[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: 7.8 uploads to experimental (pandoc, leksah, new packages)



Hello,

On Sun, Dec 21 2014, Joachim Breitner <nomeata@debian.org> wrote:
> Hi,
>
> Am Sonntag, den 21.12.2014, 09:31 -0600 schrieb Christopher Reichert:
>> > Again, packages that are not in Darcs are the cause of serious slow-down
>> > in such mass-managing of files, so I need some help here.
>>
>> Just for my own curiosity. Do you mean packages which have not yet been
>> uploaded to the debian darcs repos? e.g. new dependencies?
>
> no, packages that happen to be in git repositories. I only know of Agda
> (which is ok, because it is a leaf package), pandoc and swish, and of
> that only pandoc is really a problem, because it is a dependency of a
> few packages.
>
>> > Also, there are quite a few new packages:
>> >
>> > DRBG, auto-update, charsetdetect-ae, cipher-aes128, dynamic-state,
>> > errorcall-eq-instance, gtksourceview3, haddock-api, haddock-library,
>> > network-uri, nonce, oo-prototypes, prettyclass, regex-tdfa-text,
>> > th-expand-syns, th-reify-many, vado, vcsgui, vcswrapper, webkitgtk3,
>> > word-trie, yi-language, yi-rope, ghcjs-dom, ghcjs-codemirror,
>> > webkitgtk3-javascriptcore, jsaddle
>> >
>>
>> Noted. I have some time over the holiday and I can knock a few of these
>> in the next several days. Thanks for the heads up.
>
> Make sure to check for a darcs repo first, I might to a few of them
> myself first.
>
>

I tried my best to get some of the new packages pushed. I tested all my
builds before bumping haskell-devscripts to >0.9 so I could check builds
and warnings.


Here are some notes I made along the way:

charsetdetect-ae:

lintian gave warnings regarding references to lgpl and the RPATH.  I
think I fixed the lgpl warnings by adding a link to
/usr/share/common-licenses/LGPL-2. I don't know yet how to fix the RPATH
issues.


dynamic-state: this package did not have a copyright date. Should I make
a request upstream? I believe it's part of the yi-editor packages.


gtksourceview3: cabal-debian throws an error on this package. I've made
an issue: https://github.com/ddssff/cabal-debian/issues/25.


hsqml: This package is still a bit tricky for me.

I have the package building using QT_SELECT=5 (as suggested). However,
the qt-kde maintainers gave me the following advice:

```
[...] Now you might hit a problem at runtime. For example, if you call
moc, by default Qt4's moc will be used. So you need to somehow tell the
system you need the Qt5 version.

Take a look at qtchooser's man page, but it basically boils down to
either passing -qt5 to each qt-related tool invocation or setting
QT_SELECT to Qt5 just to your program (for example, using a script to
set the env variable and then calling the binary).
```

Any further thoughts here? It may be possible to patch HsQML's calls to
moc. The ubuntu HsQML maintainer did not mention any runtime details.


I'll keep going if there are any packages left tomorrow evening. I'll be
happy to make any corrections if I missed something.


Cheers,

-Christopher

>> This brings up a question I have. What is the best way for me to be
>> notified when a new package I help maintain is available? I know there
>> is distro-watch but I'm not certain how everyone else is using it. I
>> usually just watch the hackage rss feed and hope I don't miss anything.
>
> that is a good question. We have nothing automated yet. If you add
> yourself as an Uploader to the packages you care about, then
> https://qa.debian.org/developer.php?login=creichert07@gmail.com
> is a decent overview.
>
> Similarly http://pkg-haskell.alioth.debian.org/cgi-bin/pet.cgi lists
> packages to upgrade.
>
> https://jenkins.debian.net/view/haskell/job/haskell-package-plan/plot/
> shows an angry blue line that we should get closer to 0.
>
> And update-suggestions in the package plan suggests „trivial updates“.
>
> Greetings,
> Joachim

--
Christopher Reichert
irc: creichert
gpg: C81D 18C8 862A 3618 1376  FFA5 6BFC A992 9955 929B

Attachment: signature.asc
Description: PGP signature


Reply to: