Re: Canonical pushes upstart into user session - systemd developer complains
On Sun, Dec 02, 2012 at 11:58:00PM +0100, Vincent Lefevre wrote: Wrong. Nested blocks make config files easier to understand. Otherwise for the same feature (e.g. conditionals), one would need things like horrible state variables. For instance, think about redesigning a procmailrc with the ini format... I think redesigning a procmailrc with XML would carry over the horribleness quite well. (Sieve scripts and Exim filters are way better to understand for the casual reader. Granted, neither based on .ini nor XML but with nesting.) Kind regards Philipp Kern signature.asc Description: Digital signature
Re: Canonical pushes upstart into user session - systemd developer complains
On Dec 02, 2012, at 04:22 PM, Игорь Пашев wrote:
2012/12/2 Vincent Lefevre vinc...@vinc17.net:
No, that's not sufficient. You may want relations between key-value
pair. For instance, if you have a line with a key foo, then a line
with a key bar must also exist. Or a line with a key number must
have a value that is a number (more generally matching some regexp).
For such configs general programming languages are good.
E. g. perl:
$foo = wtf;
if ($foo !$bar) {
ohshit(...);
}
It's appealing, but not really a good option. For example, over the years
we've had many users confused with GNU Mailman's configuration file, which for
versions 3 was a Python file. People would put quotes in unnecessary places
(e.g. turning an int into a string), or forget one of the three closing triple
quotes for multi-line options, etc. There's really no reason to expect a
system administrator to be a Python or Perl programmer in order to configure
your system.
For Mailman 3 we switched to an ini file, and while that version is still in
beta and hasn't gotten nearly the same real-world exposure yet, I'm convinced
that it will be easier for end-users to manipulate than either Python or XML
syntax.
Cheers,
-Barry
signature.asc
Description: PGP signature
Re: Canonical pushes upstart into user session - systemd developer complains
On 2012-12-01 10:16:54 +0100, Wouter Verhelst wrote: On Fri, Nov 30, 2012 at 02:18:04AM +0100, Vincent Lefevre wrote: At least for Perl, I can't see anything related to validation. That's because validating an ini file is trivially easy: the line is a comment line, which must start with a # after optional whitespace, or it is a section header, where all data must be surrounded by [], or it is a key-value pair, where the key must be one word and be separated by the value by a = or it is invalid. No, that's not sufficient. You may want relations between key-value pair. For instance, if you have a line with a key foo, then a line with a key bar must also exist. Or a line with a key number must have a value that is a number (more generally matching some regexp). There, validation. To validate an XML file, much more is involved, including checks of nested tags and escaped characters. That's well-formedness (which also corresponds to validation with an empty schema). BTW, how do you do nested blocks in .ini files? You can't, and that's a feature. Instead, you have keys where the value is the name of another section (or possibly another ini file) containing the nested data. So, there is a good reason to use XML (or some other format with similar features) instead of ini: if one needs nested blocks or may need them in the future. -- Vincent Lefèvre vinc...@vinc17.net - Web: http://www.vinc17.net/ 100% accessible validated (X)HTML - Blog: http://www.vinc17.net/blog/ Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon) -- 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/20121202113059.gq5...@xvii.vinc17.org
Re: Canonical pushes upstart into user session - systemd developer complains
2012/12/2 Vincent Lefevre vinc...@vinc17.net:
No, that's not sufficient. You may want relations between key-value
pair. For instance, if you have a line with a key foo, then a line
with a key bar must also exist. Or a line with a key number must
have a value that is a number (more generally matching some regexp).
For such configs general programming languages are good.
E. g. perl:
$foo = wtf;
if ($foo !$bar) {
ohshit(...);
}
...
--
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/CALL-Q8zGhKjGBgp=i9BX5KVdQGrUce4Y0UAZzYuo3pYWb=u...@mail.gmail.com
Re: Canonical pushes upstart into user session - systemd developer complains
On Sun, Dec 02, 2012 at 12:31:00PM +0100, Vincent Lefevre wrote: On 2012-12-01 10:16:54 +0100, Wouter Verhelst wrote: On Fri, Nov 30, 2012 at 02:18:04AM +0100, Vincent Lefevre wrote: At least for Perl, I can't see anything related to validation. That's because validating an ini file is trivially easy: the line is a comment line, which must start with a # after optional whitespace, or it is a section header, where all data must be surrounded by [], or it is a key-value pair, where the key must be one word and be separated by the value by a = or it is invalid. No, that's not sufficient. Of course it is. You may want relations between key-value pair. For instance, if you have a line with a key foo, then a line with a key bar must also exist. Or a line with a key number must have a value that is a number (more generally matching some regexp). That's not validation of the format, that's validation of the data. Something like that is done by the actual code, not by a data format. A library may be able to help, but that's all (and yes, there are plenty of libraries that can do such thing for ini files). There, validation. To validate an XML file, much more is involved, including checks of nested tags and escaped characters. That's well-formedness (which also corresponds to validation with an empty schema). BTW, how do you do nested blocks in .ini files? You can't, and that's a feature. Instead, you have keys where the value is the name of another section (or possibly another ini file) containing the nested data. So, there is a good reason to use XML (or some other format with similar features) instead of ini: if one needs nested blocks or may need them in the future. I've never seen any config file with nested config blocks that didn't make the file more complex and less easy to understand. Like I said, that's a feature. -- Copyshops should do vouchers. So that next time some bureaucracy requires you to mail a form in triplicate, you can mail it just once, add a voucher, and save on postage. -- 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/20121202210452.go9...@grep.be
