Scheduled Maintenance: We are aware of an issue with Google, AOL, and Yahoo services as email providers which are blocking new registrations. We are trying to fix the issue and we have several internal and external support tickets in process to resolve the issue. Please see: viewtopic.php?t=158230
Tips on how to find a sponsor for an archiver
Tips on how to find a sponsor for an archiver
The creator of a package included in Debian (zpaq) has (sadly) abandoned it.
https://packages.debian.org/stable/utils/zpaq
It is, in very large lines, an archiver (similar to 7z and rar to simplify) that stores data in versions, or snapshots.
https://en.wikipedia.org/wiki/ZPAQ
http://mattmahoney.net/dc/zpaq.html
Dr. Mahoney is the original creator of the PAQ series of archivers, and is quite well known in the world of compression
He is of a certain age and is now retired
Wanting to use this program on platforms that support ZFS (in particular with" visible" .zfs folders), in my case FreeBSD, I created, as is done in these cases, a fork of some success
https://github.com/fcorbelli/zpaqfranz
Now zfs is becoming popular in Debian too
My fork, by the way, is written to NOT have particular dependencies (glib and pthread in fact), not to require a Makefile etc
Runs fine on Debian, CentOS, Fedora, Windows, FreeBSD, OpenBSD, Solaris, OmniOS, MacOS, Android and (although I update it less often) NAS Qnap and even ESXi [not sure about BIG ENDIAN systems, I don't have any to do with test]
Something like that (to show how to make a packaged version should be quick and painless, it is just a small project)
g++ -O3 -Dunix zpaqfranz.cpp -pthread -o zpaqfranz -static
So far, so good
---
I would now like to propose the upload in the the debian packages of my fork
About a year ago I sent an e-mail to the maintainer of zpaq in Debian (it seemed like the right place to "start"), skitt@debian.org, but with no reply
---
I am not familiar with "how" Debian is packaged but, from what I understand, it is necessary that there is a "sponsor", aka an ALREADY accreditated developer of Debian, who check if the project is interesting and, if so, worthy of being added to the packages (or not).
---
Here's the problem: https://db.debian.org/
There is a LOT of developers, without "area of interest"
Some, ie, works on database, some works on X, some on Wi-fi etc
---
Here the neophyte request:
could someone please point out to me some possible sponsor that deals with compression, archiving and so on?
I would like to avoid sending a bunch of e-mails "blindly", as a kind of "spammer"
Something like "try asking Tizio or Caio, they take care of these things"
Of course the question will seem trivial, but remember that I am not "debianized", I have more experience in other areas
Thanks for any reply
https://packages.debian.org/stable/utils/zpaq
It is, in very large lines, an archiver (similar to 7z and rar to simplify) that stores data in versions, or snapshots.
https://en.wikipedia.org/wiki/ZPAQ
http://mattmahoney.net/dc/zpaq.html
Dr. Mahoney is the original creator of the PAQ series of archivers, and is quite well known in the world of compression
He is of a certain age and is now retired
Wanting to use this program on platforms that support ZFS (in particular with" visible" .zfs folders), in my case FreeBSD, I created, as is done in these cases, a fork of some success
https://github.com/fcorbelli/zpaqfranz
Now zfs is becoming popular in Debian too
My fork, by the way, is written to NOT have particular dependencies (glib and pthread in fact), not to require a Makefile etc
Runs fine on Debian, CentOS, Fedora, Windows, FreeBSD, OpenBSD, Solaris, OmniOS, MacOS, Android and (although I update it less often) NAS Qnap and even ESXi [not sure about BIG ENDIAN systems, I don't have any to do with test]
Something like that (to show how to make a packaged version should be quick and painless, it is just a small project)
g++ -O3 -Dunix zpaqfranz.cpp -pthread -o zpaqfranz -static
So far, so good
---
I would now like to propose the upload in the the debian packages of my fork
About a year ago I sent an e-mail to the maintainer of zpaq in Debian (it seemed like the right place to "start"), skitt@debian.org, but with no reply
---
I am not familiar with "how" Debian is packaged but, from what I understand, it is necessary that there is a "sponsor", aka an ALREADY accreditated developer of Debian, who check if the project is interesting and, if so, worthy of being added to the packages (or not).
---
Here's the problem: https://db.debian.org/
There is a LOT of developers, without "area of interest"
Some, ie, works on database, some works on X, some on Wi-fi etc
---
Here the neophyte request:
could someone please point out to me some possible sponsor that deals with compression, archiving and so on?
I would like to avoid sending a bunch of e-mails "blindly", as a kind of "spammer"
Something like "try asking Tizio or Caio, they take care of these things"
Of course the question will seem trivial, but remember that I am not "debianized", I have more experience in other areas
Thanks for any reply
-
- Global Moderator
- Posts: 2933
- Joined: 2014-07-20 18:12
- Location: Europe
- Has thanked: 72 times
- Been thanked: 401 times
Re: Tips on how to find a sponsor for an archiver
Hello,
Here [1] [2] you can find the requested information about sponsored maintainers. The assumption is that you have packaged your program for Debian before proposing it: this document [3] could be worth reading.
On the contrary, if you don't want package your program for Debian, you can send a Request for Package (RFP) proposal [4] about it.
HTH
---
[1] https://wiki.debian.org/SponsoredMaintainer
[2] https://mentors.debian.net/
[3] https://www.debian.org/doc/manuals/maint-guide/
[4] https://wiki.debian.org/RFP
Here [1] [2] you can find the requested information about sponsored maintainers. The assumption is that you have packaged your program for Debian before proposing it: this document [3] could be worth reading.
On the contrary, if you don't want package your program for Debian, you can send a Request for Package (RFP) proposal [4] about it.
HTH
---
[1] https://wiki.debian.org/SponsoredMaintainer
[2] https://mentors.debian.net/
[3] https://www.debian.org/doc/manuals/maint-guide/
[4] https://wiki.debian.org/RFP
Last edited by Aki on 2022-08-15 11:37, edited 1 time in total.
Re: Tips on how to find a sponsor for an archiver
In fact no, I do no get the software packaged for Debian
Is it necessary to become a SponsoredMaintainer?
The procedure is
- make the package
- upload
- look for a sponsor (how?)
- if he like he will upload
and/or
- find a sponsor (how? who?)
- send him the source
- if he likes he will make the package and upload
?
They may seem trivial questions, but I'm used to other "worlds", not better or worse, just "different"
Thank you
Is it necessary to become a SponsoredMaintainer?
The procedure is
- make the package
- upload
- look for a sponsor (how?)
- if he like he will upload
and/or
- find a sponsor (how? who?)
- send him the source
- if he likes he will make the package and upload
?
They may seem trivial questions, but I'm used to other "worlds", not better or worse, just "different"
Thank you
I never use the GUI on BSD, Linux and Solaris. Ever.
-
- Global Moderator
- Posts: 2933
- Joined: 2014-07-20 18:12
- Location: Europe
- Has thanked: 72 times
- Been thanked: 401 times
Re: Tips on how to find a sponsor for an archiver
Hello,
As reported in [1] it should be:
Step 1: Learn to create a package (package it the best you can according to Debian rules)
1.1 You have read the Documentation.
1.2 You know that you can join a team.
1.3 You have found an interesting package
Step 2: find a sponsor (use https://mentors.debian.net as a portal to propose your program and ask for sponsorship)
2.1 You have created an account on mentors.debian.net
2.2 You have uploaded your package on mentors.debian.net (read more)
2.3 You can use Salsa for your package.
On the contrary, if you don't want package your program for Debian by yourself, you can send a Request for Package (RFP) proposal [2] about it.
---
[1] https://wiki.debian.org/SponsoredMaintainer
[2] https://wiki.debian.org/RFP
As reported in [1] it should be:
Step 1: Learn to create a package (package it the best you can according to Debian rules)
1.1 You have read the Documentation.
1.2 You know that you can join a team.
1.3 You have found an interesting package
Step 2: find a sponsor (use https://mentors.debian.net as a portal to propose your program and ask for sponsorship)
2.1 You have created an account on mentors.debian.net
2.2 You have uploaded your package on mentors.debian.net (read more)
2.3 You can use Salsa for your package.
On the contrary, if you don't want package your program for Debian by yourself, you can send a Request for Package (RFP) proposal [2] about it.
