Bug#784050: RFA: ocaml-deriving -- deriving functions from type declarations in OCaml
Package: wnpp Severity: normal Hello, Currently, ocaml-deriving has no human maintainers. It is maintained by the Debian OCaml team because of the need of transition coordination, but needs more love and a dedicated maintainer. A potential maintainer should get familiar with [1], and in particular with our policy and practices, and get subscribed to our mailing-list. [1] http://wiki.debian.org/Teams/OCamlTaskForce Cheers, -- Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150502141044.20046.95016.report...@zetta.zrh.le-gall.net
Bug#784042: RFA: ocaml-data-notation -- Store data using OCaml notation
Package: wnpp Severity: normal Hello, Currently, ocaml-data-notation has no human maintainers. It is maintained by the Debian OCaml team because of the need of transition coordination, but needs more love and a dedicated maintainer. A potential maintainer should get familiar with [1], and in particular with our policy and practices, and get subscribed to our mailing-list. [1] http://wiki.debian.org/Teams/OCamlTaskForce Cheers, -- Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150502124739.16844.39340.report...@zetta.zrh.le-gall.net
Bug#784039: RFA: ocaml-gettext -- OCaml internationalization library
Package: wnpp Severity: normal Hello, Currently, ocaml-gettext has no human maintainers. It is maintained by the Debian OCaml team because of the need of transition coordination, but needs more love and a dedicated maintainer. A potential maintainer should get familiar with [1], and in particular with our policy and practices, and get subscribed to our mailing-list. [1] http://wiki.debian.org/Teams/OCamlTaskForce Cheers, -- Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150502124348.15304.36016.report...@zetta.zrh.le-gall.net
Bug#784037: RFA: ocaml-inotify -- OCaml bindings for the inotify API
Package: wnpp Severity: normal Hello, Currently, ocaml-inotify has no human maintainers. It is maintained by the Debian OCaml team because of the need of transition coordination, but needs more love and a dedicated maintainer. A potential maintainer should get familiar with [1], and in particular with our policy and practices, and get subscribed to our mailing-list. [1] http://wiki.debian.org/Teams/OCamlTaskForce Cheers, -- Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150502124052.14785.74570.report...@zetta.zrh.le-gall.net
Bug#784036: RFA: ocamlgsl -- GNU scientific library for OCaml
Package: wnpp Severity: normal Hello, Currently, ocamlgsl has no human maintainers. It is maintained by the Debian OCaml team because of the need of transition coordination, but needs more love and a dedicated maintainer. A potential maintainer should get familiar with [1], and in particular with our policy and practices, and get subscribed to our mailing-list. [1] http://wiki.debian.org/Teams/OCamlTaskForce Cheers, -- Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150502123931.14516.70112.report...@zetta.zrh.le-gall.net
Bug#784035: RFA: ocamlify -- include files in OCaml code
Package: wnpp Severity: normal Hello, Currently, ocamlify has no human maintainers. It is maintained by the Debian OCaml team because of the need of transition coordination, but needs more love and a dedicated maintainer. A potential maintainer should get familiar with [1], and in particular with our policy and practices, and get subscribed to our mailing-list. [1] http://wiki.debian.org/Teams/OCamlTaskForce Cheers, -- Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150502123813.14178.93456.report...@zetta.zrh.le-gall.net
Bug#784033: RFA: gmetadom -- GDome2 DOM implementation
Package: wnpp Severity: normal Hello, Currently, gmetadom has no human maintainers. It is maintained by the Debian OCaml team because of the need of transition coordination, but needs more love and a dedicated maintainer. A potential maintainer should get familiar with [1], and in particular with our policy and practices, and get subscribed to our mailing-list. [1] http://wiki.debian.org/Teams/OCamlTaskForce Cheers, -- Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150502123451.13492.79443.report...@zetta.zrh.le-gall.net
Bug#784034: RFA: ocamlmod -- generate OCaml modules from source files
Package: wnpp Severity: normal Hello, Currently, ocamlmod has no human maintainers. It is maintained by the Debian OCaml team because of the need of transition coordination, but needs more love and a dedicated maintainer. A potential maintainer should get familiar with [1], and in particular with our policy and practices, and get subscribed to our mailing-list. [1] http://wiki.debian.org/Teams/OCamlTaskForce Cheers, -- Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150502123626.13836.10462.report...@zetta.zrh.le-gall.net
Bug#784032: RFA: xstr -- OCaml library for frequent string operations
Package: wnpp Severity: normal Hello, Currently, xstr has no human maintainers. It is maintained by the Debian OCaml team because of the need of transition coordination, but needs more love and a dedicated maintainer. A potential maintainer should get familiar with [1], and in particular with our policy and practices, and get subscribed to our mailing-list. [1] http://wiki.debian.org/Teams/OCamlTaskForce Cheers, -- Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150502123247.12956.98455.report...@zetta.zrh.le-gall.net
Bug#784031: RFA: camltemplate -- library for generating text from templates
Package: wnpp Severity: normal Hello, Currently, camltemplate has no human maintainers. It is maintained by the Debian OCaml team because of the need of transition coordination, but needs more love and a dedicated maintainer. A potential maintainer should get familiar with [1], and in particular with our policy and practices, and get subscribed to our mailing-list. [1] http://wiki.debian.org/Teams/OCamlTaskForce Cheers, -- Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150502122550.11366.46484.report...@zetta.zrh.le-gall.net
Re: RFC ounit 2.0.0, ocaml-expect 0.0.5, oasis 0.4.2
Anyone can have a look at this? Thanks Sylvain On 06-03-2014, Sylvain Le Gall wrote: > Hi all, > > I just pushed to git the latest version of ounit/ocaml-expect and oasis. > Can any Debian OCaml Maintainer, who should be less rusty than me, have > a look at this commit and tell me if it is ok to proceed with the > upload. > > http://anonscm.debian.org/gitweb/?p=pkg-ocaml-maint/packages/ounit.git > http://anonscm.debian.org/gitweb/?p=pkg-ocaml-maint/packages/ocaml-expect.git > http://anonscm.debian.org/gitweb/?p=pkg-ocaml-maint/packages/oasis.git > > Thanks, > Sylvain Le Gall > -- > Website: http://sylvain.le-gall.net/ > > Cheers, Sylvain Le Gall -- Website: http://sylvain.le-gall.net/ -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/slrnli1mhg.2rq.gil...@le-gall.net
RFC ounit 2.0.0, ocaml-expect 0.0.5, oasis 0.4.2
Hi all, I just pushed to git the latest version of ounit/ocaml-expect and oasis. Can any Debian OCaml Maintainer, who should be less rusty than me, have a look at this commit and tell me if it is ok to proceed with the upload. http://anonscm.debian.org/gitweb/?p=pkg-ocaml-maint/packages/ounit.git http://anonscm.debian.org/gitweb/?p=pkg-ocaml-maint/packages/ocaml-expect.git http://anonscm.debian.org/gitweb/?p=pkg-ocaml-maint/packages/oasis.git Thanks, Sylvain Le Gall -- Website: http://sylvain.le-gall.net/ -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/slrnlhfgoi.dsa.gil...@le-gall.net
Re: Please review: ocamlmod 0.0.4
On 26-06-2013, Hendrik Tews wrote: > Sylvain Le Gall writes: > >> >> - update man page (option -o, a single module file_) >> > >Not sure to understand that. > > ocamlmod -help describes option -o, which is missing in the man > page. > > Additionally, the man page says "a single module files", which > should be "a single module file". > >> + fix hyphen-used-as-minus-sign > > this is the lintian tag, see > http://lintian.debian.org/tags/hyphen-used-as-minus-sign.html > > What I don't understand about this tag is that lintian reports it > for pandoc generated man pages but not for my handwritten ones. > > BTW, I am glad you (seem to have) changed your mind and left your > name in Uploaders! > Unfortunately, I didn't really change my mind. This is one of the last upload I do -- and I need to have a human uploaders in this field. After the upload I should remove it and let someone else pick it up. However, I am thinking about setting myself as DM to still be able to help on particular topics. But you know, my retirement is more about honesty wrt to all the OCaml packages maintainers than because I want to leave. Staying in the group and thinking that I can do actual work would be a lie and I don't want to prevent other people to do a good job. Cheers, Sylvain Le Gall -- Website: http://sylvain.le-gall.net/ -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnkslg8s.rfh.gil...@le-gall.net
Re: Please review: ocamlmod 0.0.4
On 25-06-2013, Hendrik Tews wrote: > Hi, > > I don't know the standard, but I would > > - use debhelper compat level 9 > Done > - delete the cleaner line in gbp.conf > Done > - enable the tests (-configure --enable-tests) > They will fail, I need to fix them (although this is just a very minor problem). > - enable native compilation > This is not the default upstream. I keep a TODO for the next version. > - not call ocamlc setup -doc Done > > - update man page (option -o, a single module file_) > Not sure to understand that. > - fix most of the lintian warnings: > + fix Vcs fields Done > + update the copyright file Done > + add a watch file Done > + remove empty /usr/lib/OCaml Done. > + fix hyphen-used-as-minus-sign Don't understand this one. > + fix allows to in man page Done. > > For many of these points you can just copy from ocaml-fileutils. > > Bye, > > Hendrik > > Cheers, Sylvain Le Gall -- Website: http://sylvain.le-gall.net/ -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnksk9ok.rfh.gil...@le-gall.net
Re: Please review: ocamlmod 0.0.4
On 25-06-2013, Hendrik Tews wrote: > Hi, > > I don't know the standard, but I would > > - use debhelper compat level 9 > > - delete the cleaner line in gbp.conf > > - enable the tests (-configure --enable-tests) > > - enable native compilation > > - not call ocamlc setup -doc > > - update man page (option -o, a single module file_) > > - fix most of the lintian warnings: > + fix Vcs fields > + update the copyright file > + add a watch file > + remove empty /usr/lib/OCaml > + fix hyphen-used-as-minus-sign > + fix allows to in man page > How do I get this warnings ? Running on my system: $ lintian ../ocamlmod_0.0.4-1.dsc Gives me no warning at all $ lintian --version Lintian v2.5.13 > For many of these points you can just copy from ocaml-fileutils. > > Bye, > > Hendrik > > Cheers, Sylvain Le Gall -- Website: http://sylvain.le-gall.net/ -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnksk813.rfh.gil...@le-gall.net
Please review: ocamlify 0.0.2
Same reason as for ocamlmod 0.0.4, if someone can review the package before for upload. As an upstream, I did a fresh release so that we won't have any problem with OCaml 4 migration wrt setup.ml and Scanf bug. Cheers, Sylvain Le Gall -- Website: http://sylvain.le-gall.net/ -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnksk5np.rfh.gil...@le-gall.net
Please review: ocamlmod 0.0.4
Since I didn't do any packaging for a long time, I would like that someone review my latest commit to ocamlmod packages. This is a one liner in term of code (I have just injected the orig tarball and updated the changelog entry), but maybe can see if it meets the latest OCaml packaging standards. Thanks in advance, Sylvain Le Gall -- Website: http://sylvain.le-gall.net/ -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnksk349.rfh.gil...@le-gall.net
Bug#711452: liboasis-ocaml: Needs to depend on ocamlbuild in the META file.
Package: liboasis-ocaml Version: 0.3.0-1 Severity: normal Tags: patch Dear Maintainer, *** Please consider answering these questions, where appropriate *** * What led up to the situation? $ oasis setup -setup-update dynamic $ make ocaml setup.ml -build File "setup.ml", line 1, characters 0-1: Error: Reference to undefined global `Ocamlbuild_plugin' make: *** [build] Erreur 2 There is a missing dependency on ocamlbuild in the META file. The bug will be fixed upstream as well. The patch attached fix the bug. Regards Sylvain *** End of the template - remove these lines *** -- System Information: Debian Release: jessie/sid APT prefers unstable APT policy: (500, 'unstable') Architecture: amd64 (x86_64) Kernel: Linux 3.2.0-0.bpo.4-amd64 (SMP w/8 CPU cores) Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages liboasis-ocaml depends on: ii libc6 2.17-5 ii libfindlib-ocaml [libfindlib-ocaml-8p7u5] 1.3.1-1 ii libodn-ocaml [libodn-ocaml-9gxf8] 0.0.8-1 ii ocaml-base-nox [ocaml-base-nox-3.12.1] 3.12.1-4 liboasis-ocaml recommends no packages. Versions of packages liboasis-ocaml suggests: pn liboasis-ocaml-doc -- no debconf information diff --git a/src/oasis/META b/src/oasis/META index 88e7165..903ea3a 100644 --- a/src/oasis/META +++ b/src/oasis/META @@ -20,7 +20,7 @@ # OASIS_START -# DO NOT EDIT (digest: d471c9d7d5a1ce73f9fe3fdfbd92627e) +# DO NOT EDIT (digest: b0f33f22f3fba881e821ce61ae4d4b8b) version = "0.3.1" description = "_oasis file functions" requires = "unix odn" @@ -52,7 +52,7 @@ package "cli" ( package "builtin-plugins" ( version = "0.3.1" description = "_oasis file functions" - requires = "oasis oasis.base" + requires = "oasis oasis.base ocamlbuild" archive(byte) = "builtin-plugins.cma" archive(byte, plugin) = "builtin-plugins.cma" archive(native) = "builtin-plugins.cmxa"
Bug#711451: oasis: Depends on liboasis-ocaml-dev to enable dynamic mode.