Re: Canonical pushes upstart into user session - systemd developer complains
On 2012-12-02 22:04:52 +0100, Wouter Verhelst wrote: On Sun, Dec 02, 2012 at 12:31:00PM +0100, Vincent Lefevre wrote: On 2012-12-01 10:16:54 +0100, Wouter Verhelst wrote: On Fri, Nov 30, 2012 at 02:18:04AM +0100, Vincent Lefevre wrote: At least for Perl, I can't see anything related to validation. That's because validating an ini file is trivially easy: the line is a comment line, which must start with a # after optional whitespace, or it is a section header, where all data must be surrounded by [], or it is a key-value pair, where the key must be one word and be separated by the value by a = or it is invalid. No, that's not sufficient. Of course it is. Perhaps for you. Not for me. Not for users who care about checking errors in their files. You may want relations between key-value pair. For instance, if you have a line with a key foo, then a line with a key bar must also exist. Or a line with a key number must have a value that is a number (more generally matching some regexp). That's not validation of the format, that's validation of the data. That's validation of the config file. This is what XML validation does: it checks whether the file is valid according to some schema. I've never seen any config file with nested config blocks that didn't make the file more complex and less easy to understand. Wrong. Nested blocks make config files easier to understand. Otherwise for the same feature (e.g. conditionals), one would need things like horrible state variables. For instance, think about redesigning a procmailrc with the ini format... -- Vincent Lefèvre vinc...@vinc17.net - Web: http://www.vinc17.net/ 100% accessible validated (X)HTML - Blog: http://www.vinc17.net/blog/ Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon) -- 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/20121202225800.ga14...@xvii.vinc17.org
Re: Canonical pushes upstart into user session - systemd developer complains
On Thu, Nov 29, 2012 at 08:37:19AM -0800, Kelly Clowers wrote: On Thu, Nov 29, 2012 at 7:23 AM, Vincent Lefevre vinc...@vinc17.net wrote: And interfaces in various programming languages? http://docs.python.org/2/library/configparser.html http://search.cpan.org/~shlomif/Config-IniFiles-2.78/lib/Config/IniFiles.pm https://rubygems.org/gems/inifile http://ini4j.sourceforge.net/ https://github.com/2ion/ini.lua https://code.google.com/p/inih/ https://github.com/shockie/node-iniparser http://php.net/manual/en/function.parse-ini-file.php http://www.codeproject.com/Articles/10809/A-Small-Class-to-Read-INI-File http://www.boost.org/doc/libs/1_45_0/doc/html/property_tree.html http://doc.qt.digia.com/qt/qsettings.html You've forgotten http://developer.gnome.org/glib/2.34/glib-Key-value-file-parser.html (which I use in nbd-server to great effect) -- Copyshops should do vouchers. So that next time some bureaucracy requires you to mail a form in triplicate, you can mail it just once, add a voucher, and save on postage. -- 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/20121201085931.ge23...@grep.be
Re: Canonical pushes upstart into user session - systemd developer complains
On Fri, Nov 30, 2012 at 02:18:04AM +0100, Vincent Lefevre wrote: On 2012-11-29 08:37:19 -0800, Kelly Clowers wrote: On Thu, Nov 29, 2012 at 7:23 AM, Vincent Lefevre vinc...@vinc17.net wrote: And interfaces in various programming languages? http://search.cpan.org/~shlomif/Config-IniFiles-2.78/lib/Config/IniFiles.pm At least for Perl, I can't see anything related to validation. That's because validating an ini file is trivially easy: the line is a comment line, which must start with a # after optional whitespace, or it is a section header, where all data must be surrounded by [], or it is a key-value pair, where the key must be one word and be separated by the value by a = or it is invalid. There, validation. To validate an XML file, much more is involved, including checks of nested tags and escaped characters. BTW, how do you do nested blocks in .ini files? You can't, and that's a feature. Instead, you have keys where the value is the name of another section (or possibly another ini file) containing the nested data. Anyway, this is fairly off-topic for -devel, so EOT for me. -- Copyshops should do vouchers. So that next time some bureaucracy requires you to mail a form in triplicate, you can mail it just once, add a voucher, and save on postage. -- 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/20121201091654.go23...@grep.be
Re: Canonical pushes upstart into user session - systemd developer complains
On Sun, 25 Nov 2012 13:06:23 -0200, Henrique de Moraes Holschuh h...@debian.org wrote: And this is not a systemd issue, even git uses that crap instead of something better like xml, or simpler, like the hierarchical format used by apt that resembles C++ classes. Are there ready-made perl, python or ruby modules that can parse (or even write?) that ISC-style config file format that I really like? libconfig-inifiles-perl does a good job for the ugly format. Greetings Marc -- -- !! No courtesy copies, please !! - Marc Haber |Questions are the | Mailadresse im Header Mannheim, Germany | Beginning of Wisdom | http://www.zugschlus.de/ Nordisch by Nature | Lt. Worf, TNG Rightful Heir | Fon: *49 621 72739834 -- 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/e1teqpv-0003mm...@swivel.zugschlus.de
Re: Canonical pushes upstart into user session - systemd developer complains
Vincent Lefevre vinc...@vinc17.net writes: No, you don't have the structure from the XML point of view. I've seen this in production. We use XML? Check! -- Stig Sandbeck Mathisen -- 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/87lidi52ms@dagon.fnord.no
Re: Canonical pushes upstart into user session - systemd developer complains
