Re: Autopkgtest for Yanagiba and unc-fish
On 31.07.21 14:47, Nilesh Patra wrote: On 31/07/21 11:23 AM, Shruti Sridhar wrote: yanagiba is almost ready, since you downloaded data from ncbi database, can you write a script for fetching data by running it? Just like the way I did it for pilercr[5] I don't have a good idea of navigating this file at the right location Note that doing this is ``optional``, and let me know if it cannot be done I have pushed the get-test-data script. That would be in complete analogy to the createmanpages script (except for the hyphens). I like that! Could you please have a look at it? Thank you! Looks pretty good. The fastq-dump command piped with head was giving me a weird SIGPIPE error so I split it into two operations instead. And the data seemed to have changed (just a little) so I updated this and pushed. Nice! Steffen
Re: Autopkgtest for Yanagiba and unc-fish
On 31/07/21 11:23 AM, Shruti Sridhar wrote: > > yanagiba is almost ready, since you downloaded data from ncbi database, > > can you write a script for fetching data by running it? > > Just like the way I did it for pilercr[5] > > I don't have a good idea of navigating this file at the right location > > Note that doing this is ``optional``, and let me know if it cannot be done > > > I have pushed the get-test-data script. Could you please have a look at it? > Thank you! Looks pretty good. The fastq-dump command piped with head was giving me a weird SIGPIPE error so I split it into two operations instead. And the data seemed to have changed (just a little) so I updated this and pushed. Nilesh signature.asc Description: PGP signature
Re: Autopkgtest for Yanagiba and unc-fish
> > > > Hello! > > > > I have written autopkgtests for yanagiba[1] and unc-fish[2]. > > > > The yanagiba package seems to fail the ci pipeline. The error seems to be > > from - > > > > dpkg: yanagiba-build-deps: dependency problems, but configuring anyway as > > you requested: > > yanagiba-build-deps depends on dh-python; however: > > Package dh-python is not installed. > > yanagiba-build-deps depends on python3-setuptools; however: > > Package python3-setuptools is not installed. > > yanagiba-build-deps depends on python3-pandas; however: > > Package python3-pandas is not installed. > > yanagiba-build-deps depends on python3-biopython; however: > > Package python3-biopython is not installed. > > yanagiba-build-deps depends on python3-nanomath; however: > > Package python3-nanomath is not installed. > > => For yanagiba > > Not quite, the error instead comes from: > > dpkg-source: error: unwanted binary file: debian/tests/data/data.fastq.gz > dpkg-source: error: detected 1 unwanted binary file (add it in > debian/source/include-binaries to allow its inclusion). > > So whenever you include "binaries" (with sources) in d/tests/data or > d/ > you need to allow its inclusion via d/source/include-binaries. Here it > is just a compressed file, hence should be fine > > Other than that, copyright was wrong. There should not be a space before > ":" > (colon) and after Files, i.e. it has to be "Files: ..." and not "Files : > " > > And "Team Upload" need to be the top entry, not somewhere in the middle. > Fixed all of these and pushed. > > Actually, there is a tool called lintian[3] which helps find out these > things, and every package should be as much lintian clean as possible. > > Although it is best to run it locally and check, > It also shows up in the salsa-CI - although it passes, but that is > usually not enough. The bare minimum is to fix error + warnings. There > can be determined w/ statements beginning with "W:" > as an example, in your salsa CI log[4] for unc-fish, this is a warning > > "W: unc-fish source: no-nmu-in-changelog" > > yanagiba is almost ready, since you downloaded data from ncbi database, > can you write a script for fetching data by running it? > Just like the way I did it for pilercr[5] > I don't have a good idea of navigating this file at the right location > Note that doing this is ``optional``, and let me know if it cannot be done I have pushed the get-test-data script. Could you please have a look at it? Thank you! > > => Now, For unc-fish: > > Looks fine, but as you might notice, this package is not in the archive > yet. > The reason is the TODO mentioned in d/changelog: > > """ > TODO: Find reason for failure: > $ /usr/lib/unc-fish/fish -b blocks.txt -B simpleblocks.txt -g > grids.txt -C contigs.txt > Segmentation fault > """ > > I modified this command placing all relevant files in the directory I > amd running it, and I get: > > $ unc-fish -f control.txt -b blocks.txt -B simpleblocks.txt -g grids.txt > -C contigs.txt > > corrupted size vs. prev_size > [1]30290 abort unc-fish -f control.txt -b blocks.txt -B > simpleblocks.txt -g grids.txt -C > > I did not take a deeper look at it, can you check once? > > Also, @Andreas, why is test data stored in "/var/lib/fish" and not > "/usr/share/doc"? This looks > sort of weird, is that really intended? > > > [1] https://salsa.debian.org/med-team/yanagiba > > [2] https://salsa.debian.org/med-team/unc-fish > [3]: https://wiki.debian.org/HowToPackageForDebian#Lintian > [4]: https://salsa.debian.org/med-team/unc-fish/-/jobs/1782347 > [5]: > https://salsa.debian.org/med-team/pilercr/-/blob/master/debian/tests/get-test-data > > Nilesh > ᐧ
Re: Autopkgtest for Yanagiba and unc-fish
On Thu, Jul 29, 2021 at 10:53:43PM +0530, Nilesh Patra wrote: > => Now, For unc-fish: > > Looks fine, but as you might notice, this package is not in the archive > yet. > The reason is the TODO mentioned in d/changelog: > > """ > TODO: Find reason for failure: > $ /usr/lib/unc-fish/fish -b blocks.txt -B simpleblocks.txt -g grids.txt > -C contigs.txt > Segmentation fault > """ > > I modified this command placing all relevant files in the directory I > amd running it, and I get: > > $ unc-fish -f control.txt -b blocks.txt -B simpleblocks.txt -g grids.txt -C > contigs.txt > > corrupted size vs. prev_size > [1]30290 abort unc-fish -f control.txt -b blocks.txt -B > simpleblocks.txt -g grids.txt -C > > I did not take a deeper look at it, can you check once? > > Also, @Andreas, why is test data stored in "/var/lib/fish" and not > "/usr/share/doc"? This looks > sort of weird, is that really intended? I think I once pushed this package as some precondition of ngopt. However, this package is in such a bad state[1] that I gave up on it and thus unc-fish - which is pretty suspicious code itself - was never uploaded. I simply forgot why I had this unusual solution. I'd recommend not to spent much time on it. BTW, as a strong recommendation to Shruti to find sendible targets for writing autopkgtests. I've once written a script[2] to sort the packages of tasks of a Blend according to its popularity contest. You can sensibly call it by 0-missing-autopkgtest.sh debian-med bio bio-dev It created some output file you can inspect - may be by filtering those packages that just have some autopkgtest out via $ grep -v autopkgtest debian-med-bio-tests.out | head -n20 source | vote | testsuite | tags --+--+-+ hdf5 | 359 | | networkx | 275 | | avogadro | 67 | | {uitoolkit::qt} jmol | 62 | | rasmol | 50 | | {interface::x11,uitoolkit::gtk,x11::application} htsjdk | 25 | | viewmol | 21 | | {x11::application,uitoolkit::motif,interface::x11} ghemical | 13 | | {x11::application,interface::x11,uitoolkit::gtk} clustalx | 12 | | {interface::x11,uitoolkit::motif,x11::application} garlic | 11 | | {interface::x11,uitoolkit::xlib,x11::application} gamgi| 10 | | {uitoolkit::gtk} adun.app |8 | | {interface::x11,x11::application,uitoolkit::gnustep} adun.app |8 | | ngs-sdk |8 | | ngs-sdk |7 | | qutemol |7 | | {uitoolkit::wxwidgets,interface::x11,uitoolkit::glut,x11::application} salmid |6 | | treeviewx|6 | | {interface::x11,uitoolkit::wxwidgets,x11::application} The graphical interface usually means that it is hard to write a test - but may be a second look might help. (I'm sorry that the script has some duplicates like (adun.app and ngs-sdk - probably a broken join, but it should not harm here and might help to find the most urgent targets for tests anyway.) Kind regards Andreas. [1] https://salsa.debian.org/med-team/ngopt/-/blob/master/debian/README.source [2] https://salsa.debian.org/blends-team/website/-/blob/master/misc/sql/0-missing-autopkgtest.sh -- http://fam-tille.de
Re: Autopkgtest for Yanagiba and unc-fish