Package: oasis Version: 0.3.0-1 Severity: normal Dear Maintainer, *** Please consider answering these questions, where appropriate *** * What led up to the situation? Setting up a project with: $> apt-get install oasis $> oasis setup -setup-update dynamic $> make ocaml setup.ml -build File "setup.ml", line 7, characters 0-16: Error: Unbound module OASISDynRun make: *** [build] Erreur 2 Solution: $> apt-get install liboasis-ocaml-dev Turns out that setup.ml is a toplevel script and that in order to use OASISDynRun, you need OASISDynRun.cmi. You can solve the problem by depending on liboasis-ocaml-dev or by moving the *.cmi of the -dev package to liboasis-ocaml. Regards Sylvain *** End of the template - remove these lines *** -- System Information: Debian Release: jessie/sid APT prefers unstable APT policy: (500, 'unstable') Architecture: amd64 (x86_64) Kernel: Linux 3.2.0-0.bpo.4-amd64 (SMP w/8 CPU cores) Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages oasis depends on: ii libfindlib-ocaml [libfindlib-ocaml-8p7u5] 1.3.1-1 ii liboasis-ocaml [liboasis-ocaml-iw5j4] 0.3.0-1 ii libodn-ocaml [libodn-ocaml-9gxf8] 0.0.8-1 ii ocaml-base-nox [ocaml-base-nox-3.12.1] 3.12.1-4 oasis recommends no packages. Versions of packages oasis suggests: pn liboasis-ocaml-doc -- no debconf information -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20130606223038.28073.53897.report...@zetta.zrh.le-gall.net
Re: ocaml-dangling-cmi for ocamlify generated files
2013/5/29 Hendrik Tews : > Hi, > > oasis 0.3 src/oasis/OASISData.mlify from which ocamlify generates > a .ml, which is then compiled. When installing OASISData.cmi in > the -dev package lintian reports a ocaml-dangling-cmi. The > trouble is that there is no .mli available or generated and the > generated .ml file is not readable. > > In order to deal with this warning, I would ask upstream to > supply a .mli file, either by writing it manually or by enhancing > ocamlify to generate it. One could of course also ignore or > override this tag for generated .ml files. That is a good point, I never thought about (as an upstream of oasis and ocamlify). So here you have an immediate solution to fix the problem: - run "ocamlbuild src/oasis/OASISData.inferred.mli" after building the project (ocaml setup.ml -build) - copy _build/src/oasis/OASISData.inferred.mli to src/oasis/OASISData.mli This way you fix the problem of the lintian report, without any upstream release. To solve the problem on a longer term, there are two solutions: 1. ocamlify should take as an input file OASISData.mli, with this content: var oasissys_ml: string (* source "OASISSys.ml" *) var oasissyslight_ml: string (* source "OASISSysLight.ml" *) ... 2. ocamlify generate the .mli and oasis look for the .mli or .inferred.mli in _build 2. is probably easier to implement but less readable than 1. Cheers Sylvain -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CAOCAUGPrhWWjcBcDUpt8G7axX8b=czuL-AFJoW-ZcgojD=q...@mail.gmail.com
Re: [OASIS-devel] updated oasis
2013/5/29 Hendrik Tews : > Hi, > > I imported a new upstream version of oasis and fixed several > points in the package. > I was planning to do it, but I am glad to see you pick the job. > I had to omit the library builtin-plugins.cma from the packaging > because of a peculiar dh_ocaml error. builtin-plugins.cma is > built via > > ocamlfind ocamlc -a -I +ocamlbuild ocamlbuildlib.cma ... -o > src/builtin-plugins.cma > > It therefore includes a copy of ocamlbuildlib.cma and dh_ocaml > says > > E: Error: unit Ocamlbuild_plugin exported in > liboasis-ocaml-dev/liboasis-ocaml v0.3.0-1 but already exported by > ocaml-nox/ocaml-base-nox v3.12.1-4 > E: Error running /usr/bin/ocaml-md5sums --md5sums-dir > debian/liboasis-ocaml-dev//var/lib/ocaml/md5sums --load-info > debian/liboasis-ocaml-dev.oinfo.debhelper dep at /usr/bin/dh_ocaml line 462. > > > I believe the inclusion of ocamlbuildlib.cma is an error in the > oasis build procedure. Note that builtin-plugins.cmxs is build > without a copy of ocamlbuildlib and is therefore included in the > package. > True, the error was on line 212 in _tags. This is fixed. I have pushed every fix to my repository on github: https://github.com/gildor478/oasis/commits/master Thanks for the report. Cheers Sylvain -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CAOCAUGNTm=gqoqbkmbrd8qbrn0mbe4yt__sptwmtnrtnng9...@mail.gmail.com
Time to retire?
Hi fellows Ocaml package maintainers, Writing this email is not really easy for me. Over the past couple of years, I wished to devote more time to Debian and my OSS activities. I must acknowledge that this is beyond what I can do. So rather than continue to lie and send false signals that I will resume my Debian activities, it is maybe time to retire. Tonight, I'll remove myself from the few packages I am still an uploader. I'll send the corresponding email to officialy retire in a couple of days. I will maybe provide a little help for packaging (as I continue to maintain some packages on my own server), but you should consider me as retired. If I can keep my account on alioth, I'll maybe do a few commit directly, ping me if you think this is not ok. Thanks to everyone for the great OCaml packages and the years of fun doing it together. Cheers, Sylvain Le Gall -- Website: http://sylvain.le-gall.net/ -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnkq6t29.905.gil...@le-gall.net
Re: OCaml transition plans
On 08-05-2013, Lifeng Sun wrote: > > > * packages with latest version compatible with ocaml-3.12 > package latest ver ocaml-expect 0.0.3 ocamlmod 0.0.3 ocaml-csv 1.2.6 ocaml-fileutils0.4.4 oasis 0.3.0 ounit 1.1.2 Will take care of these one. > * packages with latest version requires ocaml >=3D 4.00.0 > > package latest verbuild-dep ocaml-data-notation0.0.10 type-conv >=3D 108.07.01 Will take care of this one as well. > > > * packages without new upstream release > > camlbz2, ocamlify > There was probably an update of ocamlify in between, if not I'll do it in due time. Cheers, Sylvain Le Gall -- Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnkolb7h.57k.gil...@gallu.homelinux.org
Re: OCaml transition plans
On 08-05-2013, Stéphane Glondu wrote: > Hello, > > > [1] https://wiki.debian.org/Teams/OCamlTaskForce/OCamlTransition The page is 404. Cheers, Sylvain Le Gall -- Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnkolb2q.57k.gil...@gallu.homelinux.org
Re: Orphaning some of my packages
Hi, 2012/9/21 Hendrik Tews : > Sylvain Le Gall writes: > >Here is the list of packages that have no other uploaders >after this operation: > >* ocamlp3l > > I can find camlp3l and ocamlp3l, but no Debian package of that > name. Where is this? In the git repository only, I suppose. I just went through my "git checkout"ed packages, so maybe it has never been released. > > Sylvain, for the orphaned packages, could you comment on which > are actively maintained upstream? > > This is only IMHO, as a matter of fact an OCaml package that has not been updated for years doesn't mean it is orphaned upstream: Active upstream with releases: * camomile * ocaml-extunix * ocaml-sqlexpr * ounit * sexplib310 * uuidm Active upstream without many releases: * ocaml-benchmark * atdgen * biniou * caml2html * camlmix * cppo * easy-format * mikmatch * ocaml-atd * ocamlgsl Inactive or MIA upstream: * ocaml-deriving * ocaml-dbus * ocaml-inifiles * ocaml-inotify * tophide * xstr * xstrp4 * yojson * ocamlp3l Don't know: * gmetadom >* sexplib310 > > Sexplib and typeconf has been included into Jane Streets Core. Am > I right in assuming that the next janest-core packge will provide > sexplib310 and type-conf? > It remains separate packages. Jane Street provides (AFAIK), either a big archive containing all its libraries or a set of smaller archives (type-conv, binprot, sexplib, core...). Also, since they all have adopted the same numbering scheme (108.01), you can choose to migrate to the whole archive. > > To judge the importance of the orphaned packages, I made the > following table: > > Popcon Squeeze? reverse dep > > atdgen 9 > biniou 27 atdgen yojson > caml2html6 easy-format biniou ocaml-atd yojson atdgen > camlmix 1 caml2html easy-format biniou ocaml-atd > yojson atdgen > camomile 413 yes ocaml-batteries > cppo 5 > easy-format 1 biniou OCaml-atd yojson atdgen > gmetadom 14410 yes lablgtkmathview > mikmatch 5 > ocaml-atd9 > ocaml-benchmark 17 yes > ocaml-dbus 36 yes > ocaml-deriving 4 > ocaml-extunix 11 > ocamlgsl66 yes orpie > ocaml-inifiles 7 > ocaml-inotify 18 yes > ocaml-sqlexpr5 > ounit 225 yes cudf ocaml-expect ocaml-extunix > ocaml-reins > sexplib310 58 yes occinelle > tophide 5 > uuidm 24 yes > xstr70 yes > xstrp4 3 > yojson 26 atdgen > > > Many of these packages have a rather low popcon count, but many > of them are not in Squeeze, so their popularity might rise after > the release. Popcon count for OCaml packages, in general, is quite low. OCaml libraries remain in the long tail of package installed. I think this is related to the "static compilation" scheme of OCaml. Since you don't need to install a library along the program that uses it, you only count "dev machine" that will install your package. And OCaml Debian computer used for development around the world is probably a small percent of Debian computer using the libraries. Although, this is only my take on this. > > In the above list, only sexplib310, camomile, gmetadom and > ocamlgsl have reverse dependencies not in the list. That is only > removing sexplib310, camomile, gmetadom or ocamlgsl will affect > packages not in this list. > > Depending on time I probably have a look at one or the other > package. > Thx for the analysis. Sylvain -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CAOCAUGNMmB3T-GLvTgp76Z7nPpzp1WfC4GOVpFsVWcsit8=s...@mail.gmail.com
Orphaning some of my packages
Hi, It is not a big news that I am a lot less involved in Debian. I wish to focus on OASIS a lot more and in order to do that I need to give up on certain tasks. I remove myself as an uploader from most of my packages. Here is the list of packages that have no other uploaders after this operation: * atdgen * biniou * caml2html * camlmix * camomile * cppo * easy-format * gmetadom * mikmatch * ocaml-atd * ocaml-benchmark * ocaml-dbus * ocaml-deriving * ocaml-extunix * ocamlgsl * ocaml-inifiles * ocaml-inotify * ocamlp3l * ocaml-sqlexpr * ounit * sexplib310 * tophide * uuidm * xstr * xstrp4 * yojson I am still an active Debian user of these packages, but don't have enough time to maintain the Debian packaging seriously. If you think that some of these packages are not useful, just ask for their removal of the archive, but maybe keep their git repository. I will only continue to spend my packaging time on the following packages: * oasis * ocamlmod * ocamlify * ocaml-gettext * ocaml-fileutils * ocaml-data-notation If I ever wanted to create new packages, I'll submit an RFP and will probably provide a git repository for it with initial packaging done (like ocamlrss). But I will not upload it, so that the Debian OCaml packaging team will not have an extra burden. I have removed my name from most of the packages, but won't upload them until the freeze is over. Please CC me, I don't read frequently debian-ocaml-maint@l.d.o. Cheers Sylvain -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/caocaugpjjw0dhkvx0mfrm2ae0a96jmda0caga0sbqozrvsw...@mail.gmail.com
Re: oasis 0.3.0, tonight (GMT)
On 28-06-2012, Mehdi Dogguy wrote: > On 28/06/12 15:19, Sylvain Le Gall wrote: >> >> oasis 0.2 is quite old and has one important bug that will prevent >> any setup.ml generated with it to compile with OCaml 4.00. > > Good thing that Wheezy doesn't have OCaml 4.00.x. Whatever will bring > OCaml 4.00.x to Wheezy users can also bring oasis 3.0. > Seems you make your choice, so. I'll need your help tonight for a couple of package review (oasis 0.2.0-X+1 with Scanf patch ~1line, oasis 0.3.0-1 and ocamlmod). Will you be available ? Do we plan a meeting at a certain time tonight ? Cheers, Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnjuoog9.drt.gil...@gallu.homelinux.org
Re: oasis 0.3.0, tonight (GMT)
Hello, On 28-06-2012, Mehdi Dogguy wrote: > On 28/06/12 13:23, Sylvain Le Gall wrote: >> >> Does anybody have reasons to veto a new upstream so close to the >> freeze? >> > > I do wonder if doing this race is really a serious thing to do… > > Sid's oasis doesn't look broken and has been tested/used by our users. > Putting a new upstream release in sid so close to the freeze doesn't > sound a great idea to me, IMHO. To serve stable users interests, you may > upload oasis 3.0 in sid once Wheezy released and then upload it in > backports so that stable users can install it. For now, putting oasis > 3.0 in experimental seems to be our best option. IMHO, being a leaf > package doesn't justify an upload to sid. > oasis 0.2 is quite old and has one important bug that will prevent any setup.ml generated with it to compile with OCaml 4.00. Although, I won't argue on the short notice. If you feel more comfortable with an old version of oasis, it is ok for me. I just think it will be "almost" useless for any dev done with Wheezy, targetting OCaml 3.12 and 4.00. So here 2 options: 1. upload 0.3 to experimental and patch the OCaml 4.00 bug in oasis 0.2 2. take the risk of oasis 0.3 in sid (mitigated, it is not like we have thousands of critical users or possible security bugs or rev-deps). I'll propose you a package tonight and you'll just have to pick the best option for you. Cheers, Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnjuomek.drt.gil...@gallu.homelinux.org