---
[1] https://wiki.debian.org/SponsoredMaintainer
[2] https://wiki.debian.org/RFP
-
- Debian Developer
- Posts: 442
- Joined: 2022-07-12 14:10
- Has thanked: 1 time
- Been thanked: 87 times
Re: Tips on how to find a sponsor for an archiver
If you are upstream for this piece of software it might not make sense to learn to create packages just for the sake of this one package. You can imagine how this does not scale easily if you want to get the software included to many distros. So finding an interested existing debian developer might be easier. I was initially interested but when you mentioned ZFS and kernel modules I got bit scared as there is a huge potential of data corruption if something goes wrong.
Re: Tips on how to find a sponsor for an archiver
I agree, infact I just send an e-mail to the "most-system-admin" Debian-registered italian developer (!)lindi wrote: ↑2022-08-15 12:13 If you are upstream for this piece of software it might not make sense to learn to create packages just for the sake of this one package. You can imagine how this does not scale easily if you want to get the software included to many distros. So finding an interested existing debian developer might be easier.
Maybe he will point me in the right direction
In fact no, because my fork is... exactly written to check almost everything, to "paranoid++" levelI was initially interested but when you mentioned ZFS and kernel modules I got bit scared as there is a huge potential of data corruption if something goes wrong.
Doveryay, no proveryay; trust, but verify; fidarsi e' bene, non fidarsi e' meglio
---
In two word: zpaq works just like 7z, creating an archive of compressed data, listing, extracting etc
BUT
everytime you run, it will upgrade the archive, deduplicating the data with a rolling-SHA-1 hash
You crontab something like (zpaq IS in the repository, it is easy to try)
zpaq a /mnt/thebackup.zpaq /home/mydata/*
that's all
OK, the "old" zpaq does not understand visible .zfs folders
Suppose to have the /tank/d getting /tank/d/.zfs/snapshot
If you run zpaq (or 7z or tar or rar or whatevery) like that7z a /temporaneo/test.7z /tank/d nasty things start.
zpaqfranz simply (by default) ignore .zfs folders, plus other functions (like zfsadd, to freeze snapshots inside archive, zfslist - showing autocreated snapshot by zfsnap or whatever, and zfspurge, massive deleting of snapshots)
---
So this is NOT a "zfs utility that can destroy your filesystem", but a "7z on steroid that work just like snapshots"
If you are curious, try it, from zpaq
or
zpaqfranz, single .cpp, no make, no "mess"
wget https://github.com/fcorbelli/zpaqfranz/ ... qfranz.cpp
g++ -O3 -Dunix zpaqfranz.cpp -pthread -o zpaqfranz -static
I never use the GUI on BSD, Linux and Solaris. Ever.
Re: Tips on how to find a sponsor for an archiver
Maybe it's better if I rewrite it more clearly
zpaq is a program developed by another person (Dr. Mahoney) to whom, gradually, I (and others) have submitted various improvements and modifications, over the years, which have been "incorporated" in the various versions of the program.
The last (7.15 of 2016) is "inside" Debian (albeit not very well known)
Then this didn't happen anymore (now Dr. Mahoney... just... run!), and I had to do it myself
So the technology behind (zpaq) it is absolutely the merit of the initial author, I tried several times to ask him to make changes, but to no avail.
I basically evolved it from a compressor [something that compress and extract files] to an archiver [something you use for backups]
The key point (of forking an archiver) is backward compatibility
The files that are created by zpaqfranz can be read and especially extracted by "old" zpaq, because the new data (CRC-32 and hashes) are "hidden" (for zpaq)
So there will be "forever" the certainty of being able to get back your data ... with zpaq instead of zpaqfranz
I must say that I don't have BIG ENDIAN platforms on which to do functional tests, for those LITTLE ENDIAN (arm) there should be no problems.
I have a very nice user who also tries it on Android aarch64
zpaq is a program developed by another person (Dr. Mahoney) to whom, gradually, I (and others) have submitted various improvements and modifications, over the years, which have been "incorporated" in the various versions of the program.
The last (7.15 of 2016) is "inside" Debian (albeit not very well known)
Then this didn't happen anymore (now Dr. Mahoney... just... run!), and I had to do it myself
So the technology behind (zpaq) it is absolutely the merit of the initial author, I tried several times to ask him to make changes, but to no avail.
I basically evolved it from a compressor [something that compress and extract files] to an archiver [something you use for backups]
The key point (of forking an archiver) is backward compatibility
The files that are created by zpaqfranz can be read and especially extracted by "old" zpaq, because the new data (CRC-32 and hashes) are "hidden" (for zpaq)
So there will be "forever" the certainty of being able to get back your data ... with zpaq instead of zpaqfranz
I must say that I don't have BIG ENDIAN platforms on which to do functional tests, for those LITTLE ENDIAN (arm) there should be no problems.
I have a very nice user who also tries it on Android aarch64
I never use the GUI on BSD, Linux and Solaris. Ever.
-
- Debian Developer
- Posts: 442
- Joined: 2022-07-12 14:10
- Has thanked: 1 time
- Been thanked: 87 times
Re: Tips on how to find a sponsor for an archiver
I would start by filing a wishlist bug against the package to explain that the upstream is not longer developing the software and that there is a proposed fork.
Re: Tips on how to find a sponsor for an archiver
In reality, the fork is not simply a re-edition of the old one, it basically solves (almost) all the limitations of the old program, as well as adding many functions that were not foreseen.
However, I'm not sure if this type of program would be interesting for Debian (it's not that every single "piece" of software should be included in packages, of course)
Maybe it can be considered that the software in question is of too specific use (a deduplication-enabled "backupper"), or simply it is not "good enough"
Just to better explain how it is a sort of "Swiss Army Knife" for a disaster manager (my job) if the executable is called "dir" it works about the same as a (very simplified) Windows dir (!)
Why? Because on *nix (better GNU) there is not a simply way to get recursively file size AND free space
Something a backup administrator does 1.000 times a day
Yes, of course it is possible to "pipe" a string of commands, make a script etc.
But "dir" is much more easy to remember, at least for me
The r command is an enhanced "robocopy /mir" or rsync, but with ONE master vs MULTIPLE slaves (because usually 3x mirroring is done, one local and two different NASes, or more)
The s command take cumulative size and free space (to be sended by e-mail to check free space on servers)
Because, again, a disaster manager must know exactly "how big" are the folders to be taken, and "how much" free space there is,
The b (benchmark) calc the "franzomips" to quickly check VPS-CPU speed, mono or multithread
Because if you administer a remote "something" of a client, it is useful to understand how fast the CPU is
to get a quick evaluation, a small scale of comparison with other popular machines
It is not important to have a perfect benchmark, but in exactly 60 seconds you can see something like
(BTW even the free RAM, another things you need every time)
In this exaple this "thing" is just as fast as a Xeon E52620V4
The utf command "sanitize" non latin filenames (Chinese, Russian, but Italian too), to fix rsync-to-different-fx
Command z delete every empty dir in folder
Command c compare dirs with multithread (to check against rsync or robocopy). One master folder against N slave folders in parallel. Because, normally, you will have more than one copy, and various instruments like diff -qr operate on ONE master and ONE slave at a time
Command d will deduplicate without mercy some folder
Command f will fill all the available space, to check write speed (again for VPS quick-and-dirty evaluation)
Command rsync (!) will delete orphaned rsync file
The cp command will... copy files, but with ETA. Useful with huge .vmdks over slow LANs, and an optional check for untrusted networks
Command n will take the last X files (purge different archives made by tar.gz or whatever marked by timestamp, as usual in *nix scripting)
zfsadd-zfslist-zfspurge freeze, show and delete snapshots with a kind of "query" (for automated-created-crontabbed snapshots, by zfsadd or whatever)
(...)
there are functions for backup encrypted files (aka: truecrypt), for testing "extraibility" on different filesystems, for working with huge files (different kind of ETAs for .vmdk)...
...SFX modules (on Windows), support for (kind of) "full-C" data backup (Windows)... multithreaded whenever possible, with support for large-RAM elaboration (it happens to have servers maybe with 768GB of RAM, it's a shame not to use it)
(...)
function to search encrypted files (part of my job is forensic processing)
(...)
function to search similar files (aka: almost duplicated .EML with just different X-something headers)
(...)
recoverability on Windows for rapid disaster recovery of complex servers
(...)
self-archiving of older backups when they become larger than a TOT
(...)
writing to more than one output device
(...)
output redirection without the need for > (for busybox systems where there is no "real" shell)
(...)
Short version: a tool basically for systems engineers who deal with fileservers and virtual machines' backup, and work with zfs, rsync+rclone
I don't think it's the most efficient or the fastest, or "the better", neither "the best".
