Bug#998697: ITP: bees -- a btrfs deduplication agent
Package: wnpp Severity: wishlist Owner: Alexander GQ Gerasiov X-Debbugs-Cc: debian-devel@lists.debian.org * Package name: bees Version : 0.7 Upstream Author : Zygo Blaxell b...@furryterror.org * URL : https://github.com/Zygo/bees * License : GPL-3+ Programming Lang: C++ Description : a btrfs deduplication agent Best-Effort Extent-Same, a btrfs deduplication agent. bees is a block-oriented userspace deduplication agent designed for large btrfs filesystems. It is an offline dedupe combined with an incremental data scan capability to minimize time data spends on disk from write to dedupe.
Bug#982741: ITP: rtl8821cu -- dkms source for the rtl8821cu network driver
Package: wnpp Severity: wishlist Owner: Alexander GQ Gerasiov X-Debbugs-Cc: debian-devel@lists.debian.org * Package name: rtl8821cu Version : 5.8.1.7 Upstream Author : RealTek * URL : https://github.com/gerasiov/rtl8821cu * License : GPL-2 with firmware BLOB Programming Lang: C Description : dkms source for the rtl8821cu network driver This is the driver for the RealTek RTL8811CU/RTL8821CU USB WiFi controllers, which are not supported by mainline Linux kernel this days. Since there is the generated firmware BLOB is the source files, the package will go to non-free/kernel section of the archive.
Bug#931291: ITP: pdqsort -- pdqsort is a drop-in replacement for std::sort
Package: wnpp Severity: wishlist Owner: Alexander GQ Gerasiov * Package name: pdqsort Version : git snapshot Upstream Author : Orson Peters * URL : https://github.com/orlp/pdqsort * License : zlib Programming Lang: C++ Description : pattern-defeating quicksort compile-time c++ library Pattern-defeating quicksort (pdqsort) is a novel sorting algorithm that combines the fast average case of randomized quicksort with the fast worst case of heapsort, while achieving linear time on inputs with certain patterns. pdqsort is an extension and improvement of David Mussers introsort. This package provides c++ header with drop-in replacement for std::sort.
Bug#931289: ITP: croaring -- Portable Roaring bitmaps in C (and C++)
Package: wnpp Severity: wishlist Owner: Alexander GQ Gerasiov * Package name: croaring Version : 0.2.63 Upstream Author : Daniel Lemire and others * URL : https://github.com/RoaringBitmap/CRoaring * License : Apache 2.0 Programming Lang: C, C++ Description : Portable Roaring bitmaps in C (and C++) Bitsets, also called bitmaps, are commonly used as fast data structures. Unfortunately, they can use too much memory. To compensate, we often use compressed bitmaps. Roaring bitmaps are compressed bitmaps which tend to outperform conventional compressed bitmaps such as WAH, EWAH or Concise.
Bug#918942: ITP: pcg-cpp -- PCG Random Number Generation (C++ Edition)
Package: wnpp Severity: wishlist Owner: Alexander GQ Gerasiov * Package name: pcg-cpp Version : 0.98.1 Upstream Author : Melissa O'Neill and PCG Project contributors * URL : https://github.com/imneme/pcg-cpp * License : Apache-2.0 OR MIT Programming Lang: C, C++ Description : PCG Random Number Generation (C++ Edition) implementation of the PCG family of random number generators, which are fast, statistically excellent, and offer a number of useful features. Full details can be found at the PCG-Random website. This version of the code provides many family members -- if you just want one simple generator, you may prefer the minimal C version of the library. pcg random lib is needed by ClickHouse DBMS
Bug#918941: ITP: metrohash -- library with a set of state-of-the-art hash functions for non-cryptographic use cases
Package: wnpp Severity: wishlist Owner: Alexander GQ Gerasiov * Package name: metrohash Version : 1.1.3 Upstream Author : Andrew Rogers * URL : https://github.com/jandrewrogers/MetroHash * License : Apache Programming Lang: C++ Description : library with a set of state-of-the-art hash functions for non-cryptographic use cases MetroHash is a set of state-of-the-art hash functions for non-cryptographic use cases. They are notable for being algorithmically generated in addition to their exceptional performance. The set of published hash functions may be expanded in the future, having been selected from a very large set of hash functions that have been constructed this way. Fastest general-purpose functions for bulk hashing. Fastest general-purpose functions for small, variable length keys. Robust statistical bias profile, similar to the MD5 cryptographic hash. Hashes can be constructed incrementally (new) 64-bit, 128-bit, and 128-bit CRC variants currently available. Optimized for modern x86-64 microarchitectures. Elegant, compact, readable functions. metrohash lib is needed by ClickHouse DBMS
Bug#918932: ITP: libdivide -- header-only C/C++ library for optimizing integer division
Package: wnpp Severity: wishlist Owner: Alexander GQ Gerasiov * Package name: libdivide Version : 1.0 Upstream Author : ridiculous_fish * URL : https://github.com/ridiculousfish/libdivide * License : Zlib or Boost Programming Lang: C/C++ Description : header-only C/C++ library for optimizing integer division libdivide.h is a header-only C/C++ library for optimizing integer division, it has both a C API and a C++ API. This is a summary of how to use libdivide's testing tools to develop on libdivide itself. See https://libdivide.com for more information on libdivide. Package is used by ClickHouse dbms.