oasis 0.3.0, tonight (GMT)
Hi all, I plan to release oasis 0.3.0 tonight. Since we are ultra-close to the freeze (and that my Debian packaging skills needs to be upgraded), is it possible to get a quick review of my packaging when I will be done tonight? Does anybody have reasons to veto a new upstream so close to the freeze? oasis should have not rev-deps so it won't create problems all around. I am also planning upgrades of the package requirement (hence ocamlmod/ocamlify which has no other reverse deps has well). Cheers, Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnjuofkv.drt.gil...@gallu.homelinux.org
Bug#670474: libpgocaml-ocaml-dev: Conflicts between extlib and camomile about UChar
Hi, 2012/4/27 Mehdi Dogguy > On 04/26/2012 02:07 AM, Sylvain Le Gall wrote: > > > > extlib and camomile both exports UChar and doesn't agree on this > > interface. This prevents to use pgocaml with any packages that uses > > batteries (e.g. sqlexpr). > > > > FWIW, I don't think building pgocaml with Batteries instead of extlib is > the right fix for this issue ; this does only move the problem elsewhere > (as you said in your report). So I'd rather leave this bugreport open > for now. > > If both libraries export a same module name but with different > signature/implementation, such clashes are expected. That's how OCaml is > designed. You can either re-assign this bug to both extlib *and* > camomille and have both upstream agree on a different module name than > UChar (or rename it in only one of the libraries)… or reassign to OCaml. > I guess the former is easier to fix than the latter :) > I submitted the bug upstream: http://code.google.com/p/ocaml-extlib/issues/detail?id=23 Though, since pgocaml 1.5 offers the possibility to switch, it is possible to switch... That is the reason why this bug is about pgocaml rather than extlib. Regards Sylvain Le Gall
Bug#670474: libpgocaml-ocaml-dev: Conflicts between extlib and camomile about UChar
Package: libpgocaml-ocaml-dev Version: 1.4-2+b1 Severity: important Dear Maintainer, extlib and camomile both exports UChar and doesn't agree on this interface. This prevents to use pgocaml with any packages that uses batteries (e.g. sqlexpr). Step to reproduce the bug: $ ocaml #use "topfind";; #thread;; #require "pgocaml";; #require "sqlexpr";; The files /usr/lib/ocaml/camomile/camomile.cma and /usr/lib/ocaml/extlib/extLib.cma disagree over interface UChar It seems possible to build pgocaml using batteries, starting with version 1.5. It will solve this conflict, but the conflict betweens extlib and camomile will remain. Cheers Sylvain Le Gall -- System Information: Debian Release: wheezy/sid APT prefers unstable APT policy: (500, 'unstable') Architecture: amd64 (x86_64) Kernel: Linux 2.6.32-5-amd64 (SMP w/3 CPU cores) Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages libpgocaml-ocaml-dev depends on: ii camlp4 [camlp4-3.12.1] 3.12.1-2 ii libc62.13-30 ii libcalendar-ocaml-dev [libcalendar-ocaml-dev-pm5s8] 2.03-1+b2 ii libcsv-ocaml-dev [libcsv-ocaml-dev-dk5h1]1.2.2-1+b1 ii libextlib-ocaml-dev [libextlib-ocaml-dev-jshx0] 1.5.2-1+b1 ii libpcre-ocaml-dev [libpcre-ocaml-dev-xu8a5] 6.2.5-1 ii libpcre3 1:8.30-4 ii libpgocaml-ocaml [libpgocaml-ocaml-8yz73]1.4-2+b1 ii ocaml-nox [ocaml-nox-3.12.1] 3.12.1-2 Versions of packages libpgocaml-ocaml-dev recommends: ii ocaml-findlib 1.2.8+debian-1 Versions of packages libpgocaml-ocaml-dev suggests: pn postgresql -- no debconf information -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120426000749.9984.63398.reportbug@localhost
Re: Help sought for compiling guestfs-browser
Hello, On 31-07-2011, Hilko Bengen wrote: > Hi, > > I am trying to compile guestfs-browser[1], at the moment because I want > to see if it's usable, but with the possible intent to package it. (I > have just uploaded libguestfs and various bindings to experimental.) > > I have installed everything needed by the configure script > (libcamomile-ocaml-data and libcamomile-ocaml-dev from experimental) and > that works nicely. However, I get a bunch of compiler errors, as shown > below[2]. > > Does anyone have an idea what might be going wrong? > > Cheers, > -Hilko > > [1] > http://people.redhat.com/~rjones/guestfs-browser/files/guestfs-browser-0.2.1.tar.gz > > [2] > > $ make > ocamlfind ocamldep -package > libvirt,guestfs,hivex,lablgtk2,extlib,xml-light,calendar,camomile,threads,bitstring,bitstring.syntax > -syntax bitstring cmdline.mli config.mli deviceSet.mli filetree_markup.mli > filetree.mli menu_about.mli menu_open_disk.mli menu_open_uri.mli > op_checksum_file.mli op_copy_regvalue.mli op_disk_usage.mli > op_download_as_reg.mli op_download_dir_find0.mli op_download_dir_tarball.mli > op_download_file.mli op_file_information.mli op_file_properties.mli > op_inspection_dialog.mli op_view_file.mli slave.mli slave_types.mli > slave_utils.mli utils.mli window.mli cmdline.ml deviceSet.ml > filetree_markup.ml filetree.ml main.ml menu_about.ml menu_open_disk.ml > menu_open_uri.ml op_checksum_file.ml op_copy_regvalue.ml op_disk_usage.ml > op_download_as_reg.ml op_download_dir_find0.ml op_download_dir_tarball.ml > op_download_file.ml op_file_information.ml op_file_properties.ml > op_inspection_dialog.ml op_view_file.ml slave.ml slave_types.ml > slave_utils.ml throbber.ml utils.ml w indow.ml | \ > /bin/sed -e :a -e '/ *\\$/N; s/ *\\\n */ /; ta' | \ > sort > .depend-t Can you give us the version of bitstring, ocaml and camomile? BTW, your command line seems wrong, it should be -syntax camlp4o rather than -syntax bitstring. > No level labelled ";" in entry "expr" > Failure: "Grammar.extend" > Preprocessing error on file cmdline.mli > No level labelled ";" in entry "expr" > Failure: "Grammar.extend" Cheers, Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnj3ckhi.9th.gil...@gallu.homelinux.org
Re: Automatic dependencies for shared modules
On 22-06-2011, Mehdi Dogguy wrote: > On 06/22/2011 10:30 PM, Stéphane Glondu wrote: >> Le 22/06/2011 22:27, Mehdi Dogguy a écrit : >>>> I'd like to upload an updated dh_ocaml package soon, so as to fix >>>> the experimental liquidsoap package. Let me know if this ok for the >>>> team. >>>> >>> >>> We don't have any major transition going on, so I don't see why we >>> should hold back this upload. So, please go ahead! >> >> Actually, there is the type-conv transition ongoing... but I don't >> foresee any issue in uploading dh-ocaml now. >> I don't think there will be any problem about that. > > That's why I said "major" :p > Hey!!! That is a major transition for me ;-) Cheers, Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnj04mr2.9th.gil...@gallu.homelinux.org
Re: oasis
Hello, On 02-06-2011, Daniel Bünzli wrote: > Hello Sylvain, > > I didn't get your answer (please keep me on the cc). > >> As far as oasis and uuidm is concerned, the migration to oasis is quite >> simple. I already crafted a tool to manage oasis package in debian >> (oasis2debian: > > So packagers only need an _oasis file describing the library and voilà ? > > The situation now is that I have an ad-hoc build script which has > various targets to build and install the libraries. My plan is to > remove these scripts from the distributions and replace them by an > _oasis file. Is that all what you need ? > Well, there are two parts: 1 - _oasis describes build dependencies, from this file a tool like oasis2debian can deduce build dependencies and what documentation is built 2 - setup.ml, generated from _oasis using "oasis setup", is a standard entry point in the build system. But all this remains helpers to write a debian packages from _oasis, the tool is still in development and IHMO a Debian packager will still need to review by hand the changes made by the tool. But I think, it is safe enough to say that _oasis + setup.ml is enough to have a good package. BTW, you can keep you custom scripts and call them through oasis using "BuildType: custom" "XCustomBuild: myscript.sh". Cheers, Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrniuh5q2.ks.gil...@gallu.homelinux.org
Re: oasis
Hello, On 30-05-2011, Daniel Bünzli wrote: > Hello, > > I will eventually update all my software packages to use oasis. Since > people already did package some of this software [1] I'd be glad to > hear from them on what is the most appropriate and simple way to work > for them given the oasis hypothesis (I plan to remove my ad-hoc > installation scripts). > > http://packages.debian.org/squeeze/libxmlm-ocaml-dev > http://packages.debian.org/squeeze/libuuidm-ocaml-dev > http://packages.debian.org/squeeze/libreact-ocaml-dev As far as oasis and uuidm is concerned, the migration to oasis is quite simple. I already crafted a tool to manage oasis package in debian (oasis2debian: http://anonscm.debian.org/gitweb/?p=pkg-ocaml-maint/packages/oasis2debian.git;a=summary ) For other people, you should have a look at debian/rules of oasis package: http://anonscm.debian.org/gitweb/?p=pkg-ocaml-maint/packages/oasis.git;a=blob;f=debian/rules;h=adbb6d7bb351adb37118bea0f4a0a35fd03194e4;hb=HEAD Right now, it uses override, but it should become a real buildsystem (as dh understand it). Cheers, Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrniu82ns.ks.gil...@gallu.homelinux.org
[VAC] 16th -> 24th April
Hello all, I will be on vacation for a week. All my packages are team maintained, see with debian-ocaml-maint@l.d.o for NMU. No internet access at all because it will be real vacation ;-) Cheers Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110411221431.gi3...@yocto.gallu.homelinux.org
Bug#620838: dh-ocaml: dh_ocaml doesn't depend on -dev packages even when runtime package requires it
Package: dh-ocaml Version: 0.9.6 Severity: normal While building the package oasis, I saw a difference in liboasis-ocaml and liboasis-ocaml-dev dependencies list: ocamlgraph and fileutils are missing from the runtime package. It should be related to the fact that they don't ship runtime packages (pure ocaml packages). However, the runtime package of liboasis-ocaml contains *.cma libraries that really need the libraries provided by libocamlgraph-ocaml-dev and libfileutils-ocaml-dev. Cheers Sylvain Le Gall -- System Information: Debian Release: wheezy/sid APT prefers unstable APT policy: (500, 'unstable'), (1, 'experimental') Architecture: amd64 (x86_64) Kernel: Linux 2.6.26-2-vserver-amd64 (SMP w/2 CPU cores) Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/bash dh-ocaml depends on no packages. Versions of packages dh-ocaml recommends: ii debhelper 8.1.2 helper programs for debian/rules ii ocaml-nox 3.11.2-4 ML implementation with a class-bas Versions of packages dh-ocaml suggests: ii git [git-core] 1:1.7.4.1-5 fast, scalable, distributed revisi ii git-core 1:1.7.4.1-5 fast, scalable, distributed revisi -- no debconf information -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110404151613.15999.23157.report...@yocto.gallu.homelinux.org
Re: Transition to OCaml 3.12.0...
Hello, On 10-03-2011, David MENTRE wrote: > Hello Stéphane, > > 2011/3/10 Stéphane Glondu : >> After #613848 is done. > > Could we have a message on this list when the transition is started, > so has to avoid upgrading or installing packages during it? Thank you. What would be really great, would be to have a state per ocaml package telling you which one you should not upgrade in fact and which one you should upgrade. For example, I am always afraid to broke current transition by uploading new package versions. I think it should be ok to upload "leaf" packages, but I am not sure ("leaf" is e.g. ocaml-sqlexpr). Cheers, Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrninjtik.81n.gil...@gallu.homelinux.org
Re: dh and ocamlvars.mk
Hello, On 26-02-2011, Eric Cooper wrote: > I'm changing the ocaml packages I maintain from cdbs to dh. > > The comment at the top of /usr/share/ocaml/ocamlvars.mk makes it sound > cdbs-specific, but the actual variable definitions would be useful for > dh as well. Can I assume that it will remain safe to include this > file for dh as well as cdbs debian/rules files? > It was cdbs specific, but should be useable now with dh as well. Do you know that there is a "--with ocaml" for dh 7 ? Cheers, Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnimj1cb.81n.gil...@gallu.homelinux.org
Bug#615260: libinifiles-ocaml-dev: missing html documentation
Package: libinifiles-ocaml-dev Version: 1.2-1 Severity: normal We should create and install the HTML documentation. Cheers Sylvain Le Gall -- System Information: Debian Release: wheezy/sid APT prefers unstable APT policy: (500, 'unstable') Architecture: amd64 (x86_64) Kernel: Linux 2.6.32-5-amd64 (SMP w/3 CPU cores) Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages libinifiles-ocaml-dev depends on: ii libinifiles-ocaml [libinifile 1.2-1 read and write .ini files for OCam ii libpcre-ocaml-dev [libpcre-oc 6.0.1-3OCaml bindings for PCRE (Perl Comp ii ocaml-nox [ocaml-nox-3.11.2] 3.11.2-2 ML implementation with a class-bas Versions of packages libinifiles-ocaml-dev recommends: ii ocaml-findlib 1.2.6+debian-1 management tool for OCaml librarie libinifiles-ocaml-dev suggests no packages. -- no debconf information -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110226150709.5785.10006.reportbug@localhost
Re: Bug#613848: Transition to OCaml 3.12.0...
On 18-02-2011, Stéphane Glondu wrote: > Le 10/02/2011 10:12, Sylvain Le Gall a écrit : >> I also have a pair of packages to update (sexplib, type-conv, bin-prot, >> ounit). Since some of them will break the distribution and should >> trigger the need to binNMU some other packages, I am not sure how to >> proceed: 1) upload new version during the transition, 2) do it before >> and ask for binNMU or 3) do it before and don't ask for binNMU > > I've just realized that the last upstream release of type-conv and > sexplib310 don't compile with ocaml 3.11.2. So for these two, we have no > choice, we have to do 1). > > I've updated sexplib310 packaging in git, and upstream changes look > pretty invasive (we are several versions behind). Work has to be done to > make sure all rdeps build, and I'd rather stick to the version currently > in unstable if this work is not done before we start the OCaml transition. > > The latest upstream of type-conv seems to be already packaged in git. > Sylvain, what are your feeling about updating it during OCaml > transition? If you are unsure, I've got a patch to make the version > currently in unstable compile with OCaml 3.12. I am ok for an upload during 3.12 transition of sexplib/type-conv, rdeps are not that big. It will only delay janest-core AFAIK. > > ounit is already in experimental and seems ready to be part of #613848 > (but I'd like to have confirmation). I can upload it to unstable, it is in a good shape. Cheers, Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnilsj6r.krm.gil...@gallu.homelinux.org
Re: build-dep-graph and ocaml-debian-status disabled
On 16-02-2011, Stefano Zacchiroli wrote: > > Prodded by the alioth admins due to disk space utilization, I discovered > that I was still maintaining in my homepage a couple of services linked >=66rom the OCaml team (wiki) page: > > - build-dep-graph (the tool in our svn at /trunk/tools/build-dep-graph) > - ocaml-debian-status (ditto at /trunk/tools/ocaml-debian-status) > > I've no idea whether they were still used or not (and I didn't have time > to check up to now), but for sure they are disabled since this morning > and I've removed them from my home dir. The code is still safe in SVN > though, if you discover they are needed, it would be nice to setup them > again in the home dir of the pkg-ocaml-maint project (where I should > have set up them in the very beginning). All in all, I've the impression > they are no longer needed since quite some time, superseded by the > transition monitor. > > Cheers. > > PS please Cc:-me on reply *if* you want to get my attention I setup the build-dep-graph stuff last year: gildor@alioth:~/build-dep-graph$ crontab -l # m h dom mon dow command 05 00 * * * (cd $HOME/build-dep-graph && svn up ; make update install) > /dev/null Seems we were overwriting each other stuff since one year. At least there is nothing to change after you stop your script, for this part. Concerning the ocaml-debian-status, I think the transition monitor is made for this task. Cheers, Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrniln6ma.krm.gil...@gallu.homelinux.org
Re: Transition to OCaml 3.12.0...
On 10-02-2011, Ralf Treinen wrote: > Hello, > > On Mon, Feb 07, 2011 at 03:14:52PM +0100, Stéphane Glondu wrote: > >> All packages not in the table should be safely binNMU-ed. The table >> summarizes work that will need to be done manually. Please have a >> look, check your packages, and put your name in the slots you want >> to handle yourself. > > I have some packages that would be binNMU-ed, but that I would like to > update before starting the process. What would be the timeframe for > doing that? If we could have at least one week or so before the migration > starts that would be great. > I also have a pair of packages to update (sexplib, type-conv, bin-prot, ounit). Since some of them will break the distribution and should trigger the need to binNMU some other packages, I am not sure how to proceed: 1) upload new version during the transition, 2) do it before and ask for binNMU or 3) do it before and don't ask for binNMU 1) let us use the transition binNMU for ocaml 3.12 upgrade and package upgrade which makes sense because some new versions are here to fix 3.12 build. But with 1), if something fails we will have a longer 3.12 upgrade. 2) should be the standard path outside the 3.12 upgrade but can delay significantly the start of the upgrade 3) will let unstable in a broken state until the transition finished... What should I (we) do? Cheers, Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnil7av3.981.gil...@gallu.homelinux.org
Re: Transition to OCaml 3.12.0...
Hi, On 08-02-2011, Stéphane Glondu wrote: > Le 08/02/2011 09:02, Mehdi Dogguy a écrit : > >> AFAIK, other packages FTBFSing have patches too... so, it should be easy >> to get around those build failures. (Unless I misunderstood something). > > Could you point me to those patches? I agree, some of the FTBFSes can be > easily fixed, but ocamlgsl one (for example) doesn't seem obvious to me, > and I don't have time to investigate it myself. > The ocamlgsl package is already a PAS and should not probably not be built on armel. See the error in the build log: #error "Architectures with double-word alignment for doubles are not supported" But maybe you are referring to something else. Cheers, Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnil22dt.981.gil...@gallu.homelinux.org
Bug#611213: unison-gtk: Missing dependency on unison package
Hello, On Wed, Jan 26, 2011 at 09:14:35PM +0100, Dominique Dumont wrote: > Package: unison-gtk > Version: 2.32.52-1 > Severity: normal > > Hello > > Looks like unison-gtk is not usable if unison is not installed. > But unison is not listed in the dependency list. > > Could you add this dependency ? > That should not happen, because unison-gtk is unison + gtk interface. Could you give me the output that lead you to this conclusion ? N.B. if you use unison with ssh, unison must be installed on the target computer as well. Cheers Sylvain -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110126212142.ge27...@yocto.gallu.homelinux.org
Slow retirement
Hi all fellow pkg-ocaml-maintainers, Since a few days, I am considering to retire from various extra activities I had, like Debian. The reasons are quite simple: I want to spend more time with my family and I just buy a house that will need some work. My first child was born 3 years ago, and I have been struggling to keep my Debian activities at an acceptable level since then. This is not a formal retirement, I will still be around for at least 6 months. However, I would like that other people begin to take care of some of my important packages, esp. unison. We will have time to talk about this after squeeze release. Cheers, Sylvain Le Gall -- My company: http://www.ocamlcore.com Linkedin: http://fr.linkedin.com/in/sylvainlegall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrniib9dp.118.gil...@gallu.homelinux.org
Re: List of FTBFS in Ubuntu
On 03-12-2010, Stéphane Glondu wrote: > This is a multi-part message in MIME format. > --030406090201010307010801 > Content-Type: text/plain; charset=UTF-8 > Content-Transfer-Encoding: quoted-printable > > Le 03/12/2010 16:24, Sylvain Le Gall a =C3=A9crit : >>> There is a problem with inversion of flags for library in pcre-ocaml. = > I [...] >> (the patch is from Magnus Therning) > > I independently came up with the attached patch against the version in > git (didn't check yours), and checked that cameleon, cduce, ceve, galax, > json-wheel, matita, ocsigen and pkglab build successfully. I pushed it > for reference; however, the next sid version will (most likely) be a new > upstream release. Has a patch been forwarded upstream, BTW? > I don't think so. I was part of the discussion and probably asked Magnus, which is the patch's author and Arch Linux packager, to forward it -- though I don't remember precisely. Maybe you can contact Magnus about this and see if your patch apply to the version in Arch Linux. AFAIK, the version in Arch is the latest version and the bug is still there. Regards Sylvain -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnifj1ul.r67.gil...@gallu.homelinux.org
Re: List of FTBFS in Ubuntu
On 03-12-2010, Sylvain Le Gall wrote: > On 03-12-2010, Ralf Treinen wrote: >> On Fri, Dec 03, 2010 at 08:43:23AM +0100, Lucas Nussbaum wrote: >> >>> >From time to time, I do archive rebuilds for Ubuntu too, and I thought >>> that the results would be interesting for DDs as well, since a FTBFS in >>> Ubuntu might indicate that the package will FTBFS in the future in >>> Debian, due to toolchain changes for example. >> >>> http://udd.debian.org/cgi-bin/ubuntu_ftbfs.cgi. >> >> Thanks, Lucas, that is useful. >> >> There are quite some problems related to pcre (several ocaml packages >> are bitten by that, but not only). Any idea to what this may be due? >> > > There is a problem with inversion of flags for library in pcre-ocaml. I > had a discussion about this issue with Arch Linux pcre-ocaml (and oasis) > maintainer. > On 06/11/10 21:10, Sylvain Le Gall wrote: > On Sat, Nov 06, 2010 at 08:54:13PM +, Magnus Therning wrote: > > >> On 06/11/10 17:34, Sylvain Le Gall wrote: >> I found a rather hackish way to fix it. 'ocamlobjinfo' now says >> >> >> >> >> >>Extra C object files: -lpcre_stubs -L/usr/lib -lpcre -lpcre_stubs >> >> >> >> >> >> Not the clean fix I would have liked, but it soves the problem. :-) >> >> >> >> >> >> Now I only have some minor clean-ups and then I can update OASIS and its >> >> >> dependencies. >> >> >> >> >> >> Thanks for your help with this! >> >> >> >> >> > > > > Do you fix it in the pcre-ocaml package? or just for ocaml-expect build? >
Re: List of FTBFS in Ubuntu
On 03-12-2010, Ralf Treinen wrote: > On Fri, Dec 03, 2010 at 08:43:23AM +0100, Lucas Nussbaum wrote: > >> >From time to time, I do archive rebuilds for Ubuntu too, and I thought >> that the results would be interesting for DDs as well, since a FTBFS in >> Ubuntu might indicate that the package will FTBFS in the future in >> Debian, due to toolchain changes for example. > >> http://udd.debian.org/cgi-bin/ubuntu_ftbfs.cgi. > > Thanks, Lucas, that is useful. > > There are quite some problems related to pcre (several ocaml packages > are bitten by that, but not only). Any idea to what this may be due? > There is a problem with inversion of flags for library in pcre-ocaml. I had a discussion about this issue with Arch Linux pcre-ocaml (and oasis) maintainer. Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnifi2hc.r67.gil...@gallu.homelinux.org
Bug#605741: ITP: caml2html -- HTML and LaTeX colored syntax from OCaml source files
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: caml2html Version : 1.4.1 Upstream Author : Martin Jambon * URL : http://martin.jambon.free.fr/caml2html.html * License : GPL-2 Programming Lang: OCaml Description : HTML and LaTeX colored syntax from OCaml source files Caml2html provides a command-line executable which converts a set of OCaml source files into a HTML or LaTeX document with colored syntax. A library is also provided for building web-page generators that would color OCaml code appropriately. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101202224132.32344.62165.report...@zetta.gallu.homelinux.org
Bug#605683: ITP: ocaml-sqlexpr -- type-safe access to SQL DB in OCaml
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: ocaml-sqlexpr Version : 0.2.3 Upstream Author : Mauricio Fernandez * URL : http://github.com/mfp/ocaml-sqlexpr * License : LGPL-2.1 with OCaml linking exception Programming Lang: OCaml Description : type-safe access to SQLite DB in OCaml Minimalistic library and syntax extension for type-safe, convenient execution of SQL statements. Currently compatible with Sqlite3. . Sqlexpr features: * automated prepared statement caching, param binding, data extraction, error checking (including automatic stmt reset to avoid BUSY/LOCKED errors in subsequent queries), stmt finalization on db close, etc. * HOFs like iter, fold, transaction * support for different concurrency models: everything is functorized over a THREAD monad, so you can for instance do concurrent folds/iters with Lwt * support for SQL stmt syntax checks and some extra semantic checking (column names, etc) -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101202113319.21761.95696.report...@zetta.gallu.homelinux.org
Bug#605682: ITP: ocaml-deriving -- deriving functions from type declarations in OCaml
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: ocaml-deriving Version : 0.1.1a Upstream Author : Jeremy Yallop * URL : http://code.google.com/p/deriving/ * License : MIT Programming Lang: OCaml Description : deriving functions from type declarations in OCaml Camlp4 extension to OCaml for deriving functions from type declarations. Includes derivers for pretty-printing, type-safe marshalling with structure-sharing, dynamic typing, equality, and more. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101202112747.21519.42292.report...@zetta.gallu.homelinux.org
Bug#605681: ITP: yojson -- JSON library for OCaml
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: yojson Version : 0.8.1 Upstream Author : Martin Jambon * URL : http://martin.jambon.free.fr/yojson.html * License : BSD3 Programming Lang: OCaml Description : JSON library for OCaml Yojson is an optimized parsing and printing library for the JSON format. It addresses a few shortcomings of json-wheel including 3x speed improvement, polymorphic variants and optional syntax for tuples and variants. . It is a replacement for json-wheel (libjson-wheel-ocaml-dev). -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101202112053.21353.22031.report...@zetta.gallu.homelinux.org
Bug#605680: ITP: easy-format -- text generation with indentation made easy(ier) for OCaml
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: easy-format Version : 1.0.0 Upstream Author : Martin Jambon * URL : http://martin.jambon.free.fr/easy-format.html * License : BSD3 Programming Lang: OCaml Description : text generation with indentation made easy(ier) for OCaml This module offers a simplified interface to the Format module of the standard library. Input data must be converted into a tree using 3 kinds of nodes: atoms, lists and labelled nodes. Each node is bound to its own formatting parameters and a single function call produces the formatted output. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101202111757.21306.45343.report...@zetta.gallu.homelinux.org
Bug#605677: ITP: cppo -- Cpp for OCaml
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: cppo Version : 0.9.0 Upstream Author : Martin Jambon * URL : http://martin.jambon.free.fr/cppo.html * License : BSD3 Programming Lang: OCaml Description : Cpp for OCaml Cppo is an OCaml-friendly implementation of cpp, the C preprocessor. It can replace camlp4 for preprocessing. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101202111507.21261.25651.report...@zetta.gallu.homelinux.org
Bug#605675: ITP: camltemplate -- configurable library for generating text from templates in OCaml
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: camltemplate Version : 1.0.2 Upstream Author : Benjamin Geer * URL : http://camltemplate.forge.ocamlcore.org/ * License : GPL-2+ Programming Lang: OCaml Description : configurable library for generating text from templates in OCaml CamlTemplate is library for generating text from templates in OCaml. It can be used to generate web pages, scripts, SQL queries, XML documents and other sorts of text. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/2010120253.21155.17807.report...@zetta.gallu.homelinux.org
Bug#605674: ITP: camlmix -- preprocessor which converts text with embedded OCaml
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: camlmix Version : 1.3.0 Upstream Author : Martin Jambon * URL : http://martin.jambon.free.fr/camlmix/ * License : BSD3 Programming Lang: OCaml Description : preprocessor which converts text with embedded OCaml Camlmix is a generic preprocessor which converts text with embedded OCaml into an OCaml program with embedded text. It produces text documents from one or several templates. OCaml toplevel statements are inserted between '## ... ##', and OCaml string expressions between '##= ... ##'. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101202110447.19888.12933.report...@zetta.gallu.homelinux.org
Bug#605672: ITP: biniou -- Flexible binary data format in OCaml
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: biniou Version : 0.9.1 Upstream Author : Martin Jambon * URL : http://martin.jambon.free.fr/biniou.html * License : BSD3 Programming Lang: OCaml Description : Flexible binary data format in OCaml Biniou is a binary data format designed for speed, safety, ease of use and backward compatibility as protocols evolve. Biniou is vastly equivalent to JSON in terms of functionality but allows implementations about 4 times as fast (see godi-yojson for comparison), with 25-35% space savings. Biniou data can be decoded into human-readable form without knowledge of type definitions except for field and variant names which are represented by 31-bit hashes. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101202110121.19529.31069.report...@zetta.gallu.homelinux.org
Bug#605671: ITP: atdgen -- Code generator for biniou and JSON serialization
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: atdgen Version : 1.0.1 Upstream Author : Martin Jambon * URL : http://oss.wink.com/atdgen/ * License : BSD3 Programming Lang: OCaml Description : Code generator for biniou and JSON serialization Atdgen is a command-line program that takes as input type definitions in the ATD syntax and produces OCaml code suitable for data serialization and deserialization. Two data formats are currently supported, these are biniou and JSON. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101202105843.19472.99478.report...@zetta.gallu.homelinux.org
Bug#605670: ITP: atd -- Syntax for cross-language data types in OCaml
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: atd Version : 0.9.2 Upstream Author : Martin Jambon * URL : http://oss.wink.com/atd/ * License : BSD3 Programming Lang: OCaml Description : Syntax for cross-language data types in OCaml ATD stands for Adjustable Type Definitions. It is a type definition language designed to accomodate a variety of programming languages and data formats by the means of target-specific annotations. It supports sum types, parametrized types and inheritance. The library provides a parser and other tools useful for manipulating ATD type definitions. The reference manual gives a complete description of the syntax. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101202105611.19337.78289.report...@zetta.gallu.homelinux.org
Bug#605669: ITP: tophide -- Hides toplevel values whose name starts with an underscore.
