The conclusion I draw from the threads above is that the license above is DFSG-free. I agree with you, but also with those who observed that the notion of a federal agency holding copyright is a bit odd; published US government work falls into the public domain, and I would expect a work for hire (by a contractor) to have a corresponding copyright holder. IANAL, but I do work directly for a different US government agency (the National Center for Biotechnology Information) as a software developer. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlbp4t5ssq@dr-wily.mit.edu
Re: mathgl 1.11 upload
Sylvestre Ledru sylves...@debian.org writes: I tried to build it into a chroot and it fails: /usr/include/hdf/hdfi.h:1011: error: conflicting declaration 'typedef long int intf' /usr/include/zconf.h:340: error: 'intf' has a previous declaration as 'typedef int intf' This looks like an architecture-specific issue; hdfi.h typedefs intf to be long on i386 and int on amd64, whereas zconf.h uses int regardless. The types are the same size on i386, but still formally distinct, hence the error. To work around it, you may have to use the preprocessor to divert one library's definition: [...] #include zlib.h #ifdef HAVE_HDF5 #include hdf5.h #endif #ifdef HAVE_HDF4 #define intf hdf4_intf #include hdf/mfhdf.h #undef intf #endif [...] -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udl7hgemwci@dr-wily.mit.edu
Re: atlas, lapack, blas
[Addressing you directly because I'm not sure you're subscribed; apologies in advance for any resulting duplication.] are these packages coming back in a future version, or does the libatlas3gf-base package now include libs optimised for opteron? AIUI, the CPU-type-specific packages proved more trouble than they were worth; although there may be some benefit to be had from accounting for processor type, especially on architectures such as i386 with wide ranges, there are other parameters for which ATLAS should ideally be tuned, including the number of available cores per system and the sizes of various CPU caches. (In particular, please note that even the baseline AMD64 instruction set features SSE2 instructions.) Instead, ATLAS ships a README.Debian documenting how to build packages optimized for your exact hardware: http://svn.debian.org/viewsvn/debian-science/packages/atlas/trunk/debian/README.Debian?revision=36882view=markup This is a semi-manual process for the time being, but Sylvestre has mentioned plans to supply a binary package containing sources to be built automatically upon installation or perhaps invocation of some script. (At any rate, given your experience, I expect you should have no trouble following README.Debian's directions.) -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlocag4i2o@dr-wily.mit.edu
Re: how to deal with bibtex with long fields
Jameson Rollins jroll...@finestructure.net writes: Hey, Aaron. Thanks for the suggestions. No problem. UnicodeDecodeError: 'utf8' codec can't decode byte 0xd7 in position 118533: invalid continuation byte That looks like an encoding issue; I would suggest either converting your .bib file to UTF-8 or running pybtex -e ISO-8859-1. 1 $ nbibtex thesis Call failed: /usr/share/nbibtex/bibtex.lua:1460: module 'nbib-plain' not found: no field package.preload['nbib-plain'] That looks more like a packaging bug. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udl7hj660dc@dr-wily.mit.edu
Re: how to deal with bibtex with long fields
Jameson Rollins jroll...@finestructure.net writes: Hey, Jim. Yeah, I did see that but yikes! I should have put that on my list of unacceptable solutions! Someone else suggested this might be the only solution, but they thought it might also require rebuilding all of tex as well. Not something I have any intention of doing. Take a closer look; the the FAQ also mentions the possibility of running bibtex8, which Debian ships (right in texlive-binaries). It appears to be only optionally encoding-aware, so you may be able to dodge that issue while still requesting capacity increases with -B, -H, -W, or individual --m* flags. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlpqwy4iug@dr-wily.mit.edu
Re: how to deal with bibtex with long fields
Jameson Rollins jroll...@finestructure.net writes: replacement. In any event, using ISO-8859-1 didn't work either: Bleh; so much for that idea. Maybe after I defend... Priorities, priorities. ;-) -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlzkw24j0u@dr-wily.mit.edu
Re: Salomé rules file and make wildcards
Adam C Powell IV hazel...@debian.org writes: I've been beating my head on $(wildcard...) in a rules file for nearly a week now, and can't make any sense of it. The rules file is attached, In general, using $(wildcard ...) differs in two ways from using shell wildcards directly (which I would recommend simply doing here): - make expands it before running *any* commands; that allows its use in some contexts (particularly target lists), but backfires here because there won't necessarily be a *-packages directory when make starts. - When there are no matches, it expands to nothing, which is not default shell behavior (but available with bash's shopt nullglob). Your observations are consistent with that; in particular, your echo statement lists site-packages only because you fed the shell an unquoted wildcard. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/udlhbp6ysvz@dr-wily.mit.edu
Re: Changing section from math to science?
Manuel Prinz man...@debian.org writes: I do not see an urgent need to change the section. If you think that Yorick can be useful in scientific areas other than maths the science section seems to be a good candidate. I do not think interpreter is not the right section to put it in. Agreed. Since this is no urgent change, I suggest you simply change it in the next regular upload. IIRC, the section info known to apt et al. comes from ftpmaster overrides, which you can get changed by filing a bug against ftp.debian.org. I'm not sure whether the .debs should already specify the desired sections, though. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-science-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: GCTP license
Ivan Shmakov [EMAIL PROTECTED] writes: Is it in public domain? Probably, though not necessarily, as the USGS may have contracted the work out. Have you tried to get in touch with the authors? -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/[EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: package sponsorship request
Rudi Cilibrasi [EMAIL PROTECTED] writes: Michael Koch believes it is inappropriate for me to be changing AUTHORS and THANKS. Actually, I made those up quite some time ago to allow for automake to be used in packaging (optionally). But, he As far as this particular issue goes, have you considered simply running automake --foreign ? -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/[EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: exceptions in gnu c++
Paul E Condon [EMAIL PROTECTED] writes: row: 21100 128Killed Killed almost certainly indicates that something (likely the kernel) sent your application a SIGKILL signal, which instantly terminates your application, with the message itself coming from your shell. (Running your program under strace should confirm this.) Moreover, Unix signals are generally a separate matter from exceptions, and converting one to the other is easier said than done. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) Finger [EMAIL PROTECTED] (NOT a valid e-mail address) for more info. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How much interest in a debian-science.org repository?
Thomas Walter [EMAIL PROTECTED] writes: By the way, what is the license status of 'contrib'? Is it: not 'non-free' and not 'free' in the debian spirit. So where in this 3-way free _per se_, but tied down by a dependency (at build time or runtime) on non-free software. Many Java packages that were once in contrib are now in main because free JVMs and runtimes have gotten better. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) Finger [EMAIL PROTECTED] (NOT a valid e-mail address) for more info. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: New source and binary Debian packages for ROOT.
Christian Holm Christensen [EMAIL PROTECTED] writes: dpkg-gencontrol: error: current build architecture powerpc does not appear in package's list (i386 ia64) If you're using debhelper, I believe you can avoid this problem by replacing any uses of -a with its smarter variant -s. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) Finger [EMAIL PROTECTED] (NOT a valid e-mail address) for more info. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Random numbers generation
Lisandro Damian Nicanor Perez Meyer [EMAIL PROTECTED] writes: ¿Is there any way of making differents results without having to change the seed from the outside? It looks like you can also call gsl_rng_set(r, some_seed); or assign directly to gsl_rng_default_seed, at least with the version of gsl available in sid (unstable). -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) Finger [EMAIL PROTECTED] (NOT a valid e-mail address) for more info. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]