Bug#824152: ITP: ofxstatement -- Tool to convert proprietary bank statement to OFX format, suitable for importing to GnuCash.
Package: wnpp Severity: wishlist Owner: Alexander GQ Gerasiov * Package name: ofxstatement Version : 0.5.0 Upstream Author : Andrey Lebedev * URL : https://github.com/kedder/ofxstatement * License : GPL Programming Lang: Python Description : Tool to convert proprietary bank statement to OFX format, suitable for importing to GnuCash. Most internet banking systems are capable of exporting account transactions to some sort of computer readable formats, but few supports standard data formats, like OFX. On the other hand, personal accounting tools, such as GnuCash support standard formats only, and will probably never support proprietary statement formats of online banking systems. To bridge the gap between them, ofxstatement tool was created.
Re: Bug#568303: can-utils Debian package
Mon, 24 Feb 2014 08:08:25 +0100 John Paul Adrian Glaubitz wrote: > Hello Alexander! > > On Mon, Feb 24, 2014 at 10:44:48AM +0400, Alexander GQ Gerasiov wrote: > > Thu, 20 Feb 2014 17:30:05 +0400 > > Alexander Gerasiov wrote: > > > > > Sorry guys. I'm totally busy with other tasks last month :-(. As I > > > remember your package is quite ready for upload, so I'll do it > > > tomorrow (without my changes, just add myself as an uploader if > > > you don't mind). Ok? > > > > As Uwe is worried about correctness I'd like to mention, that I did > > not use his package as the base for last upload, but incorporated > > some of lines he wrote into my package. > > That doesn't sound plausible to me given the fact that you imported > his debian/rules file and commented out his overrride_dh_* statements > [1]. Unfortunately I missed to fixup pair of changes when build package, that's why second revision was uploaded right after first. You are talking about revision -1 which really had some issues. Please comment on the last version available. > > If you had incorporated some of his changes, you'd just have use the > default rules file from the template. > > You should also have asked yourself why Uwe had added those overrides > and not just silently commented them out. If someone adds extra > overrides, he usually has very good reasons. You should have asked Uwe > about that. I think you did not get clean with this or that's Uwe who mislead you. Those commented out overrides were left in rules file from my previous experiments, and not needed anymore. And they have nothing with Uwe's package, I believe. I can count all changes I took from his work: Arch: linux-any (totally forget that SocketCAN is Linux specific) Several strings in description field. And that's all. =\ > > The package currently also includes the debian/README.source template > and git-related files (.gitignore, gbp.conf). Template README.source was also removed in -2 revision. As for .gitignore and gbp.conf, this package is maintained under git and git-buildpackage and I see no reason, why thees files should not be included in debian/ I think some gbp related info should goes to README.source. One day I'll write it. > > As someone who is sponsoring very often and has some experience with > reviewing packages now, can-utils wouldn't have passed my quality > requirements in its current state. I could not agree with you if we speak about revision -2. > > I can therefore fully understand that Uwe is upset and I think it > would be best if you asked the FTP Masters to have the package set to > REJECT in NEW and get into touch with Uwe to coordinate improving the > package. [spend 15 minutes, explaining situation, but as Uwe prefer to continue privately, I removed it] Conclusion: Looks like Uwe decided that I modified his package, removed him from Maintainer and broke all around. And he started offense instead of discussion. Ok, I disagree, but understand why did he react like this. I think it's possible to discuss this privately. > > > Package is in new queue right now and will be soon available in > > unstable repository. > > It's actually been set to not be reviewed before February 28th to be > able to discuss this matter first. Well, I remember time when packages were held in new for 2-3 months =) -- Regards, Alexander -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20140224125933.44d8ab7b@snail
Re: Bug#568303: can-utils Debian package