There are so many very good systems
AFAIK - I certainly cannot say that I know any software in the world, so it's just my opinion - it is among the most comfortable to use for the system administrator
However, I'm not sure if this type of program would be interesting for Debian (it's not that every single "piece" of software should be included in packages, of course)
Maybe it can be considered that the software in question is of too specific use (a deduplication-enabled "backupper"), or simply it is not "good enough"
Just to better explain how it is a sort of "Swiss Army Knife" for a disaster manager (my job) if the executable is called "dir" it works about the same as a (very simplified) Windows dir (!)
Code: Select all
root@aserver:~ # dir /temporaneo/ /os
==== Scanning dir <</temporaneo/>> /os
15/08/2022 12:53 <DIR> /temporaneo/
06/08/2021 12:46 <DIR> /temporaneo/ciucciato/
17/08/2022 00:00 <DIR> /temporaneo/dedup/
15/08/2022 13:01 <DIR> /temporaneo/ugo/
01/08/2021 15:05 3.992 /temporaneo/7.txt
12/08/2021 11:58 2.079.947.186 /temporaneo/daprog2.zpaq
2 File 2.079.951.178 byte
4 dir 79.723.231.232 bytes (74.25 GB) free
0.000 seconds (00:00:00) (all OK)
Something a backup administrator does 1.000 times a day
Yes, of course it is possible to "pipe" a string of commands, make a script etc.
But "dir" is much more easy to remember, at least for me
The r command is an enhanced "robocopy /mir" or rsync, but with ONE master vs MULTIPLE slaves (because usually 3x mirroring is done, one local and two different NASes, or more)
The s command take cumulative size and free space (to be sended by e-mail to check free space on servers)
Code: Select all
root@aserver:~ # zpaqfranz s /tmp /temporaneo /tank
zpaqfranz v55.2c-experimental archiver, compiled Jul 14 2022
Get directory size ignoring .zfs and :$DATA
17/08/2022 14:33:01 Scan dir <</tmp/>>
17/08/2022 14:33:01 Scan dir <</temporaneo/>>
17/08/2022 14:33:30 Scan dir <</tank/>>
======================================================================================
Free 0 53.227.646.976 ( 49.57 GB) <</tmp/>>
Free 1 79.723.231.232 ( 74.25 GB) <</temporaneo/>>
Free 2 222.897.623.040 ( 207.59 GB) <</tank/>>
======================================================================================
Dir 0 15.323.178.995 | Delta bytes| | 762| 0.01 <</tmp/>>
Dir 1 503.386.700.264 | 454.54 GB| | 605.935| 29.09 <</temporaneo/>>
Dir 2 243.761.039.995 | 212.75 GB| | 157.902| 5.32 <</tank/>>
======================================================================================
Total | 762.470.919.254| (710.11 GB)
Delta | 716.501.382.269| 762.313|files
34.806 seconds (000:00:34) (all OK)
The b (benchmark) calc the "franzomips" to quickly check VPS-CPU speed, mono or multithread
Because if you administer a remote "something" of a client, it is useful to understand how fast the CPU is
to get a quick evaluation, a small scale of comparison with other popular machines
It is not important to have a perfect benchmark, but in exactly 60 seconds you can see something like
(BTW even the free RAM, another things you need every time)
Code: Select all
root@aserver:~ # zpaqfranz b
zpaqfranz v55.2c-experimental archiver, compiled Jul 14 2022
RAM seems 36.923.277.312
Benchmarks: XXHASH64 XXH3 SHA-1 SHA-256 BLAKE3 CRC-32 CRC-32C WYHASH WHIRLPOOL MD5 SHA-3 NILSIMSA
Time limit 5 s (-n X)
Chunks of 390.62 KB (-minsize Y)
00000005 s XXHASH64: speed ( 1.98 GB/s)
00000005 s XXH3: speed ( 2.29 GB/s)
00000005 s SHA-1: speed ( 299.61 MB/s)
00000005 s SHA-256: speed ( 77.10 MB/s)
(...)
