On 20/03/16 03:12, Mattia Rizzolo wrote: >> The issues that you commented on above are fixed in this repository. >> >> 1: https://anonscm.debian.org/cgit/python-modules/packages/faker.git > > of what I wrote before these still need fixing: > > * debian/control: > + I: fake-factory source: duplicate-long-description python-fake-factory > python3-fake-factory > + I: python-fake-factory: extended-description-is-probably-too-short > + I: python3-fake-factory: extended-description-is-probably-too-short
I have added the Python version to the extended description, which fixes these lintian issues. > of the manpage erros you output, I can see man has to complain about > this: > > <standard input>:8: warning [p 1, 1.7i]: cannot adjust line > <standard input>:8: warning [p 1, 1.8i]: can't break line > <standard input>:33: warning [p 1, 5.3i, div `an-div', 0.0i]: cannot adjust > line > <standard input>:33: warning [p 1, 5.3i, div `an-div', 0.2i]: can't break line > <standard input>:33: warning [p 1, 5.3i, div `an-div', 0.3i]: cannot adjust > line > <standard input>:33: warning [p 1, 5.3i, div `an-div', 0.5i]: can't break line > > Indeed `faker --help` outputs this > > -l > {lt_LT,lv_LV,de_AT,tr_TR,sv_SE,no_NO,pt_PT,it_IT,en_US,en_CA,en,zh_TW,fi_FI,hi_IN,hr_HR,fa_IR,bg_BG,zh_CN,sk_SK,fr_FR,en_GB,dk_DK,es,en_AU,sl_SI,el_GR,la,de_DE,bs_BA,uk_UA,pt_BR,nl_NL,pl_PL,ru_RU,cs_CZ,ja_JP,es_ES,ko_KR,es_MX,ne_NP}, > --lang > {lt_LT,lv_LV,de_AT,tr_TR,sv_SE,no_NO,pt_PT,it_IT,en_US,en_CA,en,zh_TW,fi_FI,hi_IN,hr_HR,fa_IR,bg_BG,zh_CN,sk_SK,fr_FR,en_GB,dk_DK,es,en_AU,sl_SI,el_GR,la,de_DE,bs_BA,uk_UA,pt_BR,nl_NL,pl_PL,ru_RU,cs_CZ,ja_JP,es_ES,ko_KR,es_MX,ne_NP} > specify the language for a localized provider (e.g. > de_DE) > > well, that's awful. > did you manually fix this in the build manpage by any chance? Yes > imho the proper fix for this thing would be to have argparse (or > whatever thing this library/app is using to parse the options) to not > output the recognized options, and output them formatted in a saner way > (e.g. separated by ', ') at the bottom of the --help, then maybe > help2man will need some help to handle that correctly, though. > > And BTW, I have another feeling that the list of languages is stored > inside a set, and that is passed to argparse as a list of valid options; > well, that makes argparse output non-reproducible, and as such the > generated manpage is non reproducible; please make that either a real > list (so the elements have a fixed position) or sort the set before > passing it over to argparse, so that the output is reproducible between > runs (and as such between rebuilds). > https://paste.debian.net/plain/417202 (this is not built with the > reproducible toolchain, and as such there is noise about filesystem > order (actually there is not, but this is by chance) and timestamps > inside the {control,data}.tar member of the .deb, that in the > reproducible toolchain is fixed by a patched dpkg not yet included in > dpkg mainstream) I have made a pull request to the upstream project, improving the help output in a way that will fix these issues [1]. Its a bit awkward, but is the best step to take next, to wait for upstream to do a release with these changes? A possible alternative would be to include a manpage made with help2man, and not generate it for this version (and switch to generating it later)? 1: https://github.com/joke2k/faker/pull/342
signature.asc
Description: OpenPGP digital signature