Thu, 20 Feb 2014 17:30:05 +0400 Alexander Gerasiov wrote: > Sorry guys. I'm totally busy with other tasks last month :-(. As I > remember your package is quite ready for upload, so I'll do it > tomorrow (without my changes, just add myself as an uploader if you > don't mind). Ok? As Uwe is worried about correctness I'd like to mention, that I did not use his package as the base for last upload, but incorporated some of lines he wrote into my package. Package is in new queue right now and will be soon available in unstable repository. Thanks for your patience. -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20140224104448.66d46355@snail
Systemwide desktop icons
Hi there. I need to setup some software which need to be available for all users as desktop link. For kde3 (yep, we're still using it) there was /usr/share/apps/kdesktop/Desktop dir, where I could put .desktop files. But I also want Gnome and XFCE users to have such links, and can't find any standard places (in xdg or somewhere else). The only ideas I got: add some autostart xdg entry which will add this icon using xdg-desktop-icon, but this is too annoying for those users, who don't want such links. Did I miss something? PS. I'm speaking not about general package for Debian distribution, but about deploying some professional software for customer, using Debian. They really need this icon on their desktops. -- Best regards, Alexander GQ Gerasiov Contacts: e-mail:g...@cs.msu.su Jabber: g...@jabber.ru Homepage: http://gq.net.ru ICQ: 7272757 PGP fingerprint: 04B5 9D90 DF7C C2AB CD49 BAEA CA87 E9E8 2AAC 33F1 -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120302121807.253538c3@desktopvm.lvknet
user removal in postrm
Hi there. I have the question about user removal in package's postrm script. Is this correct? I see the following scenario: Package creates user for internal use only, and there will be no files owned by this user left after package removal. In this case there will be no problem if we remove user on package purge. But some packages do not remove all files on removal. E.g. mailman or mysql may leave it's databases, or torrent client (in system-wide daemon mode) should not remove files it downloaded. If such packages remove user, files will be owned by unnamed uid, and this uid will be used by the next installed package. I think this is not correct. So what is the common way? I see nothing about this in Policy. -- Best regards, Alexander GQ Gerasiov Contacts: e-mail:g...@cs.msu.su Jabber: g...@jabber.ru Homepage: http://gq.net.ru ICQ: 7272757 PGP fingerprint: 04B5 9D90 DF7C C2AB CD49 BAEA CA87 E9E8 2AAC 33F1 -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120225130214.1de37aa6@thus
Debian Policy "10.7.4 Sharing configuration files" question
Hello there. I found, that package rtpg-www modifies /etc/hosts on installation/purge http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=608451 I thing this violate 10.7.4 "The maintainer scripts must not alter a conffile of any package, including the one the scripts belong to." But maintainer disagrees with me arguing that preinst asks user, and that's user who modify /etc/hosts (using preinst, eah). Please comment this issue. Here or in #608451 -- Best regards, Alexander GQ Gerasiov Contacts: e-mail:g...@cs.msu.su Jabber: g...@jabber.ru Homepage: http://gq.net.ru ICQ: 7272757 PGP fingerprint: 04B5 9D90 DF7C C2AB CD49 BAEA CA87 E9E8 2AAC 33F1 -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/2011011557.2f8d8...@desktopvm.lvknet
Best practice for packaging *spell
Hi there. I'd like to adopt rus-ispell - Russian dictionary for aspell/ispell/hunspell/myspell [1] I'm not familiar with all this *spell staff and like to ask: is there any Policy or best practice for such packages? Is there somewhere the description of formats and tools used to generate one dictionary from another? E.g. I'd like to generate 2 versions of dictionaries with diacritical marks and without, how to do this in more natural way? To put both dictionary into one binary package and allow user's to choose one of them in there cli/gui? [1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=542011 -- Best regards, Alexander GQ Gerasiov Contacts: e-mail:g...@cs.msu.su Jabber: g...@jabber.ru Homepage: http://gq.net.ru ICQ: 7272757 PGP fingerprint: 04B5 9D90 DF7C C2AB CD49 BAEA CA87 E9E8 2AAC 33F1 -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Preparing lecture about Debian. Help needed.