Results:
WHIRLPOOL: 54.23 MB/s (done 271.23 MB)
SHA-256: 77.10 MB/s (done 385.67 MB)
SHA-3: 182.99 MB/s (done 915.15 MB)
BLAKE3: 200.38 MB/s (done 1002.12 MB)
SHA-1: 299.61 MB/s (done 1.46 GB)
MD5: 303.42 MB/s (done 1.48 GB)
CRC-32C: 839.71 MB/s (done 4.10 GB)
XXHASH64: 1.98 GB/s (done 9.92 GB)
XXH3: 2.29 GB/s (done 11.45 GB)
WYHASH: 3.86 GB/s (done 19.31 GB)
CRC-32: 4.06 GB/s (done 20.31 GB)
NILSIMSA: 4.07 GB/s (done 20.32 GB)
franzomips single thread index 1.755 (quick CPU check, raw 1.755)
Atom N2800 (phy) 4 431.22 %
Xeon E3 1245 V2 (vir) 4 72.67 %
i5-6200U (phy) 2 92.23 %
Xeon E5 2620 V4 (phy) 8 94.97 %
Xeon E5 2630 V4 (phy) 10 113.08 %
Xeon D-1541 (vir) 8 86.41 %
i5-3570 (phy) 4 59.37 %
i7-4790K (phy) 4 53.75 %
i7-8700K (phy) 6 52.27 %
i9-10900 (phy) 10 47.37 %
AMD-5950X (phy) 16 36.58 %
60.005 seconds (000:01:00) (all OK)
The utf command "sanitize" non latin filenames (Chinese, Russian, but Italian too), to fix rsync-to-different-fx
Command z delete every empty dir in folder
Command c compare dirs with multithread (to check against rsync or robocopy). One master folder against N slave folders in parallel. Because, normally, you will have more than one copy, and various instruments like diff -qr operate on ONE master and ONE slave at a time
Command d will deduplicate without mercy some folder
Command f will fill all the available space, to check write speed (again for VPS quick-and-dirty evaluation)
Command rsync (!) will delete orphaned rsync file
The cp command will... copy files, but with ETA. Useful with huge .vmdks over slow LANs, and an optional check for untrusted networks
Command n will take the last X files (purge different archives made by tar.gz or whatever marked by timestamp, as usual in *nix scripting)
zfsadd-zfslist-zfspurge freeze, show and delete snapshots with a kind of "query" (for automated-created-crontabbed snapshots, by zfsadd or whatever)
(...)
there are functions for backup encrypted files (aka: truecrypt), for testing "extraibility" on different filesystems, for working with huge files (different kind of ETAs for .vmdk)...
...SFX modules (on Windows), support for (kind of) "full-C" data backup (Windows)... multithreaded whenever possible, with support for large-RAM elaboration (it happens to have servers maybe with 768GB of RAM, it's a shame not to use it)
(...)
function to search encrypted files (part of my job is forensic processing)
(...)
function to search similar files (aka: almost duplicated .EML with just different X-something headers)
(...)
recoverability on Windows for rapid disaster recovery of complex servers
(...)
self-archiving of older backups when they become larger than a TOT
(...)
writing to more than one output device
(...)
output redirection without the need for > (for busybox systems where there is no "real" shell)
(...)
Short version: a tool basically for systems engineers who deal with fileservers and virtual machines' backup, and work with zfs, rsync+rclone
I don't think it's the most efficient or the fastest, or "the better", neither "the best".
There are so many very good systems
AFAIK - I certainly cannot say that I know any software in the world, so it's just my opinion - it is among the most comfortable to use for the system administrator
I never use the GUI on BSD, Linux and Solaris. Ever.
- sunrat
- Administrator
- Posts: 6476
- Joined: 2006-08-29 09:12
- Location: Melbourne, Australia
- Has thanked: 118 times
- Been thanked: 474 times
Re: Tips on how to find a sponsor for an archiver
Not trying to dampen your enthusiasm, but the functions of the software are rather irrelevant to this forum. This is a user forum and we have only minimal participation from developers and maintainers. Previous respondents have explained how to propose having your software included in Debian.
I wish you luck.
I wish you luck.
“ computer users can be divided into 2 categories:
Those who have lost data
...and those who have not lost data YET ” Remember to BACKUP!
Those who have lost data
...and those who have not lost data YET ” Remember to BACKUP!
Re: Tips on how to find a sponsor for an archiver
Not a big deal
Nothing to gain, no money, no nothing
I am just waiting for answers, if any
Nothing to gain, no money, no nothing
I am just waiting for answers, if any
I never use the GUI on BSD, Linux and Solaris. Ever.
-
- Global Moderator
- Posts: 2933
- Joined: 2014-07-20 18:12
- Location: Europe
- Has thanked: 72 times
- Been thanked: 401 times
Re: Tips on how to find a sponsor for an archiver
@fcorbelli: you could have better chances if you (as upstream author) make the software more friendly with the policy requirements of the Debian distribution [1]. For example see "Copyright considerations" [2] in regard to code copyrighted by you, contributors and third part authors (e.g., code of other people you eventually included in the source code).