On Mon, Nov 26, 2012 at 05:00:14PM +0100, Vincent Lefevre wrote: On 2012-11-26 07:27:08 +0900, Norbert Preining wrote: Ever heard of grep, sed, awk, all these nice things that make your life happy. These tools are broken when dealing with multibyte characters. No they're not. For instance, with: foo = aéb a grep 'a.b' file will find nothing in the C locale. But it will in a UTF8 locale, or in an ISO-8859-1 locale, for instance. In a C locale, the é character simply does not exist, so you can't enter it. If you entered it in a UTF8 locale and then switched to a C locale to try to parse it, that can't work. It'd be like if you said foobar/foo, and then complained to the author of your XML parser that it doesn't understand what's going on. No please - I don't mind the key = value in group config format, that is readable, usefull, easy to edit. I disagree about readable, except on small files. For instance, in the default .subversion/config file, the group names are lost among the comments. The default .subversion/config file is a piece of documentation, not a configuration file. I agree that there's far too much noise in there. However, that's not a flaw of the format, it's a flaw of the subversion default config file. And this format is also easy to break without noticing the breakage. That claim is true for any structured file format, including XML. Everything but XML. *EVERYTHING*. This is your opinion. I disagree. XML is nice for things like validation and complex operations. XML is easy (easier) to edit if you use the right tools. XML is great for the things it was made for, but it's not a very useful configuration file format. XML requires far too much noise to be put in the config file for that. In addition, your implicit claim that it's difficult to validate ini-style configuration files, or to do complex operations on them, is just false. There are plenty of libraries to parse .ini files, too, for instance. -- Copyshops should do vouchers. So that next time some bureaucracy requires you to mail a form in triplicate, you can mail it just once, add a voucher, and save on postage. -- 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/20121129144635.gd...@grep.be
Re: Canonical pushes upstart into user session - systemd developer complains
On 2012-11-29 15:46:35 +0100, Wouter Verhelst wrote: On Mon, Nov 26, 2012 at 05:00:14PM +0100, Vincent Lefevre wrote: On 2012-11-26 07:27:08 +0900, Norbert Preining wrote: Ever heard of grep, sed, awk, all these nice things that make your life happy. These tools are broken when dealing with multibyte characters. No they're not. For instance, with: foo = aéb a grep 'a.b' file will find nothing in the C locale. But it will in a UTF8 locale, Unfortunately the C locale is the only really portable one. or in an ISO-8859-1 locale, for instance. In a C locale, the é character simply does not exist, so you can't enter it. The config file may have been generated under some locale (or in an application where locales are ignored or can be ignored, say GNU Emacs), but scripts may run in other locales, in particular C for more portability. If you entered it in a UTF8 locale and then switched to a C locale to try to parse it, that can't work. There's no such problem with XML, where the encoding of the documents are handled internally and locales do not matter. So, one can handle XML whatever the environment. Let's get back to XML? The default .subversion/config file is a piece of documentation, not a configuration file. I agree that there's far too much noise in there. However, that's not a flaw of the format, it's a flaw of the subversion default config file. But comments may be useful, and again, there's no such problem with XML. XML tools can hide comments and so on. So, you have config and the documentation at the same place, which is fine. And this format is also easy to break without noticing the breakage. That claim is true for any structured file format, including XML. Less likely with XML, because of validation (you have well-formedness checking in standard, without anything special to write). This is from my experience. Everything but XML. *EVERYTHING*. This is your opinion. I disagree. XML is nice for things like validation and complex operations. XML is easy (easier) to edit if you use the right tools. XML is great for the things it was made for, but it's not a very useful configuration file format. XML requires far too much noise to be put in the config file for that. This verbosity, or redundancy (I wouldn't call that noise), is useful to avoid some form of undetected breakage. That's a choice. In addition, your implicit claim that it's difficult to validate ini-style configuration files, or to do complex operations on them, is just false. There are plenty of libraries to parse .ini files, too, for instance. And interfaces in various programming languages? And command-line tools? The advantage of XML is that it is more common. -- Vincent Lefèvre vinc...@vinc17.net - Web: http://www.vinc17.net/ 100% accessible validated (X)HTML - Blog: http://www.vinc17.net/blog/ Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon) -- 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/20121129152303.gh5...@xvii.vinc17.org
Re: Canonical pushes upstart into user session - systemd developer complains
On Thu, Nov 29, 2012 at 04:23:03PM +0100, Vincent Lefevre wrote: The default .subversion/config file is a piece of documentation, not a configuration file. I agree that there's far too much noise in there. However, that's not a flaw of the format, it's a flaw of the subversion default config file. But comments may be useful, and again, there's no such problem with XML. XML tools can hide comments and so on. So, you have config and the documentation at the same place, which is fine. Imagine a .ini tool that can fold #-comments. -- WBR, wRAR signature.asc Description: Digital signature
Re: Canonical pushes upstart into user session - systemd developer complains