Hi there. I'd like to ask you guys for some help. Here in Moscow State University there is a course "Software maintenance in Linux Distribution." It is dedicated to general question of software packaging. As example they use rpm-based community repository Sisyphus (related to AltLinux distribution). I'm going to lecture there (2 hours) about Debian project, deb packages, repositories, release cycle etc. So that would be something like "debian developer's reference + debian new maintainers' guide in pictures" :) If you have some materials (mostly I need presentation) which could help me in preparing, please mail. -- Best regards, Alexander GQ Gerasiov Contacts: e-mail:g...@cs.msu.su Jabber: g...@jabber.ru Homepage: http://gq.net.ru ICQ: 7272757 PGP fingerprint: 04B5 9D90 DF7C C2AB CD49 BAEA CA87 E9E8 2AAC 33F1 -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Re: Bug#496938: ITP: libcrypt-generatepassword-perl -- perl module to generate secure random passwords
On Fri, 29 Aug 2008 08:09:20 -0300, David Bremner <[EMAIL PROTECTED]> wrote: > > At Thu, 28 Aug 2008 20:27:44 +0400, > Alexander Gerasiov wrote: >> >> Package: wnpp >> Severity: wishlist >> Owner: Alexander Gerasiov <[EMAIL PROTECTED]> >> >> * Package name: libcrypt-generatepassword-perl >> Version : 0.03 >> >> I don't really know is this module is usefull for anybody else, but >> it is needed by netams ITP #496636. > > Hi Alexander; > > I'm sure you would be welcome to maintain this as part of the Debian > perl module packaging team. http://pkg-perl.alioth.debian.org/ > Then sponsors come looking for you... Nice, thanks for advice. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: What to do with old configs
На Sat, 22 Mar 2008 20:09:47 +0100
Adeodato Simó <[EMAIL PROTECTED]> записано:
> * Alexander GQ Gerasiov [Sat, 22 Mar 2008 22:04:38 +0300]:
>
> > Hi there.
>
> Hello Alexander. Check <http://wiki.debian.org/DpkgConffileHandling>.
Hello again. I found problem with those scripts:
$ export CONFFILE=/etc/xneur/fr/proto
$ dpkg-query -W -f='${Conffiles}' $PKGNAME | sed -n -e "\\' $CONFFILE'{s/
obsolete$//;s/.* //p}"
18a3b3b86bf3e5b0ca94407472466764
ae9d7e984350e929fd6820bf46ea6efe
That's because I have /etc/xneur/fr/proto and /etc/xneur/fr/proto3
So I suppose it would be better to add space after $CONFFILE in sed expression:
dpkg-query -W -f='${Conffiles}' $PKGNAME | sed -n -e "\\' $CONFFILE '{s/
obsolete$//;s/.* //p}"
--
Best regards,
Alexander GQ Gerasiov
Contacts:
e-mail:[EMAIL PROTECTED] Jabber: [EMAIL PROTECTED]
Homepage: http://gq.net.ru ICQ: 7272757
PGP fingerprint: 0628 ACC7 291A D4AA 6D7D 79B8 0641 D82A E3E3 CE1D
Re: What to do with old configs
На Sat, 22 Mar 2008 20:09:47 +0100 Adeodato Simó <[EMAIL PROTECTED]> записано: > * Alexander GQ Gerasiov [Sat, 22 Mar 2008 22:04:38 +0300]: > > > Hi there. > > Hello Alexander. Check <http://wiki.debian.org/DpkgConffileHandling>. Oh, thanks. That's exactly what I was interested in. -- Best regards, Alexander GQ Gerasiov Contacts: e-mail:[EMAIL PROTECTED] Jabber: [EMAIL PROTECTED] Homepage: http://gq.net.ru ICQ: 7272757 PGP fingerprint: 0628 ACC7 291A D4AA 6D7D 79B8 0641 D82A E3E3 CE1D
What to do with old configs
Hi there. I create package A version 1. It provides config file /etc/A/X. In version 2 there are no more /etc/A/X. If version 2 see /etc/A/X it will work wrong. But usual from A=1 to A=2 upgrade will leave /etc/A/X. What can I do in such situation? Can I remove /etc/A/X in preinst of version 2? -- Best regards, Alexander GQ Gerasiov Contacts: e-mail:[EMAIL PROTECTED] Jabber: [EMAIL PROTECTED] Homepage: http://gq.net.ru ICQ: 7272757 PGP fingerprint: 0628 ACC7 291A D4AA 6D7D 79B8 0641 D82A E3E3 CE1D -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Some questions about eaccelerator-package