On 29/07/21 09:40 PM, Shruti Sridhar wrote: > Hello! > > I have written autopkgtests for yanagiba[1] and unc-fish[2]. > > The yanagiba package seems to fail the ci pipeline. The error seems to be > from - > > dpkg: yanagiba-build-deps: dependency problems, but configuring anyway as > you requested: > yanagiba-build-deps depends on dh-python; however: > Package dh-python is not installed. > yanagiba-build-deps depends on python3-setuptools; however: > Package python3-setuptools is not installed. > yanagiba-build-deps depends on python3-pandas; however: > Package python3-pandas is not installed. > yanagiba-build-deps depends on python3-biopython; however: > Package python3-biopython is not installed. > yanagiba-build-deps depends on python3-nanomath; however: > Package python3-nanomath is not installed. => For yanagiba Not quite, the error instead comes from: dpkg-source: error: unwanted binary file: debian/tests/data/data.fastq.gz dpkg-source: error: detected 1 unwanted binary file (add it in debian/source/include-binaries to allow its inclusion). So whenever you include "binaries" (with sources) in d/tests/data or d/ you need to allow its inclusion via d/source/include-binaries. Here it is just a compressed file, hence should be fine Other than that, copyright was wrong. There should not be a space before ":" (colon) and after Files, i.e. it has to be "Files: ..." and not "Files : " And "Team Upload" need to be the top entry, not somewhere in the middle. Fixed all of these and pushed. Actually, there is a tool called lintian[3] which helps find out these things, and every package should be as much lintian clean as possible. Although it is best to run it locally and check, It also shows up in the salsa-CI - although it passes, but that is usually not enough. The bare minimum is to fix error + warnings. There can be determined w/ statements beginning with "W:" as an example, in your salsa CI log[4] for unc-fish, this is a warning "W: unc-fish source: no-nmu-in-changelog" yanagiba is almost ready, since you downloaded data from ncbi database, can you write a script for fetching data by running it? Just like the way I did it for pilercr[5] I don't have a good idea of navigating this file at the right location Note that doing this is ``optional``, and let me know if it cannot be done => Now, For unc-fish: Looks fine, but as you might notice, this package is not in the archive yet. The reason is the TODO mentioned in d/changelog: """ TODO: Find reason for failure: $ /usr/lib/unc-fish/fish -b blocks.txt -B simpleblocks.txt -g grids.txt -C contigs.txt Segmentation fault """ I modified this command placing all relevant files in the directory I amd running it, and I get: $ unc-fish -f control.txt -b blocks.txt -B simpleblocks.txt -g grids.txt -C contigs.txt corrupted size vs. prev_size [1]30290 abort unc-fish -f control.txt -b blocks.txt -B simpleblocks.txt -g grids.txt -C I did not take a deeper look at it, can you check once? Also, @Andreas, why is test data stored in "/var/lib/fish" and not "/usr/share/doc"? This looks sort of weird, is that really intended? > [1] https://salsa.debian.org/med-team/yanagiba > [2] https://salsa.debian.org/med-team/unc-fish [3]: https://wiki.debian.org/HowToPackageForDebian#Lintian [4]: https://salsa.debian.org/med-team/unc-fish/-/jobs/1782347 [5]: https://salsa.debian.org/med-team/pilercr/-/blob/master/debian/tests/get-test-data Nilesh signature.asc Description: PGP signature
Autopkgtest for Yanagiba and unc-fish
Hello! I have written autopkgtests for yanagiba[1] and unc-fish[2]. The yanagiba package seems to fail the ci pipeline. The error seems to be from - dpkg: yanagiba-build-deps: dependency problems, but configuring anyway as you requested: yanagiba-build-deps depends on dh-python; however: Package dh-python is not installed. yanagiba-build-deps depends on python3-setuptools; however: Package python3-setuptools is not installed. yanagiba-build-deps depends on python3-pandas; however: Package python3-pandas is not installed. yanagiba-build-deps depends on python3-biopython; however: Package python3-biopython is not installed. yanagiba-build-deps depends on python3-nanomath; however: Package python3-nanomath is not installed. But these are already present as build dependencies in d/control. Kindly review and let me know how I can proceed. Thank you! Shruti [1] https://salsa.debian.org/med-team/yanagiba [2] https://salsa.debian.org/med-team/unc-fish ᐧ