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: tophide Version : 1.0.0 Upstream Author : Martin Jambon * URL : http://martin.jambon.free.fr/ocaml.html * License : BSD3 Programming Lang: OCaml Description : Hides toplevel values whose name starts with an underscore. Tophide hides toplevel values whose name starts with an underscore. This is useful for some Camlp4 syntax extensions that produce lots of global identifiers that should remain hidden. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101202105226.19295.46431.report...@zetta.gallu.homelinux.org
Bug#605652: ITP: ocaml-inifiles -- read and write .ini for OCaml
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: ocaml-inifiles Version : 1.2 Upstream Author : Eric Stokes * URL : http://homepage.mac.com/letaris/ * License : LGPL-2.1+ Programming Lang: OCaml Description : read and write .ini for OCaml This library allow to read and write .ini files. It features an object oriented interface to manipulate inifiles. It allows sections listing and operation on several inifiles grouped in a directory. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101201161318.30245.24704.report...@localhost
Bug#605635: ITP: mikmatch -- camlp4 extension for pattern matching with regexps
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: mikmatch Version : 1.0.2 Upstream Author : Martin Jambon * URL : http://martin.jambon.free.fr/micmatch.html * License : BSD3 Programming Lang: OCaml Description : camlp4 extension for pattern matching with regexps Mikmatch provides enhanced pattern matching with regexps for OCaml. . The goal of Mikmatch is to make text-oriented programs even easier to write, read and run without losing the unique and powerful features of OCaml. Mikmatch provides a concise and highly readable syntax for regular expressions, and integrates it into the syntax of OCaml thanks to Camlp4. . The implementation of Mikmatch consists essentially of: * a library which is loaded by the OCaml preprocessor (Camlp4) and defines sophisticated "macros", i.e. the modified syntax; * a traditional library (runtime) which is required by the programs that use the Mikmatch syntax; * a dedicated 'mikmatch' command which can be used as a replacement for 'ocaml' in scripts or as an interactive toplevel. It performs automatically these steps: preprocessing, compilation and execution. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101202001634.3383.39917.report...@yocto.gallu.homelinux.org
Bug#605634: ITP: ocaml-inifiles -- read and write .ini files for OCaml
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: ocaml-inifiles Version : 1.2 Upstream Author : Eric Stokes * URL : http://homepage.mac.com/letaris/ * License : LGPL-2.1 Programming Lang: OCaml Description : read and write .ini files for OCaml This library allow to read and write .ini files. It features an object oriented interface to manipulate inifiles. It allows sections listing and operation on several inifiles grouped in a directory. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101202001232.3259.92799.report...@yocto.gallu.homelinux.org
Bug#605575: ITP: xstrp4 -- camlp4 extension that expands brace expansions in OCaml string
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: xstrp4 Version : 1.8 Upstream Author : Gerd Stolpmann * URL : http://projects.camlcity.org/projects/xstrp4.html * License : MIT Programming Lang: OCaml Description : camlp4 extension that expands brace expansions in OCaml string This camlp4 syntax extension interprets the dollar notation ${name} in strings and in included files. . It can: * include whole file in your OCaml code * define a format '%x' conversion to display variables * interpolate '$x' as well as '${x} * take into account record field and module names -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101201130634.14022.54375.report...@localhost
Bug#604636: libcamlimages-ocaml-dev: does not install oFreetype.cmi
On Tue, Nov 23, 2010 at 11:47:28AM +0200, ygrek wrote: > Package: libcamlimages-ocaml-dev > Version: 1:3.0.1-5+b3 > Severity: normal > > oFreetype.cmi is created during build but is not installed. So OFreetype > is present in cma but is not accessible because of cmi absence. > > $ ls src/oFreetype.* > src/oFreetype.cmi > src/oFreetype.cmo > src/oFreetype.cmx > src/oFreetype.ml > src/oFreetype.o > > $ ls /usr/lib/ocaml/camlimages/oFreetype.* > ls: cannot access /usr/lib/ocaml/camlimages/oFreetype.*: No such file or > directory > > Maybe there are some other modules with this problem.. > Upstream author seems to distribute freetype.mli to access freetype. Maybe he wants to keep oFreetype internal. Cheers Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101123101715.gp20...@yocto.gallu.homelinux.org
Re: names of distribution-branches in the git repository
On 23-11-2010, Ralf Treinen wrote: > On Mon, Nov 22, 2010 at 02:15:27PM +0000, Sylvain Le Gall wrote: >> On 22-11-2010, Mehdi Dogguy wrote: >> > On 22/11/2010 14:15, Sylvain Le Gall wrote: >> >> >> >> May I suggest my own summary, about this feature (trying to solve >> >> conflicts between all opinions): >> > >> > I think that we all agreed on Stéphane's proposal (at least, those >> > who raised their voice here). Didn't we? :) >> >> The only thing, I had against Stéphane proposal was about */upstream. >> But as you see in the summary, I change my mind about that. > > I didn't. > > All changes to the upstream branch are local, and are not supposed to be > pushed to the central repository (which usually means that they are > only temporary and should not even be comitted). When I do changes > on upstream then it is to try changes, and when I am satisfied I > put them into a quilt patch on the the debian branch and undo them on > my upstream branch. The $distrib/upstream will only contain change made by upstream, not by you. We just keep the same scheme as with master and upstream but git-buildpackage need to have an upstream branch as a reference. Example: - git-import-orig in experimental/master will commit upstream tarball in experimental/upstream - if you made change to something outside debian/ in experimental/master you can store it into debian/patches Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnien4ad.r67.gil...@gallu.homelinux.org
Re: names of distribution-branches in the git repository
On 22-11-2010, Mehdi Dogguy wrote: > On 22/11/2010 15:15, Sylvain Le Gall wrote: > > I'm not sure it's relevant then. I would not keep the history if I'm > reworking from scratch. ymmv > >>> I'm not sure that we should enforce anything on how/from what the >>> branches were created here. >>> >>>> - set debian/gbp.conf accordingly (upstream-branch and >>>> debian-branch) >>>> >>> >>> Why? or did you forget to add "in $distrib/debian"? >> >> I mean that we should add/change upstream-branch = upstream to >> upstream-branch = $distrib/upstream >> >> and debian-branch = master to debian-branch = $distrib/master >> > > Yeah, I did understand that. But, I wouldn't not commit that on "master". > The default remains "sid", IMHO. I am talking about a commit in $distrib/master not master. Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrniel5pj.r67.gil...@gallu.homelinux.org
Re: names of distribution-branches in the git repository
On 22-11-2010, Mehdi Dogguy wrote: > On 22/11/2010 14:15, Sylvain Le Gall wrote: >> >> May I suggest my own summary, about this feature (trying to solve >> conflicts between all opinions): > > I think that we all agreed on Stéphane's proposal (at least, those > who raised their voice here). Didn't we? :) The only thing, I had against Stéphane proposal was about */upstream. But as you see in the summary, I change my mind about that. Stéphane's proposal seems reasonable to me. > >> - in the Debian OCaml reference: >> >> - if we package something for anything else than sid: >> >> - create a set of branch $distrib/master, $distrib/upstream forked >> from the master, upstream branches (E.g. squeeze/master, >> squeeze/upstream) >> > > Why adding the condition "forked from"? What if I want to rework my > packaging from scratch? and What if the new upstream release has nothing > to do with previous releases? It $distrib < sid, then ok… but in this > case, it should be forked from "debian/$distrib's_version" and > "upstream/$distrib's_version" IMHO. > Forked from scratch -> you can work from scratch with history. Even if you delete everything, you can keep history of this big change... > I'm not sure that we should enforce anything on how/from what the branches > were created here. > >> - set debian/gbp.conf accordingly (upstream-branch and debian-branch) >> > > Why? or did you forget to add "in $distrib/debian"? I mean that we should add/change upstream-branch = upstream to upstream-branch = $distrib/upstream and debian-branch = master to debian-branch = $distrib/master > >> - the pristine-tar branch remains the same > > pristine-tar branch should not be branched… since it's just a directory > with known versions (as I see it). > >> - in dom-git-checkout: >> >> - if */master exists, display an information note listing other >> branches. E.g.: You checkout master but experimental/master and >> squeeze/master also exists >> >> - add an option (-t $distrib) to fetch/checkout $distrib/upstream and >> $distrib/master, with a special keyword "all" that ifetch all $distrib > > Seems reasonable for me. > >> - in dom-new-git-repo: >> >> - add an option (-t $distrib) to fork upstream and master branches >> remotely and check them out locally. >> > > How do you fork from something that doesn't exist? (e.g. master). Is this > option relevant here? > My idea is that dom-new-git-repo should do the forks for us. Maybe we should not use the script dom-new-git-repo for that. Maybe we can use a script called dom-fork-git-repo or something similar... Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrniekunv.r67.gil...@gallu.homelinux.org
Re: names of distribution-branches in the git repository
On 20-11-2010, Ralf Treinen wrote: > Hi, > > these days we are (or ar least we should be) mostly using the experimental > distribution for uploads. In that case it seems reasonable to put > modifications on a git branch that is specific to that distribution. > This of course also applies to other branches, like for backporting. > > It would be nice if we could agree on canonical names for these branches > in the git repository. Personally, I have been using "master-experimental", > which is a PITA to type, but I also have seen "experimetal" which is > probably less acurate but much more convenient. > > I suggest that we add to the dom packaging reference that branches intended > for primary release into a distribution should be named like that > distribution, that is experimental, squeeze, whatever-backport, etc, with > the exception that the branch for sid is called master. Does that sound > reasonable? > > In that case I would also suggest that dom-git-checkout also pulls the > experimental branch when it exists. > > May I suggest my own summary, about this feature (trying to solve conflicts between all opinions): - in the Debian OCaml reference: - if we package something for anything else than sid: - create a set of branch $distrib/master, $distrib/upstream forked from the master, upstream branches (E.g. squeeze/master, squeeze/upstream) - set debian/gbp.conf accordingly (upstream-branch and debian-branch) - the pristine-tar branch remains the same - in dom-git-checkout: - if */master exists, display an information note listing other branches. E.g.: You checkout master but experimental/master and squeeze/master also exists - add an option (-t $distrib) to fetch/checkout $distrib/upstream and $distrib/master, with a special keyword "all" that ifetch all $distrib - in dom-new-git-repo: - add an option (-t $distrib) to fork upstream and master branches remotely and check them out locally. What do you think about this? Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrniekr85.r67.gil...@gallu.homelinux.org
Re: names of distribution-branches in the git repository
Hello, On 22-11-2010, Stéphane Glondu wrote: > Le 22/11/2010 09:54, Sylvain Le Gall a écrit : >> Is experimental/upstream mandatory? I thought it was possible to inject >> upstream tarball in upstream branch and merge into master (or >> experimental/master). > > Then the upstream of the master branch doesn't correspond to any head? > > It is convenient to have a name for the base of the patch series, and > dom-{apply,save}-patches use that. Committing stuff in the upstream > branch not meant for master might confuse these tools (and those who are > used to them :). The upstream branch also plays a role in gbp, so it > might get confused too. For example, when you don't use pristine-tar, > which might happen temporarily when you make snapshots, gbp uses that > branch to generate a tarball. Moreover, branches are cheap in git, and > cost next to nothing when they are an ancestor of another branch. > > In short, experimental/upstream might not be mandatory, but having it is > convenient and avoids confusion IMHO. Concerning the role in gbp, it uses branch + tag: (I disable the pristine-tar) gbp:info: ocaml-extunix_0.0.1.orig.tar.gz does not exist, creating from 'upstream/0.0.1' AFAIC, I never touch anything in upstream branch, so I won't be confused ;-) My point is that upstream branch is upstream and master (or experimental/master) should refer to upstream + tag. It seems to be the case and I think it is a consistent choice. So even if branches are cheap, I would prefer to have a single upstream branch. After all we only follow one upstream. Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrniekhjf.r67.gil...@gallu.homelinux.org
Re: names of distribution-branches in the git repository