On 2012-11-29 21:33:37 +0600, Andrey Rahmatullin wrote: On Thu, Nov 29, 2012 at 04:23:03PM +0100, Vincent Lefevre wrote: The default .subversion/config file is a piece of documentation, not a configuration file. I agree that there's far too much noise in there. However, that's not a flaw of the format, it's a flaw of the subversion default config file. But comments may be useful, and again, there's no such problem with XML. XML tools can hide comments and so on. So, you have config and the documentation at the same place, which is fine. Imagine a .ini tool that can fold #-comments. Yes. But remember that the detractors of XML say that you need tools to handle it in a nice way. BTW, about the readability: ini format: [section1] key1=val1 key2=val2 [section2] foo=bar XML format: root section1 key1val1/key1 key2val2/key2 /section1 section2 foobar/foo /section2 /root It is more verbose, but I find it as readable (if you have characters that normally need to be escaped, you can still use CDATA sections, which is a way to keep the readability). -- Vincent Lefèvre vinc...@vinc17.net - Web: http://www.vinc17.net/ 100% accessible validated (X)HTML - Blog: http://www.vinc17.net/blog/ Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon) -- 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/20121129154407.gl5...@xvii.vinc17.org
Re: Canonical pushes upstart into user session - systemd developer complains
On Thu, Nov 29, 2012 at 7:23 AM, Vincent Lefevre vinc...@vinc17.net wrote: And interfaces in various programming languages? http://docs.python.org/2/library/configparser.html http://search.cpan.org/~shlomif/Config-IniFiles-2.78/lib/Config/IniFiles.pm https://rubygems.org/gems/inifile http://ini4j.sourceforge.net/ https://github.com/2ion/ini.lua https://code.google.com/p/inih/ https://github.com/shockie/node-iniparser http://php.net/manual/en/function.parse-ini-file.php http://www.codeproject.com/Articles/10809/A-Small-Class-to-Read-INI-File http://www.boost.org/doc/libs/1_45_0/doc/html/property_tree.html http://doc.qt.digia.com/qt/qsettings.html And command-line tools? An editor with syntax highlighting? Other than that, I don't know of any, but do you really need it with ini? For a specific purpose, you could probably whip something up pretty fast with one of the libraries... The advantage of XML is that it is more common. [citation needed] ini is pretty darn common... I would guess that they might be about the same. Cheers, Kelly Clowers -- 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/CAFoWM=_wgjpuuekj+g23x--g3njffpjwmybotk+e_-wtyrm...@mail.gmail.com
Re: Canonical pushes upstart into user session - systemd developer complains
Kelly Clowers, 2012-11-29 08:37:19 -0800 : [...] And command-line tools? An editor with syntax highlighting? Other than that, I don't know of any, but do you really need it with ini? For a specific purpose, you could probably whip something up pretty fast with one of the libraries... For reading a value, there's confget(1). A tool to *set* a value in an *.ini file is three lines of Python that can be inlined in a shell script (it's been mentioned elsewhere in this thread). Roland. -- Roland Mas Qui trop embrasse rate son train. -- 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/87wqx4pcr2@polymir.internal.placard.fr.eu.org
Re: Canonical pushes upstart into user session - systemd developer complains
On Thu, Nov 29, 2012 at 04:23:03PM +0100, Vincent Lefevre wrote: On 2012-11-29 15:46:35 +0100, Wouter Verhelst wrote: But it will in a UTF8 locale, Unfortunately the C locale is the only really portable one. Debian's glibc has C.UTF-8 always available these days. or in an ISO-8859-1 locale, for instance. In a C locale, the é character simply does not exist, so you can't enter it. The config file may have been generated under some locale (or in an application where locales are ignored or can be ignored, say GNU Emacs), but scripts may run in other locales, in particular C for more portability. It's high time to kill ancient encodings. They're a maintenance burden, and also, GUIs stopped paying even lip service to non-UTF8 quite some time ago. There's some discussion in #603914, although it has been derailed by minutiae of behaviour of LC_CTYPE=C, which are mostly irrelevant for getting rid of ISO-8859 and friends. Besides, even today, if someone has a config file in an encoding other than the one currently selected, that's an user error. Here XML trying to support that is a downside rather than upside. -- How to squander your resources: those silly Swedes have a sauce named hovmästarsås, the best thing ever to put on cheese, yet they waste it solely on mere salmon. -- 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/20121129170805.gb23...@angband.pl
Re: Canonical pushes upstart into user session - systemd developer complains
Vincent Lefevre vinc...@vinc17.net writes: It is more verbose, but I find it as readable (if you have characters that normally need to be escaped, you can still use CDATA sections, which is a way to keep the readability). So to keep everyone equally happy, we need: config ![CDATA[ [section1] key1=val1 key2=val2 key3=♬♫♩♩♫ [section2] foo=bar ]] /config Structure _and_ readability. -- Stig Sandbeck Mathisen -- 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/87wqx4dult@dagon.fnord.no
Re: Canonical pushes upstart into user session - systemd developer complains
On Thu, Nov 29, 2012 at 09:25:50PM +0100, Stig Sandbeck Mathisen wrote: Vincent Lefevre vinc...@vinc17.net writes: It is more verbose, but I find it as readable (if you have characters that normally need to be escaped, you can still use CDATA sections, which is a way to keep the readability). So to keep everyone equally happy, we need: config ![CDATA[ [section1] key1=val1 key2=val2 key3=♬♫♩♩♫ [section2] foo=bar ]] /config Structure _and_ readability. I wonder what the effect is of setting key3 to ♬♫♩♩♫ :D :D. Hilse fra Berlin! Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 -- 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/20121129203902.gd21...@physik.fu-berlin.de
Re: Canonical pushes upstart into user session - systemd developer complains
On Thu, Nov 29, 2012 at 09:25:50PM +0100, Stig Sandbeck Mathisen wrote:
Vincent Lefevre vinc...@vinc17.net writes:
It is more verbose, but I find it as readable (if you have characters
that normally need to be escaped, you can still use CDATA sections,
which is a way to keep the readability).