[1] https://www.debian.org/doc/debian-policy/
[2] https://www.debian.org/doc/debian-polic ... iderations
Re: Tips on how to find a sponsor for an archiver
I'm pretty "calm", from the licensing point of view, as the original software is already inside Debian, and the portions I've added are all opensource, GPL, BSD etc
For example, I inserted a "piece" of 7z, but asking the author's express permission
In the country where I live (Italy / EU), all types of software and free licenses - simply - do not exist (they are not recognized by law at all)
Of course, however, I should make a more precise list than the one that already exists
But let's say that I'm more "worried" on the technical side (ie if the software is good enough) rather than on the licensing side
However, your report is absolutely important, and I had underestimated it.
Thank you
For example, I inserted a "piece" of 7z, but asking the author's express permission
In the country where I live (Italy / EU), all types of software and free licenses - simply - do not exist (they are not recognized by law at all)
Of course, however, I should make a more precise list than the one that already exists
But let's say that I'm more "worried" on the technical side (ie if the software is good enough) rather than on the licensing side
However, your report is absolutely important, and I had underestimated it.
Thank you
I never use the GUI on BSD, Linux and Solaris. Ever.
-
- Global Moderator
- Posts: 2933
- Joined: 2014-07-20 18:12
- Location: Europe
- Has thanked: 72 times
- Been thanked: 401 times
Re: Tips on how to find a sponsor for an archiver
Hello,
My assumption is that making the software more "debian friendly" will make it a better software for other *nixes and FreeBSD in the medium/long run.
Your concerns about software reliability must be overcame (as far as possible) before packaging for Debian.
No developer/maintainer will take care of a faulty software or an untested software, because he/she will have to answer to bug reports of Debian's users, if any.
Maintainers often try to solve issue collaborating with upstream developers, but If the software is faulty or untested, it could be a nightmare for the Debian developer/maintainer: even more so if it concerns backup software, with requests for urgent answer for not accessible or corrupted backups. That’s why I presume it could better that the Debian maintainer would be the upstream developer, expecially for not so complex code base, at least at the beginning of package development.
Feel free to open a thread in the programming section of the forum to deepen those or others topics about your software.
My assumption is that making the software more "debian friendly" will make it a better software for other *nixes and FreeBSD in the medium/long run.
Your concerns about software reliability must be overcame (as far as possible) before packaging for Debian.
No developer/maintainer will take care of a faulty software or an untested software, because he/she will have to answer to bug reports of Debian's users, if any.
Maintainers often try to solve issue collaborating with upstream developers, but If the software is faulty or untested, it could be a nightmare for the Debian developer/maintainer: even more so if it concerns backup software, with requests for urgent answer for not accessible or corrupted backups. That’s why I presume it could better that the Debian maintainer would be the upstream developer, expecially for not so complex code base, at least at the beginning of package development.
Feel free to open a thread in the programming section of the forum to deepen those or others topics about your software.
Re: Tips on how to find a sponsor for an archiver
Well, I do all the day
I just bought a G4 Mini Mac trying to get it working (not easy task indeed) to certify non-Intel CPUs (on which it already works) but BIG ENDIAN.
On "normal" systems (ie amd64) I am quite confident that it works with a bug level comparable to that of similar software.
I agree and, in fact, on FreeBSD, I am (I hope I will become) the developer and the maintainerMaintainers often try to solve issue collaborating with upstream developers, but If the software is faulty or untested, it could be a nightmare for the Debian developer/maintainer... it could better that the Debian maintainer would be the upstream developer...
Typically bug reports are opened... on github, as issues
---
However, from what I understand, on Debian an already "Debian approved" programmer is mandatory to decide if the package qualifies to be included in the repository, if it is interesting (does something useful) and if, in a word, is good enough.
---
In my opinion (stressing MY opinion) my program is NOT the best program ever seen in Debian packages, but it's not the worst either
For archivers, on a scale of say 1 to 10, I'd say maybe 7 for general use, and maybe 8.5 for those it's specifically meant for (daily backups of fileservers and virtual machines)
I never use the GUI on BSD, Linux and Solaris. Ever.