On 21-11-2010, Stéphane Glondu wrote: > Le 20/11/2010 12:05, Ralf Treinen a écrit : >> I suggest that we add to the dom packaging reference that branches intended >> for primary release into a distribution should be named like that >> distribution, that is experimental, squeeze, whatever-backport, etc, with >> the exception that the branch for sid is called master. Does that sound >> reasonable? > > My own practice is to use git-buildpackage's defaults (master, upstream) > for unstable, and prefix them by "experimental/" (e.g. > experimental/master and experimental/upstream) for experimental. For > $codename, I would similarly create $codename/master and > $codename/upstream. I'd like to see this adopted by the team. > > Having two git branches (master/upstream) per Debian branch is IMHO > cleaner, and also fits better with git-buildpackage. I got used to it > and saw nothing better so far. I find the name "experimental" ambiguous, > and the words look in the wrong order in master-experimental. And > upstream/$whatever conclicts with git-buildpackage's default name for > the upstream branch. Starting names with $branch/ doesn't conflict with > gbp's defaults, and forces to use an additionnal component name that > makes the name meaningful gbp-wise. > > I don't branch pristine-tar (and BTW I don't even always commit there > tarballs I don't upload to the official archive, especially snapshots), > given the fact that files once there are there forever, and new files > don't disturb tools (gbp, pristine-tar itself). > > Is experimental/upstream mandatory? I thought it was possible to inject upstream tarball in upstream branch and merge into master (or experimental/master). Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrniekbu1.r67.gil...@gallu.homelinux.org
Re: names of distribution-branches in the git repository
On 20-11-2010, Ralf Treinen wrote: > Hi, > > these days we are (or ar least we should be) mostly using the experimental > distribution for uploads. In that case it seems reasonable to put > modifications on a git branch that is specific to that distribution. > This of course also applies to other branches, like for backporting. > > It would be nice if we could agree on canonical names for these branches > in the git repository. Personally, I have been using "master-experimental", > which is a PITA to type, but I also have seen "experimetal" which is > probably less acurate but much more convenient. > > I suggest that we add to the dom packaging reference that branches intended > for primary release into a distribution should be named like that > distribution, that is experimental, squeeze, whatever-backport, etc, with > the exception that the branch for sid is called master. Does that sound > reasonable? > > In that case I would also suggest that dom-git-checkout also pulls the > experimental branch when it exists. > I agree on everything, but would vote for the "experimental" name. My reason: you won't have upstream-experimental or pristine-tar-experimental. Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrniefbel.r67.gil...@gallu.homelinux.org
RFC: oasis2debian, a tool to create and maintain a debian/ using _oasis
Hello all, I work on this tool since a week and it is now ready for a peer review. It is a quite simple tool that helps to translate an _oasis to a debian directory. It can do several things, depending on the content of _oasis: - package split: - if only exec -> 1 binary pkg - if exec + lib -> 1 bin (libXXX-ocaml-bin) + dev ( -ocaml-dev) + runtime (-ocaml) - if more than 2 docs -> add a -ocaml-doc package - package name: - if one findlib hierarchy (e.g. expect, expect.str and expect.pcre) use its base name (e.g. expect) - otherwise name of the source with ocaml- removed - take care of creating .doc-base for documentation defined in _oasis by translating Document section of _oasis - create debian/rules using debhelper 7, override target to use "ocaml setup.ml TARGET" - create debian/control: use package split and compute build dependencies using BuildDepends of _oasis (for lib/exec that will be built), use dh_ocaml for the rest - create debian/copyright: use _oasis fields to define copyright holders and license texts to copy - create *.install and *.install.in files to move the result of the package build into various binary packages The sole "conflicting" choice, I had to make, is the split of dev/runtime. I decided to place META and *.cma into the runtime package. I took this decision, considering the fact that plugins may need it... You can fetch it from our git repository: git+ssh://git.debian.org//git/pkg-ocaml-maint/packages/oasis2debian.git You will need to patch and install xstrp4 1.7: http://projects.camlcity.org/projects/xstrp4.html (both of these will be packaged in a short term future) I have generated a debian/ directory for the following project: git+ssh://git.debian.org/git/pkg-ocaml-maint/packages/ocamlify.git git+ssh://git.debian.org/git/pkg-ocaml-maint/packages/ocaml-data-notation.git git+ssh://git.debian.org/git/pkg-ocaml-maint/packages/ocaml-expect.git git+ssh://git.debian.org/git/pkg-ocaml-maint/packages/ocaml-extunix.git Feel free to browse the already generated debian/ and make me comments about them, so that I can at least correct the tool. They are lintian clean and can build in a pbuilder. If you have any _oasis in your project, it can also be a good time to see if my oasis2debian tool works correctly on your sources. Thank you in advance for your feedback. Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrniee8q4.r67.gil...@gallu.homelinux.org
ITP: ocaml-extunix -- Extended functions for OCaml Unix module
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: ocaml-extunix Version : 0.0.1 Upstream Author : ygrek, Sylvain Le Gall, Stephane Glondu * URL : http://extunix.forge.ocamlcore.org/ * License : LGPL-2.1 with OCaml linking exception Programming Lang: OCaml Description : Extended functions for OCaml Unix module Thin bindings to various low-level system APIs (often non-portable) which are not covered by Unix module. . Example functions: * uname * statvfs * fsync * fadvise * fallocate * atfile * dirfd * eventfd * signalfd * ... -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101119223331.21843.95583.report...@yocto.gallu.homelinux.org
Bug#603830: ITP: oasis -- Architecture for building OCaml libraries and applications
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: oasis Version : 0.2.0 Upstream Author : Sylvain Le Gall * URL : http://oasis.forge.ocamlcore.org/ * License : LGPL-2.1 with OCaml linking exception Programming Lang: OCaml Description : Architecture for building OCaml libraries and applications This program generates a full configure, build and install system for your application. It starts with a simple `_oasis` file at the toplevel of your project and creates everything required. It uses external tools like OCamlbuild and it can be considered as the glue between various subsystems that do the job. It should support the following tools: - OCamlbuild - OMake - OCamlMakefile - ocaml-autoconf It also features a do-it-yourself command line invocation and an internal configure/install scheme. Libraries are managed through findlib. It has been tested on GNU Linux and Windows. It also allows to have standard entry points and description. It helps to integrates your libraries and software with third parties tools like GODI. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101117174909.22509.5029.report...@yocto.gallu.homelinux.org
Bug#603829: ITP: ocaml-expect -- Expect-like framework in OCaml
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: ocaml-expect Version : 0.0.2 Upstream Author : Sylvain Le Gall * URL : http://forge.ocamlcore.org/projects/ocaml-expect/ * License : LGPL-2.1 with OCaml linking exception Programming Lang: OCaml Description : Expect-like framework in OCaml This is a simple implementation of `expect` to help building unitary testing of interactive program. It helps to receive question and send answers from an interactive process. You can match the question using a regular expression (Str or Pcre). You can also use a timeout to ensure that the process answer in time. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101117174456.22479.89892.report...@yocto.gallu.homelinux.org
Bug#603828: ITP: ocaml-data-notation -- Store data using OCaml notation
Package: wnpp Severity: wishlist Owner: Sylvain Le Gall * Package name: ocaml-data-notation Version : 0.0.3 Upstream Author : Sylvain Le Gall * URL : http://forge.ocamlcore.org/projects/odn * License : LGPL-2.1 with OCaml linking exception Programming Lang: OCaml Description : Store data using OCaml notation This library uses type-conv to dump OCaml data structure using OCaml data notation. This kind of data dumping helps to write OCaml code generator, like OASIS. -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101117174051.22435.62869.report...@yocto.gallu.homelinux.org
Bug#591891: libdbus-ocaml-dev: Possible conflict between 2 OCaml bindings: dBus and ssl
Hello, On Mon, Nov 08, 2010 at 05:04:19PM +0100, Gregory Bellier wrote: > I forgot to say the only info I can get from "ocamldebug" are those : > [...] > > After the step 3456, it froze. I hope this can help. I acknowledge the bug. I can reproduce it, but I really don't know what is happening. When you say you contacted upstream, you talk about ocaml-ssl or ocaml-dbus? Cheers Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101109110310.gk19...@yocto.gallu.homelinux.org
Bug#591891: libdbus-ocaml-dev: Possible conflict between 2 OCaml bindings: dBus and ssl
On Tue, Nov 09, 2010 at 12:08:51PM +0100, Gregory Bellier wrote: > Hi, > > 2010/11/9 Sylvain Le Gall > > > When you say you contacted upstream, you talk about ocaml-ssl or > ocaml-dbus? > > > I meant ocaml-dbus. Try upstream of ocaml-ssl. Maybe he will be more helpful regarding this issue. Cheers Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101109112347.gl19...@yocto.gallu.homelinux.org
Bug#602314: please let unison create hardlinks
Hello, On Wed, Nov 03, 2010 at 06:14:14PM +0100, martin f krafft wrote: > Package: unison > Version: 2.32.52-1 > Severity: wishlist > > Thanks to storing content hashes, unison can SHORTCUT copying remote > files by using already present local files as source. It sounds like > it would be just a small step towards letting unison hardlink those > files instead, falling back to copying if the hardlink fails. > I am not sure to understand what you ask. Hardlinks and copies are not the same thing. AFAIU, if you hardlink a file to a different filename and make a change to one of these files, it will also appear in the other file. You don't have this effect with copies. If the remote file is not an hardlink, there is no point making the local file an hardlink. Are you really sure about this feature? Could you explain to me how you expect it to work? Cheers Sylvain Le Gall signature.asc Description: Digital signature
Bug#602298: [libounit-ocaml-dev] force "unit" return type in OUnit.bracket for typesafety
On Wed, Nov 03, 2010 at 03:38:42PM +0100, Joost Yervante Damad wrote: > On Wednesday 03 November 2010 15:34:26 Sylvain Le Gall wrote: > > forwarded 602298 > > https://forge.ocamlcore.org/tracker/index.php?func=detail&aid=799&group_id > > =162&atid=732 > > > > thanks > > > > Hello, > > > > On Wed, Nov 03, 2010 at 03:13:20PM +0100, Joost Yervante Damad wrote: > > > Package: libounit-ocaml-dev > > > Version: 1.0.3-5.0.2 > > > > What a strange version number? Did you recompile it yourself? > > Yeah, now you mention it. Things I added locally: > - stacktrace support on testcase failure due to exception > - compile ounit with -g for better stacktrace support in general > Everything has been integrated into OUnit 1.1.0. http://ounit.forge.ocamlcore.org I'll wait the end of the freeze to upload it. Cheers Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101103163325.gk28...@yocto.gallu.homelinux.org
Bug#602298: [libounit-ocaml-dev] force "unit" return type in OUnit.bracket for typesafety
forwarded 602298 https://forge.ocamlcore.org/tracker/index.php?func=detail&aid=799&group_id=162&atid=732 thanks Hello, On Wed, Nov 03, 2010 at 03:13:20PM +0100, Joost Yervante Damad wrote: > Package: libounit-ocaml-dev > Version: 1.0.3-5.0.2 What a strange version number? Did you recompile it yourself? > Severity: wishlist > Tags: patch > > --- Please enter the report below this line. --- > > It would be useful to make return value of the testfunction f and teardown in > OUnit.bracket "unit" to enforce typing to avoid accidentally returning a > monadic type causing partial execution of a testcase. (e.g. when using Lwt) > > See attached patch for a proposed solution. > I thought about it while I release OUnit 1.1.0. It will probably make it into 1.2.0. Cheers Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101103143426.gi28...@yocto.gallu.homelinux.org
Bug#600408: ocaml: Building OCaml with LOCAL_CALLBACK_BYTECODE enabled