So to keep everyone equally happy, we need:
config
![CDATA[
[section1]
key1=val1
key2=val2
key3=♬♫♩♩♫
[section2]
foo=bar
]]
/config
Structure _and_ readability.
Should be more like:
- item: |
config
![CDATA[
[section1]
key1=val1
key2=val2
[section2]
foo={bar: baz}
]]
/config
Structure, readability *and* flexibility.
Ben.
--
Ben Hutchings
We get into the habit of living before acquiring the habit of thinking.
- Albert Camus
--
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/20121129213830.ga13...@decadent.org.uk
Re: Canonical pushes upstart into user session - systemd developer complains
❦ 29 novembre 2012 17:58 CET, Roland Mas lola...@debian.org : And command-line tools? An editor with syntax highlighting? Other than that, I don't know of any, but do you really need it with ini? For a specific purpose, you could probably whip something up pretty fast with one of the libraries... For reading a value, there's confget(1). A tool to *set* a value in an *.ini file is three lines of Python that can be inlined in a shell script (it's been mentioned elsewhere in this thread). Or something like Augeas. -- Format a program to help the reader understand it. - The Elements of Programming Style (Kernighan Plauger) pgprG7b6MxLvc.pgp Description: PGP signature
Re: Canonical pushes upstart into user session - systemd developer complains
On 2012-11-29 18:08:05 +0100, Adam Borowski wrote: On Thu, Nov 29, 2012 at 04:23:03PM +0100, Vincent Lefevre wrote: On 2012-11-29 15:46:35 +0100, Wouter Verhelst wrote: But it will in a UTF8 locale, Unfortunately the C locale is the only really portable one. Debian's glibc has C.UTF-8 always available these days. This seems to be Debian only, and not for the current stable. This is not OK for portable applications... until C.UTF-8 is made standard and widely adopted. -- Vincent Lefèvre vinc...@vinc17.net - Web: http://www.vinc17.net/ 100% accessible validated (X)HTML - Blog: http://www.vinc17.net/blog/ Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon) -- 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/20121130005612.gm5...@xvii.vinc17.org
Re: Canonical pushes upstart into user session - systemd developer complains
On 2012-11-29 21:25:50 +0100, Stig Sandbeck Mathisen wrote: So to keep everyone equally happy, we need: config ![CDATA[ [section1] key1=val1 key2=val2 key3=♬♫♩♩♫ [section2] foo=bar ]] /config Structure _and_ readability. No, you don't have the structure from the XML point of view. -- Vincent Lefèvre vinc...@vinc17.net - Web: http://www.vinc17.net/ 100% accessible validated (X)HTML - Blog: http://www.vinc17.net/blog/ Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon) -- 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/20121130005851.gn5...@xvii.vinc17.org
Re: Canonical pushes upstart into user session - systemd developer complains
On 2012-11-29 08:37:19 -0800, Kelly Clowers wrote: On Thu, Nov 29, 2012 at 7:23 AM, Vincent Lefevre vinc...@vinc17.net wrote: And interfaces in various programming languages? http://search.cpan.org/~shlomif/Config-IniFiles-2.78/lib/Config/IniFiles.pm At least for Perl, I can't see anything related to validation. BTW, how do you do nested blocks in .ini files? And command-line tools? An editor with syntax highlighting? No, I mean the equivalent of xmlstarlet, e.g. to extract / change values... The advantage of XML is that it is more common. [citation needed] ini is pretty darn common... I would guess that they might be about the same. XML is not just used for config files... -- Vincent Lefèvre vinc...@vinc17.net - Web: http://www.vinc17.net/ 100% accessible validated (X)HTML - Blog: http://www.vinc17.net/blog/ Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon) -- 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/20121130011804.go5...@xvii.vinc17.org
Re: Canonical pushes upstart into user session - systemd developer complains
On Mon, 2012-11-26 at 07:27, Norbert Preining wrote: On So, 25 Nov 2012, Henrique de Moraes Holschuh wrote: [crap] foo = bar ... issue, even git uses that crap instead of something better like xml, ??? Sorry, are you realistically proposing a convolutive pile of shit like XML for simple config files? I will send *each*and*every* bug report due to misconfiguration due to human incapability of editing XML to you. No please. In upstream TeX Live which is distributing over 2G of material for 15 different arch/os combinations, we threw out XML in 2005, because it is a pain in the ass, and nothing nothing else. Ever heard of grep, sed, awk, all these nice things that make your life happy. Trash them when you are doing XML. No please - I don't mind the key = value in group config format, that is readable, usefull, easy to edit. Everything but XML. *EVERYTHING*. +1 I found YAML as the best compromise for human readability and easy for automated tool processing. -- Kind regards -- 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/20121127112623.ga2...@arvanta.net
Re: Canonical pushes upstart into user session - systemd developer complains
On 2012-11-26 07:27:08 +0900, Norbert Preining wrote: Ever heard of grep, sed, awk, all these nice things that make your life happy. These tools are broken when dealing with multibyte characters. For instance, with: foo = aéb a grep 'a.b' file will find nothing in the C locale. No please - I don't mind the key = value in group config format, that is readable, usefull, easy to edit. I disagree about readable, except on small files. For instance, in the default .subversion/config file, the group names are lost among the comments. And this format is also easy to break without noticing the breakage. Everything but XML. *EVERYTHING*. This is your opinion. I disagree. XML is nice for things like validation and complex operations. XML is easy (easier) to edit if you use the right tools. -- Vincent Lefèvre vinc...@vinc17.net - Web: http://www.vinc17.net/ 100% accessible validated (X)HTML - Blog: http://www.vinc17.net/blog/ Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon) -- 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/20121126160014.gh1...@xvii.vinc17.org