Re: Tips on how to find a sponsor for an archiver
Perhaps, however, it does not reach the minimum required level and, therefore, I can save myself from the trouble of studying the Debian "packaging" system.
This is the "old" software (2016). Stephen Kitt seems to do "something" in 2021
https://tracker.debian.org/pkg/zpaq
This is the "old" software (2016). Stephen Kitt seems to do "something" in 2021
https://tracker.debian.org/pkg/zpaq
Thanks anyway for all the comments (to date no reply to my e-mail)Standards version of the package is outdated. wishlist
news
[2021-01-16] zpaq 7.15+repack-1 MIGRATED to testing (Debian testing watch)
[2021-01-05] Accepted zpaq 7.15+repack-1 (source) into unstable (Stephen Kitt)
[2018-04-14] zpaq 7.15-1 MIGRATED to testing (Debian testing watch)
[2018-04-09] Accepted zpaq 7.15-1 (source) into unstable (Stephen Kitt)
[2016-10-07] zpaq 1.10-3 MIGRATED to testing (Debian testing watch)
[2016-10-01] Accepted zpaq 1.10-3 (source amd64) into unstable (Chris Lamb)
[2016-09-14] zpaq 1.10-2 MIGRATED to testing (Debian testing watch)
[2016-09-08] Accepted zpaq 1.10-2 (source amd64) into unstable (Allan Dixon Jr.) (signed by: Joao Eriberto Mota Filho)
[2010-06-23] zpaq 1.10-1 MIGRATED to testing (Debian testing watch)
[2010-06-12] Accepted zpaq 1.10-1 (source amd64) (Jari Aalto) (signed by: Ricardo Mones Lastra)
I never use the GUI on BSD, Linux and Solaris. Ever.
-
- Global Moderator
- Posts: 2933
- Joined: 2014-07-20 18:12
- Location: Europe
- Has thanked: 72 times
- Been thanked: 401 times
Re: Tips on how to find a sponsor for an archiver
You are welcome.
Not to bother you, but the proper way to propose a software for packaging in Debian is to send a RFP (https://wiki.debian.org/RFP). My assumption for your software is that can be considerered as new one (not an upgrade of the one already packaged for Debian).
Re: Tips on how to find a sponsor for an archiver
Please keep us informed how it is going. This is a very interesting process.
Re: Tips on how to find a sponsor for an archiver
Update
Working on this
https://www.debian.org/doc/manuals/debm ... 04.en.html
I made my very first "test" package (well, the 2nd. The first was without man), this one
http://www.francocorbelli.it/zpaqfranz/debian/due.7z
I have submitted a ITP request via the "right path"...of course no answer, no upload, no email, no nothing
In my opinion it could be an antispam problem, sent from my small domain francocorbelli.commaybe the e-mail has been filtered
OK, nothing is easy, the very first time
Recap
0) Maked a... Debian-friendly Makefile (or, I hope so)
1) Followed the instructions with a "fake" ITP code
2) debmake
3) edit some files (in the edited-files-debmake), with a particular rule (add a -D for non-Intel CPUs)
4) created a (very small) manpage file (debuild does not like packages without man)
5) done a "debuild"
---
Some kind user could try to install the deb I have attached and report me what happens?
Keep in mind that I am really new to debian packages.
Basically a new command should appear, called zpaqfranz and a new man zpaqfranz
There is nothing else (ie exactly two files, the executable and the manpage, it does not make a "dirty" system)
And no, he won't steal your credit card code and mail it to me
In every case, if someone does not trust and want to start from scratch should...
Code: Select all
1) un-tar the source tar.gz
There is src\zpaqfranz.cpp
doc\man\zpaqfranz.1
LICENSE
Makefile
2) debmake
3) cd debmake and overwrite with edited-files-debmake
(changelog, control, copyright, manpages, README.Debian, rules)
4) debuild
... I'm almost sure ... nothing at all
On the other hand, Debian is a very large system and it is understandable that there is a lot of work in progress
- Attachments
-
- submit.jpg (400.61 KiB) Viewed 1605 times
I never use the GUI on BSD, Linux and Solaris. Ever.
Re: Tips on how to find a sponsor for an archiver
Just one note: I have about 70 "stars" on github which are certainly very few compared to much larger projects, for example lz4 counts 7.000 , however - as we say in Italy - ... better than nothing
I never use the GUI on BSD, Linux and Solaris. Ever.