[Ce message a aussi été publié sur gmane.test.] On 19-10-2010, Sylvain Le Gall wrote: > Try to cc the bug as well as the mailing list (for the record). > > On 17-10-2010, Sylvain Le Gall wrote: >> On 16-10-2010, Guillaume Yziquel wrote: >>> Le Sunday 17 Oct 2010 à 00:10:41 (+0200), Stéphane Glondu a écrit : >>>> Le 16/10/2010 23:24, Guillaume Yziquel a écrit : >>>> > Package: ocaml >>>> > Version: 3.12.0-1~38 >>>> > Severity: normal >>>> >>> >>>> > [...] and digging into >>>> > the callbacks.c file, I discovered that OCaml in Debian is not built >>>> > with the LOCAL_CALLBACK_BYTECODE macro enabled. >>>> >>>> Why should it be? >>> >>> To me, the question is "why shouldn't it be?". >>> >> >> [...] >> >>>> > It seems to me that the current situation might be a can of worms and >>>> > segfaults, and I'm wondering whether it would not be a good idea to >>>> > build OCaml with LOCAL_CALLBACK_BYTECODE enabled. >>>> >>>> Where did you get that from? Is this LOCAL_CALLBACK_BYTECODE documented >>>> somewhere? The only usage I see is in byterun/callback.c, and I don't >>>> see why it should matter here (we are just using the standard bytecode >>>> interpreter). >>> >>> Haven't found documentation on LOCAL_CALLBACK_BYTECODE anywhere. I'm >>> stumbling on it doing painful gdb debugging. >>> >>> I do not think that the comments in callbacks.c are very enlightening as >>> to the proper usage of LOCAL_CALLBACK_BYTECODE. I'm not saying that it >>> should be changed, but I do not see why it should be kept this way. >>> >> >> The effect of this seems to be quite tricky and "maybe" not worth >> using a different set of options than the default OCaml one. Since, >> Debian packaging do nothing to disable this macro and that upstream >> doesn't enable it or even document it, I am not sure it is a good idea >> to change it in Debian. >> >> This doesn't mean that this is not a problem and that it doesn't cause >> segfault. I just think this issue should be dealt with upstream directly >> so that he can integrate in OCaml 3.12.1 a sane default for this option. >> >> Regards, >> Sylvain Le Gall >> >> >> -- >> To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org >> with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org >> Archive: http://lists.debian.org/slrnibligp.fmh.gil...@gallu.homelinux.org >> >> > > Regards, > Sylvain Le Gall > > > -- > To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org > with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org > Archive: http://lists.debian.org/slrnibqn5c.nk3.gil...@gallu.homelinux.org > > Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnibqngv.r67.gil...@gallu.homelinux.org
Re: Bug#600408: ocaml: Building OCaml with LOCAL_CALLBACK_BYTECODE enabled
Try to cc the bug as well as the mailing list (for the record). On 17-10-2010, Sylvain Le Gall wrote: > On 16-10-2010, Guillaume Yziquel wrote: >> Le Sunday 17 Oct 2010 à 00:10:41 (+0200), Stéphane Glondu a écrit : >>> Le 16/10/2010 23:24, Guillaume Yziquel a écrit : >>> > Package: ocaml >>> > Version: 3.12.0-1~38 >>> > Severity: normal >>> >> >>> > [...] and digging into >>> > the callbacks.c file, I discovered that OCaml in Debian is not built >>> > with the LOCAL_CALLBACK_BYTECODE macro enabled. >>> >>> Why should it be? >> >> To me, the question is "why shouldn't it be?". >> > > [...] > >>> > It seems to me that the current situation might be a can of worms and >>> > segfaults, and I'm wondering whether it would not be a good idea to >>> > build OCaml with LOCAL_CALLBACK_BYTECODE enabled. >>> >>> Where did you get that from? Is this LOCAL_CALLBACK_BYTECODE documented >>> somewhere? The only usage I see is in byterun/callback.c, and I don't >>> see why it should matter here (we are just using the standard bytecode >>> interpreter). >> >> Haven't found documentation on LOCAL_CALLBACK_BYTECODE anywhere. I'm >> stumbling on it doing painful gdb debugging. >> >> I do not think that the comments in callbacks.c are very enlightening as >> to the proper usage of LOCAL_CALLBACK_BYTECODE. I'm not saying that it >> should be changed, but I do not see why it should be kept this way. >> > > The effect of this seems to be quite tricky and "maybe" not worth > using a different set of options than the default OCaml one. Since, > Debian packaging do nothing to disable this macro and that upstream > doesn't enable it or even document it, I am not sure it is a good idea > to change it in Debian. > > This doesn't mean that this is not a problem and that it doesn't cause > segfault. I just think this issue should be dealt with upstream directly > so that he can integrate in OCaml 3.12.1 a sane default for this option. > > Regards, > Sylvain Le Gall > > > -- > To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org > with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org > Archive: http://lists.debian.org/slrnibligp.fmh.gil...@gallu.homelinux.org > > Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnibqn5c.nk3.gil...@gallu.homelinux.org
Bug#588095: downgrade severity
severity 588095 normal thanks Cheers Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101019074346.gf28...@yocto.gallu.homelinux.org
Re: Bug#588095: Should be grave?
severity 588095 normal thanks On 18-10-2010, Ivan Jager wrote: > severity 588095 grave > thanks > > I believe this should have been severity grave, on the basis that > it "makes the package in question unusable or mostly so". > > I'd say mostly so in this case, since everything in > /usr/share/doc/libcore-ocaml-doc/html/core-api/ is useless, > although at least some of the files for the extended-api seems to > have content. (Although presumably if you're using the extended > API you're also using the core API...) > > If I am wrong, feel free to change it back. I don't mean to step > on any toes, but this bug should at least get some attention > before the release. > While the problem is important and make the library hard to understand, I don't think it is a release blocking issue. The problem is taken into account but we will deal with it after the release. Thanks to highlight this issue, ping us again after the release of Squeeze. Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnibqijd.rgu.gil...@gallu.homelinux.org
Re: Bug#600408: ocaml: Building OCaml with LOCAL_CALLBACK_BYTECODE enabled
On 16-10-2010, Guillaume Yziquel wrote: > Le Sunday 17 Oct 2010 à 00:10:41 (+0200), Stéphane Glondu a écrit : >> Le 16/10/2010 23:24, Guillaume Yziquel a écrit : >> > Package: ocaml >> > Version: 3.12.0-1~38 >> > Severity: normal >> > >> > [...] and digging into >> > the callbacks.c file, I discovered that OCaml in Debian is not built >> > with the LOCAL_CALLBACK_BYTECODE macro enabled. >> >> Why should it be? > > To me, the question is "why shouldn't it be?". > [...] >> > It seems to me that the current situation might be a can of worms and >> > segfaults, and I'm wondering whether it would not be a good idea to >> > build OCaml with LOCAL_CALLBACK_BYTECODE enabled. >> >> Where did you get that from? Is this LOCAL_CALLBACK_BYTECODE documented >> somewhere? The only usage I see is in byterun/callback.c, and I don't >> see why it should matter here (we are just using the standard bytecode >> interpreter). > > Haven't found documentation on LOCAL_CALLBACK_BYTECODE anywhere. I'm > stumbling on it doing painful gdb debugging. > > I do not think that the comments in callbacks.c are very enlightening as > to the proper usage of LOCAL_CALLBACK_BYTECODE. I'm not saying that it > should be changed, but I do not see why it should be kept this way. > The effect of this seems to be quite tricky and "maybe" not worth using a different set of options than the default OCaml one. Since, Debian packaging do nothing to disable this macro and that upstream doesn't enable it or even document it, I am not sure it is a good idea to change it in Debian. This doesn't mean that this is not a problem and that it doesn't cause segfault. I just think this issue should be dealt with upstream directly so that he can integrate in OCaml 3.12.1 a sane default for this option. Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnibligp.fmh.gil...@gallu.homelinux.org
Re: Bug#599215: Homonymous modules that will conflict in META makes dh-ocaml choke.
Hello, On 05-10-2010, Guillaume Yziquel wrote: > Package: dh-ocaml > Version: 1.0.0~6 > Severity: normal > > > I've been working on an OCaml-Python binding. An I'm currently trying to > create > Debian packages for it. As Python may have multiple binaries for different > versions > of the interpreter, from 2.4 to 3.2, I'm trying to generate automatically > packages > libpython2.4-ocaml-dev to libpython3.2-dev. > > As it would impossible to load simultaneously two different versions of the > python > interpreter shared library, it seems natural that these different packages > conflict > in META files (not in Debian's control file). > > However, there will be, in each of these libpythonX.X-ocaml-dev packages, an > oCamlPython.cm[xo] binary/bytecode, without the oCamlPython.cmi file. This is > to > be able to load statically the interpreter. Much like the lablgtk.init > findlib package. > > I would not like to be able to rename the oCamlPython files upstream > (although I will > presumably be forced to). Keep in mind that they cannot be loaded > simultaneously because > of the META file conflict. The problem is that dh-ocaml fails with > > dh_ocaml -s > E: Error: unit OCamlPython exported in libpython3.2-ocaml-dev v0.90-2 but > already exported by libpython-ocaml-dev v0.90-1 > Without renaming the file, you can add a dummy function in it, that will modify the signature and allow computation of dependencies. When dh_ocaml compute it takes into account the module name and the hash exported/imported. If you add a dummy function: let is_v2_7 = true include Python.Interpreter (struct end) or let is_v3_2 = true include Python.Interpreter (struct end) You should modify the signature of the module and solve the dependency problem (and the error reported above). Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrniau2gg.fmh.gil...@gallu.homelinux.org
Bug#599284: cduce: inconsistent assumption wit curl
Package: cduce Version: 0.5.3-2+b2 Severity: grave Justification: renders package unusable When trying to compile something with cduce, I get this: File "_none_", line 1, characters 0-1: Error: Files /usr/lib/ocaml/cduce/cduce_lib.cmxa and /usr/lib/ocaml/curl/curl.cmxa make inconsistent assumptions over interface Curl A binNMU should be scheduled to solve this bug. Regards Sylvain Le Gall -- System Information: Debian Release: squeeze/sid APT prefers unstable APT policy: (500, 'unstable') Architecture: amd64 (x86_64) Kernel: Linux 2.6.32-bpo.5-amd64 (SMP w/3 CPU cores) Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages cduce depends on: ii libc62.11.2-6Embedded GNU C Library: Shared lib ii libcurl-ocaml-dev0.5.3-1 OCaml libcurl bindings (Developmen ii libcurl3-gnutls 7.21.1-1Multi-protocol file transfer libra ii libexpat-ocaml-dev 0.9.1+debian1-7 OCaml expat bindings ii libexpat12.0.1-7 XML parsing C library - runtime li ii libocamlnet-ocaml-dev2.2.9-8 OCaml application-level Internet l ii libpcre3 8.02-1.1Perl 5 Compatible Regular Expressi ii ocaml-nox [ocaml-nox-3.1 3.11.2-1ML implementation with a class-bas ii ocaml-ulex 1.1-2 OCaml lexer generator with Unicode cduce recommends no packages. cduce suggests no packages. -- no debconf information -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20101006115033.7407.96519.report...@localhost
Re: Bug#599093: dh_ocaml: warns on not resolving dependencies on modules contained in the application
On 05-10-2010, Stéphane Glondu wrote: > Le 05/10/2010 14:56, Stéphane Glondu a écrit : >>> Ralf, do you know how to make the difference between internal/external? >>> Can we trust that every unresolved symbols are externals? >> >> How can symbols be unresolved in OCaml executables? > > C stubs can be "unresolved", and IIRC dh_ocaml uses dependencies between > ocaml interfaces to find dependencies to dynamic stubs (dll*.so files). > This approach is good to find ABI-zed dependencies, but warnings > generated in this case won't be relevant most of the time. > > However, we also have access to dynamic stub dependencies in the > bytecode executables, so I propose we still use OCaml interfaces to find > ABI-zed dependencies, get rid of the warning for so-called "unresolved" > symbols on the OCaml side, and issue warnings for dll*.so that cannot be > found. > Fine for me. Since I added this warning, I was not sure if it was really mandatory or not. Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrniamb6s.fmh.gil...@gallu.homelinux.org
Re: Bug#599093: dh_ocaml: warns on not resolving dependencies on modules contained in the application
On 04-10-2010, Ralf Treinen wrote: > On Mon, Oct 04, 2010 at 08:32:21PM +0200, Mehdi Dogguy wrote: >> severity 599093 wishlist >> tags 599093 + moreinfo >> thanks >> >> On 10/04/2010 05:41 PM, Ralf Treinen wrote: >> > Package: dh-ocaml Version: 0.9.6 Severity: normal >> > >> > dh_ocaml spits out a list of warnings concerning modules that are in >> > fact defined in the application program itself, which is annoying. >> > Example bibtex2html (which is compiled to bytecode, if that >> > matters): >> > >> >> I don't know what's the purpose of this bug and how you want us to deal >> with it. So, I'm adding appropriate tags and reducing the severity. >> >> Note that those warnings are important to be sure that all external >> modules are resolved. > > They are not external. They are internal to the program. > When implementing this warning, we don't know how to be sure that we don't miss a dependency. Ralf, do you know how to make the difference between internal/external? Can we trust that every unresolved symbols are externals? Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrnialt67.fmh.gil...@gallu.homelinux.org
Bug#596622: Fwd: Re: [haXe] haXe source & ocaml library extlib