Re: Canonical pushes upstart into user session - systemd developer complains
On Sun, 25 Nov 2012 17:16:09 +0100 Guillem Jover guil...@debian.org wrote: On Sun, 2012-11-25 at 23:30:01 +0800, Chow Loong Jin wrote: [...] and the hierarchical format that apt uses doesn't have a readily-usable parser outside of apt (at least not that I know of). W/o getting into the debate of what format is better or nicer, the configuration format from which APT's is based, the ISC config format, is available in libisccfg. Well, from what I see in [1], this library depends on libcap2, libdns81, libisc83 and libisccc80, among other things. Of these only libisc83 appears to be a somewhat sensible dependency (ISC Shared Library used by BIND) -- all the others look like a clear sign of a layering violation: I fail to understand how a configuration file format parsing library might depend on POSIX capabilities, a DNS library and a BIND command channel library. It's not like I want to blame the library or its packagers, just want to point out this library, as it currently stands, does not appear to be really useful outside of the set of tools comprising the BIND suite. 1. http://packages.debian.org/wheezy/libisccfg82 -- 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/20121126210501.34a2c64c28bfd391de39e...@domain007.com
Re: Canonical pushes upstart into user session - systemd developer complains
On Sat, Nov 24, 2012 at 08:43:07PM +0200, Andrej N. Gritsenko wrote: Hello! Carlos Alberto Lopez Perez has written on Saturday, 24 November, at 19:20: FYI, Yet another episode of the Linux init drama: https://wiki.ubuntu.com/FoundationsTeam/Specs/RaringUpstartUserSessions https://plus.google.com/115547683951727699051/posts/ZZWLtq6tYdn It is only me who thinks that it is going by Windows steps? With only difference Windows has more manpower to manage that bloated monster bugs. https://lists.linuxaudio.org/mailarchive/lau/2012/11/21/194431 There is a rather bad smell regarding all this. -- If you're not careful, the newspapers will have you hating the people who are being oppressed, and loving the people who are doing the oppressing. --- Malcolm X -- 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/20121125094827.GI9596@tal
Re: Canonical pushes upstart into user session - systemd developer complains
On Sun, Nov 25, 2012 at 10:48:27PM +1300, Chris Bannister wrote: https://lists.linuxaudio.org/mailarchive/lau/2012/11/21/194431 There is a rather bad smell regarding all this. None of the systemd advocates ever mentioned for example the real reason why it uses such an ugly configuration syntax and Windoze format files.^^^ I stopped reading there Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 -- 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/20121125100359.ga11...@physik.fu-berlin.de
Re: Canonical pushes upstart into user session - systemd developer complains
On Sun, 25 Nov 2012, John Paul Adrian Glaubitz wrote: On Sun, Nov 25, 2012 at 10:48:27PM +1300, Chris Bannister wrote: https://lists.linuxaudio.org/mailarchive/lau/2012/11/21/194431 There is a rather bad smell regarding all this. None of the systemd advocates ever mentioned for example the real reason why it uses such an ugly configuration syntax and Windoze format files.^^^ The: [crap] foo = bar format for config files is widely despised. And this is not a systemd issue, even git uses that crap instead of something better like xml, or simpler, like the hierarchical format used by apt that resembles C++ classes. -- One disk to rule them all, One disk to find them. One disk to bring them all and in the darkness grind them. In the Land of Redmond where the shadows lie. -- The Silicon Valley Tarot Henrique Holschuh -- 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/20121125150623.ga8...@khazad-dum.debian.net
Re: Canonical pushes upstart into user session - systemd developer complains
On Sun, 2012-11-25 at 13:06 -0200, Henrique de Moraes Holschuh wrote: On Sun, 25 Nov 2012, John Paul Adrian Glaubitz wrote: On Sun, Nov 25, 2012 at 10:48:27PM +1300, Chris Bannister wrote: https://lists.linuxaudio.org/mailarchive/lau/2012/11/21/194431 There is a rather bad smell regarding all this. None of the systemd advocates ever mentioned for example the real reason why it uses such an ugly configuration syntax and Windoze format files.^^^ The: [crap] foo = bar format for config files is widely despised. And this is not a systemd issue, even git uses that crap instead of something better like xml, or simpler, like the hierarchical format used by apt that resembles C++ classes. This 'crap' is far easier for humans to edit (without specialised tools). And *that* is most certainly in the spirit of Unix, even if this particular format is most widely used by some parts of Windows. Ben. -- Ben Hutchings Never attribute to conspiracy what can adequately be explained by stupidity. signature.asc Description: This is a digitally signed message part
Re: Canonical pushes upstart into user session - systemd developer complains
On 25/11/2012 23:06, Henrique de Moraes Holschuh wrote: The: [crap] foo = bar format for config files is widely despised. And this is not a systemd issue, even git uses that crap instead of something better like xml, or simpler, like the hierarchical format used by apt that resembles C++ classes. It is? Why? Are there any technical reasons for this besides Windows uses it? For simple key=value-type configuration that may need grouping, I really don't see an issue. XML tends to be rather long-winded, and the hierarchical format that apt uses doesn't have a readily-usable parser outside of apt (at least not that I know of). -- Kind regards, Loong Jin signature.asc Description: OpenPGP digital signature
Re: Canonical pushes upstart into user session - systemd developer complains
* Henrique de Moraes Holschuh h...@debian.org, 2012-11-25, 13:06: The: [crap] foo = bar format for config files is widely despised. And this is not a systemd issue, even git uses that crap instead of something better like xml, or simpler, like the hierarchical format used by apt that resembles C++ classes. I guess you're not being serious here, but it's not that obvious. This problem is known as Poe's Law: Without a winking smiley or other blatant display of humour, it is impossible to create a parody of fundamentalism that someone won't mistake for the real thing. -- Jakub Wilk -- 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/20121125154804.ga...@jwilk.net
Re: Canonical pushes upstart into user session - systemd developer complains
On Sun, 2012-11-25 at 23:30:01 +0800, Chow Loong Jin wrote: [...] and the hierarchical format that apt uses doesn't have a readily-usable parser outside of apt (at least not that I know of). W/o getting into the debate of what format is better or nicer, the configuration format from which APT's is based, the ISC config format, is available in libisccfg. Regards, Guillem -- 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/20121125161609.ga31...@gaara.hadrons.org
Re: Canonical pushes upstart into user session - systemd developer complains
On Sun, Nov 25, 2012 at 04:48:04PM +0100, Jakub Wilk wrote: * Henrique de Moraes Holschuh h...@debian.org, 2012-11-25, 13:06: The: [crap] foo = bar format for config files is widely despised. And this is not a systemd issue, even git uses that crap instead of something better like xml, or simpler, like the hierarchical format used by apt that resembles C++ classes. I guess you're not being serious here, but it's not that obvious. This problem is known as Poe's Law: Without a winking smiley or other blatant display of humour, it is impossible to create a parody of fundamentalism that someone won't mistake for the real thing. YMMD :D. Cheers, Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 -- 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/20121125173246.ga14...@physik.fu-berlin.de
Re: Canonical pushes upstart into user session - systemd developer complains
On So, 25 Nov 2012, Henrique de Moraes Holschuh wrote:
[crap]
foo = bar
...
issue, even git uses that crap instead of something better like xml,
??? Sorry, are you realistically proposing a convolutive pile of shit
like XML for simple config files?
I will send *each*and*every* bug report due to misconfiguration due to
human incapability of editing XML to you.
No please.
In upstream TeX Live which is distributing over 2G of material for 15
different arch/os combinations, we threw out XML in 2005, because
it is a pain in the ass, and nothing nothing else.
Ever heard of
grep, sed, awk,
all these nice things that make your life happy. Trash them when you
are doing XML.
No please - I don't mind the key = value in group config format, that
is readable, usefull, easy to edit.
Everything but XML. *EVERYTHING*.
Best wishes
Norbert
Norbert Preiningpreining@{jaist.ac.jp, logic.at, debian.org}
JAIST, Japan TeX Live Debian Developer
DSA: 0x09C5B094 fp: 14DF 2E6C 0307 BE6D AD76 A9C0 D2BF 4AA3 09C5 B094
BROMPTON
A bromton is that which is said to have been committed when you are
convinced you are about to blow off with a resounding trumpeting noise
in a public place and all that actually slips out is a tiny 'pfpt'.
--- Douglas Adams, The Meaning of Liff
--
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/20121125222708.ge...@gamma.logic.tuwien.ac.at
Canonical pushes upstart into user session - systemd developer complains
FYI, Yet another episode of the Linux init drama: https://wiki.ubuntu.com/FoundationsTeam/Specs/RaringUpstartUserSessions https://plus.google.com/115547683951727699051/posts/ZZWLtq6tYdn signature.asc Description: OpenPGP digital signature
Re: Canonical pushes upstart into user session - systemd developer complains
Hello! Carlos Alberto Lopez Perez has written on Saturday, 24 November, at 19:20: FYI, Yet another episode of the Linux init drama: https://wiki.ubuntu.com/FoundationsTeam/Specs/RaringUpstartUserSessions https://plus.google.com/115547683951727699051/posts/ZZWLtq6tYdn It is only me who thinks that it is going by Windows steps? With only difference Windows has more manpower to manage that bloated monster bugs. WBR, Andriy. -- 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/20121124184307.gd29...@rep.kiev.ua
Re: Canonical pushes upstart into user session - systemd developer complains
On Sat, Nov 24, 2012 at 07:20:02PM +0100, Carlos Alberto Lopez Perez wrote: https://wiki.ubuntu.com/FoundationsTeam/Specs/RaringUpstartUserSessions This is actually going to be very interesting to see if they are able to extend upstart in such a way that they can use it for session management similar to systemd-loginctl. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 -- 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/20121124191735.ga6...@physik.fu-berlin.de
Re: Canonical pushes upstart into user session - systemd developer complains
On Sat, Nov 24, 2012 at 08:17:35PM +0100, John Paul Adrian Glaubitz wrote: On Sat, Nov 24, 2012 at 07:20:02PM +0100, Carlos Alberto Lopez Perez wrote: https://wiki.ubuntu.com/FoundationsTeam/Specs/RaringUpstartUserSessions This is actually going to be very interesting to see if they are able to extend upstart in such a way that they can use it for session management similar to systemd-loginctl. While reading a bit on it, I found this passage: By making use of a Linux-specific prctl(2) call, we effectively tie Upstart to systems running with a Linux kernel. This is a major restriction, but porting to other systems is already complicated by the fact that even the BSDs do not provide a full POSIX environment (missing waitid(2) for example). So it's not just systemd which runs into the situation that at some point they have to drop support for non-Linux kernels because they need a Linux-specific feature, in this case prctl. Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 -- 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/20121124192951.ga6...@physik.fu-berlin.de
Re: Canonical pushes upstart into user session - systemd developer complains
John Paul Adrian Glaubitz glaub...@physik.fu-berlin.de writes: On Sat, Nov 24, 2012 at 08:17:35PM +0100, John Paul Adrian Glaubitz wrote: By making use of a Linux-specific prctl(2) call, we effectively tie Upstart to systems running with a Linux kernel. This is a major restriction, but porting to other systems is already complicated by the fact that even the BSDs do not provide a full POSIX environment (missing waitid(2) for example). So it's not just systemd which runs into the situation that at some point they have to drop support for non-Linux kernels because they need a Linux-specific feature, in this case prctl. This is not a revelation. We've known this for, literally, years. It's been part of every previous init script discussion we've had on debian-devel. However, many of those features can be made optional or be implemented another way, not to mention that the FreeBSD kernel is not static either and can adapt and adopt new features itself. Porting any more modern, event-driven, session-aware init system to FreeBSD is going to require actual work. I don't think anyone is under any illusions to the contrary. -- Russ Allbery (r...@debian.org) http://www.eyrie.org/~eagle/ -- 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/87624ux001@windlord.stanford.edu
Re: Canonical pushes upstart into user session - systemd developer complains
On Sat, 2012-11-24 at 20:29:51 +0100, John Paul Adrian Glaubitz wrote: While reading a bit on it, I found this passage: On Sat, Nov 24, 2012 at 08:17:35PM +0100, John Paul Adrian Glaubitz wrote: By making use of a Linux-specific prctl(2) call, we effectively tie Upstart to systems running with a Linux kernel. This is a major restriction, but porting to other systems is already complicated by the fact that even the BSDs do not provide a full POSIX environment (missing waitid(2) for example). So it's not just systemd which runs into the situation that at some point they have to drop support for non-Linux kernels because they need a Linux-specific feature, in this case prctl. waitid is supported on FreeBSD. Regarding prctl(2), it seems upstart is supposed to handle Linux kernels w/o prctl(PR_SET_CHILD_SUBREAPER) support gracefully by printing a warning and just ignoring respawn statements, so the same could be applied to non-Linux kernels. Of course there's other things that might need porting currently. Regards, Guillem -- 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/20121124212846.ga10...@gaara.hadrons.org
Re: Canonical pushes upstart into user session - systemd developer complains
On Sat, Nov 24, 2012 at 10:28:46PM +0100, Guillem Jover wrote: waitid is supported on FreeBSD. Are you sure? According to their status page [1] it's not yet fully implemented. The page is dated to last October. Adrian [1] http://www.freebsd.org/projects/c99/index.html -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 -- 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/20121124214629.ga7...@physik.fu-berlin.de
Re: Canonical pushes upstart into user session - systemd developer complains
On Sat, 2012-11-24 at 22:46:29 +0100, John Paul Adrian Glaubitz wrote: On Sat, Nov 24, 2012 at 10:28:46PM +0100, Guillem Jover wrote: waitid is supported on FreeBSD. Are you sure? According to their status page [1] it's not yet fully implemented. The page is dated to last October. https://github.com/freebsd/freebsd/commit/a31680b74e1dc8fb7acf515826812d0237aba661 https://github.com/freebsd/freebsd/commit/6da05143f33f9398d99a3423476dbe49b43a98a6 https://github.com/freebsd/freebsd/commit/31050a67fbf31c3de5e21c98ea1520c6c204496e Regards, Guillem -- 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/20121124232556.ga17...@gaara.hadrons.org
Re: Canonical pushes upstart into user session - systemd developer complains
On Sat, Nov 24, 2012 at 08:29:51PM +0100, John Paul Adrian Glaubitz wrote: On Sat, Nov 24, 2012 at 08:17:35PM +0100, John Paul Adrian Glaubitz wrote: On Sat, Nov 24, 2012 at 07:20:02PM +0100, Carlos Alberto Lopez Perez wrote: https://wiki.ubuntu.com/FoundationsTeam/Specs/RaringUpstartUserSessions This is actually going to be very interesting to see if they are able to extend upstart in such a way that they can use it for session management similar to systemd-loginctl. While reading a bit on it, I found this passage: By making use of a Linux-specific prctl(2) call, we effectively tie Upstart to systems running with a Linux kernel. This is a major restriction, but porting to other systems is already complicated by the fact that even the BSDs do not provide a full POSIX environment (missing waitid(2) for example). So it's not just systemd which runs into the situation that at some point they have to drop support for non-Linux kernels because they need a Linux-specific feature, in this case prctl. Upstart is already not portable to non-Linux kernels. This is a known issue; had it not been for sensitivity to not breaking Debian's non-Linux ports, it's likely that upstart would have been the default in Debian before systemd was a glimmer in Lennart's eye. Unlike systemd, upstart is open to being ported to non-Linux kernels. Sadly, so far no one has stepped up to do this work. -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. Ubuntu Developerhttp://www.debian.org/ slanga...@ubuntu.com vor...@debian.org signature.asc Description: Digital signature