Hello there. I ask for advise. PHP uses its own OpenSource license which is incompatible with GPL for some reasons. eAccelerator[1] is a PHP extension which speeds up php scrips at 10 times and do some more things (content caching, session handling etc.) But it is licensed under GPL and unfortunately it could not be changed[2]. So binary distribution of eAccelerator would be GPL violation. I made wrapper package which contains make-eaceelerator-package script and some other stuff which allow end-user to make binary package in one command. Now I have some questions: 1st. Where should such script go? main or contrib? from one point of view, my script, eAccelerator and PHP are all free software, but from the other one there are some restriction which makes eaccelerator not 100% free. 2nd. Is it possible to have eaccelerator-src package with eaccelerator's sources in main? If not, could it enter non-free? PS Yes, I know about pine, which is available only as source package, but I think, that it's not easy for end-user to make binary packages by hand. And I prefer script which do all this stuff for them. PPS You can take a look at this package at [3]. For now it has eaccelerator sources inside and creates two binary packages: eaccelerator-package and eaccelerator-src, but this could be changed if we decide to put this binary packages in separate sections. Any suggestion are welcomed. [1] http://eaccelerator.net [2] http://lists.debian.org/debian-legal/2005/01/msg00130.html [3] http://vice.gq.net.ru/staff/eaccelerator -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Bug#445998: ITP: eaccelerator -- PHP accelerator, optimizer, and dynamic content cache
На Tue, 9 Oct 2007 15:53:36 -0400 "Roberto C. Sánchez" <[EMAIL PROTECTED]> записано: > On Tue, Oct 09, 2007 at 09:24:14PM +0200, Tim Dijkstra wrote: > > On Tue, 09 Oct 2007 21:29:38 +0400 > > Alexander Gerasiov <[EMAIL PROTECTED]> wrote: > > > > > Package: wnpp > > > Severity: wishlist > > > Owner: Alexander Gerasiov <[EMAIL PROTECTED]> > > > > > > * Package name: eaccelerator > > > Version : 0.9.5.2 > > > Upstream Author : eaccelerator team > > > http://eaccelerator.net/wiki/Team > > > * URL : http://eaccelerator.net > > > * License : GPL > > > Programming Lang: C > > > Description : PHP accelerator, optimizer, and dynamic > > > content cache > > > > > Some dummy packages available for now in my repository at > > > http://gq.net.ru/debian > > > > > > I'm going to upload it after fixing some packaging issues. Feel > > > free to kick me by mail, if I'm too slow. > > > > Are the license issues finally solved then? Last time I checked > > binaries were not distributable, see for example: > > > > http://www.mailarchives.org/list/debian-devel/msg/2005/08164 > > > > grts Tim > > Also, in one of the many discussions about this piece of software it > came out that original author of turck-mmcache (the software on which > eaccelerator is based) now works for Zend. Zend produces a > proprietary (and expensive) accelerator-sort of product for PHP. It > is doubtful that they would enable the distribution of something that > they would see as competing with their product. > > I seem to recall that the people who took over eaccelerator had it in > mind to do a complete rewrite of the eccalerator code to break any > link with turck-mmcache, allowing them to relicense eaccelerator. If > that rewrite is complete, then the software may be distributable. > However, I have not looked into it for quite a while. OMG! That's why it's still unavailable in Debian =\ I didn't knew about PHP Licence incompatibility with GPL before. So what do you guys think about wrapper package, which allows installing eaccelerator like $ make-eaccelerator-package ? As I understand all this legal issues it will be ok. -- Best regards, Alexander GQ Gerasiov Contacts: e-mail: [EMAIL PROTECTED] Homepage: http://gq.net.ru