Hello, In gmane.linux.debian.devel.ocaml, you wrote: > > Le 14/09/2010 22:24, Jens Peter Secher a =E9crit : >> Hi Nicolas, >> >> Your ocaml library extlib seems to be distributed from both Google Code >> [1] and from your Motion-Twin CVS repository, which gives me some >> problems when packaging haXe 2.06 for Debian because the official >> packaging of extlib tracks the Google Code version, not the Motion-Twin >> one, see [2]. Do you plan to release the updated version of extlib on >> Google Code, or should the extlib Debian package track the Motion-Twin >> CVS instead? > > The current extlib status is a bit awkward : I was the main contributor=20 > and project initiator, then the project went unactive. And since it was=20 > hosted on SourceForge with many slowdowns and maintenance, I moved the=20 > repository on MT CVS. A few years after, the project was moved to Google=20 > Code by other people and might have evolved differently than my local=20 > repository. > > Merging both would surely be feasible but to be honest I don't have much=20 > time for it ATM. > > Maybe you can try to get in touch with extLib maintainers and find a way=20 > to get this done if they have interest in it. > Your best option right now, is to create a patch that adds the feature you need (i.e. "the new output_strings functionality"). And submit it to the BTS of the extlib project. http://code.google.com/p/ocaml-extlib/issues/list Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20100915201746.ga12...@centi.ks368928.kimsufi.com
Bug#596842: libgettext-ocaml-dev not installable in sid
Hello, On Tue, Sep 14, 2010 at 04:12:17PM +0200, Ralf Treinen wrote: > Package: libgettext-ocaml-dev > Version: 0.3.3-1+b1 > Severity: grave > User: trei...@debian.org > Usertags: edos-uninstallable > > libgettext-ocaml-dev is currently not installable in sid, probably needs > bin-NMU due to a new version of libfileutils-ocaml-dev : > > libgettext-ocaml-dev (= 0.3.3-1+b1) depends on missing: - > libfileutils-ocaml-dev-c2hd0 > I just send a binNMU request. Cheers Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20100914162533.gb9...@yocto.gallu.homelinux.org
Re: overhaul of the debian ocaml policy
On 05-09-2010, Ralf Treinen wrote: > Hi everybody, > > On Sun, Aug 15, 2010 at 10:39:06PM +0200, Ralf Treinen wrote: > >> I think I now have a draft of an update of our policy. This work >> is simply a continuation of the work started by Sylvain who did already >> the splitting in policy and reference that we had discussed earlier. > >> git+ssh://git.debian.org//git/pkg-ocaml-maint/packages/dh-ocaml.git >> branch: policy-update > > I cannot see a consensus on the question of tagging debian-specific > META files that was brought up by Sylvain. Hence, I did not put > anything on this topic in the policy for now. > Better than to have nothing, I think that we should use a comment at the beginning of the META file, when it has been written by a Debian packager: # This META file is delivered by the Debian distribution. version = ... description = ... This doesn't imply non-standard field and is at least an information about the fact upstream doesn't provide one. It is grep-able and you can even grep -E "^# This META File is delivered by the .* distribution\." which should at least partially satisfied F. Monnier. I think it is the most reasonnable way to do things. Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrni8787r.skq.gil...@gallu.homelinux.org
Re: debian and upstream (was: overhaul of the debian ocaml policy)
I cross-post this thread to debian-proje...@l.d.o. As Ralf Treinen said, this is a general question rather than an OCaml specific question. Summary: Ralf Treinen is reviewing the OCaml packaging policy and we discussed a point about adding features to upstream package. We would like to share this discussion with other Debian packagers to know what is the best practice. Point of view 1: We can add features to upstream software. Debian goal is to deliver non-broken software and we should even improve it. It includes fixing examples, documentation and the software itself. One of the example was adding a way to detect user's locale setting which was not accepted upstream because of portability issue on Windows. Point of view 2: We should limit patches to upstream software to packaging matter: - fix the build system to make it compile on Debian - fix security bugs The benefit of this approach for a packager are: - patches only apply to things most of DD or DM know: build system for example, so it is easier for anyone to understand the meaning of patches (while adding features can be tougher to understand for newcomers) - patches are easier to maintain on the long-term - no difference of behavior for the software between Linux distributions We don't exclude doing more specific patches, but this should only be temporary and hopefully immediatly sent upstream. POV2 is a little bit more documented than POV1, because I support POV1. Maybe Ralf or others can extend POV1. Thank you for giving us your opinions/best practices/experiences about this. Regards, Sylvain Le Gall On 24-08-2010, Ralf Treinen wrote: > On Mon, Aug 23, 2010 at 03:19:21PM +0200, Stéphane Glondu wrote: >> Le 23/08/2010 11:13, Ralf Treinen a écrit : >> >>>> Whenever you patch the source, IMHO, you limit the patch to: >> >>>> - fix the build system to make it compile on Debian (source -> binary >> >>>> package, new OCaml version) >> >>>> - fix security bugs >> >>> >> >>> Certainly not. We do add features (I am speaking here of debian packages >> >>> in general, not only of ocaml libraries), and we fix things that are >> >>> broken. >> >> Not any feature. I would tend to agree with Sylvain on this one. It may >> cause portability issues with non-Debian users, and can make the >> packaging harder to maintain and update (see advi, for example). > > The situation with advi has improved a lot. Currently, the debian patch set > is very small. > > Besides, I do not see why we shouldn't add any feature. Of course, > we should try to cooperate with upstream and have changes integrated > upstream, but there are various reasons why this is not possible. In that > case, our priority should be the quality of the software provided by > our packages. > >> If the patch is committed upstream, then I guess it's ok to put it in >> the Debian package if it is backward compatible somehow... but still, >> care should be taken and I wouldn't do that if I wasn't following >> upstream development more than for my average package. >> >> Upstream development should be done upstream (I don't say that the >> packager should not contribute...). A package is perfect when the >> packaging is trivial and there are no patches. > > This is the ideal situation, I agree. However, there are two sides who > have to cooperate to maintain that ideal situation. > > But this is really getting off topic now. This discussion would be more > appropriate for debian-project. > -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrni771bb.e23.gil...@gallu.homelinux.org
Re: overhaul of the debian ocaml policy
Hello, On 24-08-2010, David MENTRE wrote: > 2010/8/23 Ralf Treinen : >> PS: I just looked up the bug report, which reveals what distro the bug >> reporter is using. > > Ubuntu. > Nevermind, the bug report was aggressive be it from Ubuntu or Debian user. >> He rather should complain to his distro when they >> are just copying our stuff and then, on top of it, get it wrong (as I >> deduce from the fact that he could not rebuild his package). > > There are no OCaml developers in Ubuntu. The bug report would come > back to Debian. And Ubuntu people are hardly ever patching Debian > OCaml packages > (https://bentobako.org/ubuntu-ocaml-status/raw/ubuntu-patches.html). > >Maybe some Debian developers should understand that a > significant part of Debian OCaml work users are on Ubuntu. Rather than > a competitor, Ubuntu offers a wider exposition to Debian Developers > work. > We understand that and we (I) asked our DPL to talk about this subject at the Ubuntu conference. Basicaly, the idea was to coordinate release for sub-projects (like OCaml packaging) to know when to do a snapshot. I don't know if this idea has been accepted. But I think the problem will remain as long as their won't be an official Ubuntu developpers attached to OCaml packages. Some people are already doing a great share of this job though (like D. Mentre ;-) Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrni77077.e23.gil...@gallu.homelinux.org
Re: overhaul of the debian ocaml policy
On 23-08-2010, Ralf Treinen wrote: > On Mon, Aug 23, 2010 at 01:30:25AM +0200, Florent Monnier wrote: >> Le dimanche 22 août 2010 01:12:58, Sylvain Le Gall a écrit : >> [...] >> > > For example if we use: >> > > origin="packaging::Debian" >> > > origin="packaging::Mandriva" >> > > >> > > then one can get the list of the added META with this single command: >> > > $ grep -l packaging `ocamlc -where`/*/META >> > > >> > > But if we use: >> > > origin="Debian" >> > > origin="Mandriva" >> > > then the user has to know which distro he's using first, before to grep >> > > it as a keyword. >> > > and in this example you can't use "origin" as a keyword too, because you >> > > can't prevent an upstream to use an "origin" field too, for example: >> > > origin="janest" >> > > OK one can grep "origin" to display its content, but not *filter* the >> > > META files that come from packagers. >> > > >> > > I hope what I mean is clearer now, is it? >> > >> > Yes it is. >> > >> > I propose something like: >> > >> > origin = "debian-packagers" >> > >> > or >> > >> > createdby = "debian-packagers" >> >> I find the string "debian-packagers" fine! > > The problem is that these strings should be removed once upstream > incorporates the META file, which should be the ultimate goal. A string > like "origin" is easily mistaken for "original author", and the risk > is that people will leave them in in order to give credits. > > For that reason, I'll rather propose something like "meta-not-yet-upstream". > This way it is clear when it has to be removed. > meta-not-yet-upstream = "debian-package-only" ? Sending the META to upstream should be done with a "grep -v meta-not-yet-upstream" We could say that this value should be one line... Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrni74gh0.e23.gil...@gallu.homelinux.org
Re: overhaul of the debian ocaml policy
On 23-08-2010, Ralf Treinen wrote: > On Mon, Aug 23, 2010 at 08:56:33AM +0000, Sylvain Le Gall wrote: >> On 23-08-2010, Ralf Treinen wrote: >> > On Sun, Aug 22, 2010 at 03:07:02PM +0000, Sylvain Le Gall wrote: >> >> On 22-08-2010, Ralf Treinen wrote: >> > >> > [...] >> >> > What is the reason to mandate a special field or comment in the META >> >> > field when it is created by debian? We are patching sources all the time >> >> > both for end user applications and develpment packages, without >> >> > attaching >> >> > an extra warning sign. What makes META files so special that warrants >> >> > an exception to that rule? >> >> > >> >> >> >> Whenever you patch the source, IMHO, you limit the patch to: >> >> - fix the build system to make it compile on Debian (source -> binary >> >> package, new OCaml version) >> >> - fix security bugs >> > >> > Certainly not. We do add features (I am speaking here of debian packages >> > in general, not only of ocaml libraries), and we fix things that are >> > broken. >> > >> >> Could you be more precise about the features we add. I don't have >> examples in mind. I don't consider that FHS compliance is a feature for >> example but I don't see any commmon case where Debian packagers add a >> new function to a library. > > Looking through the hevea changelog: > > * Modify hevea and imagen such that pictures are by default generated in > png instead of gif. Added options -gif to hevea and imagen for generation > of gif. Modify man pages accordingly. > > For that one we even had a discussion on this mailing list whether we > should stay with png even when the gif patent expired (the answer was yes) > > * removed spurious space in the definition of \cyan in html/color.hva > (patch sent in by Salvador Abreu - thanks!). > Closes: Bug#136243. > > * Added an option "-charset" to specify the charset to be printed in the > META tag. The default is the output of "locale charmap" > (closes: Bug#165447) . > > * Don't insert a full stop in info output after reference "Up: (dir)" > (file: infoRef.mll, closes: Bug#179377). > > * Installed improved imagen script by Francois-Rene Rideau > . > > and more. We also have improvements of documentation, additions of > examples, and so on. The bug fixes and the code are communicated to > upstream, who usually will use them (but a fixed release will be > published much later), or something doesn't agree. > Are you not afraid of portability problem with other distro not shipping these changes? Doc and examples are ok because they don't imply incompatibilities. I am always afraid of mixing packager/upstream task... Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrni74g7e.e23.gil...@gallu.homelinux.org
Re: overhaul of the debian ocaml policy
On 23-08-2010, Ralf Treinen wrote: > On Mon, Aug 23, 2010 at 08:56:33AM +0000, Sylvain Le Gall wrote: > >> At the beginning, I was adding wrapper scripts to mldonkey >> (create/delete users, set password) and I had been criticized for this >> because I was adding too advanced features... > > Criticized by whom? Did you give in? > My preferred one: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=484674 * Why all of the Debian-specific utilities in debian/utils? This is just vanity. Please, ship the upstream software, not your own. At this time, I was in the process to give maintainership to Mehdi, so he probably closed this bug. This is probably not the vast majority of users. But there is a little truth in what he said. Sharing patches is very important and we should try to automate this as much as possible. The field I propose can be a way to automatically send it to upstream authors. This is just an idea, don't know if it is worth. Regards, Sylvain Le Gall -- To UNSUBSCRIBE, email to debian-ocaml-maint-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/slrni74g2o.e23.gil...@gallu.homelinux.org