Re: [DNG] Avoid 'git commit -m ...' every time code is tested after editing.
On 17/02/16 13:38, Simon Wise wrote: On 17/02/16 03:38, Steve Litt wrote: Where would you suggest I find out more about the practicalities of git? I use it for my own stuff quite a bit, but don't know how to do branches and all that. look at: man gitworkflows that one is more about collaboration and a wider group of developers maybe first look at: man gittutorial and perhaps some of the other pages it references. These are much more focussed on the everyday things, with examples, than the more detailed manpages. these require package git-man, there is also git-doc for online versions of these see: https://www.kernel.org/pub/software/scm/git/docs/ or go to: http://linux.die.net/man/7/gittutorial http://linux.die.net/man/1/git http://man7.org/linux/man-pages/man7/gitworkflows.7.html Simon ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Avoid 'git commit -m ...' every time code is tested after editing.
On 17/02/16 03:38, Steve Litt wrote: Where would you suggest I find out more about the practicalities of git? I use it for my own stuff quite a bit, but don't know how to do branches and all that. look at: man gitworkflows that one is more about collaboration and a wider group of developers maybe first look at: man gittutorial and perhaps some of the other pages it references. These are much more focussed on the everyday things, with examples, than the more detailed manpages. Simon ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
On Tuesday, 16 February 2016, Steve Litt wrote: > On Tue, 16 Feb 2016 16:07:35 +0200 > Aldemir Akpinar > wrote: > > > My response is unresponsive to your question, aldemir, but this is > something I've given lots of thought to in the past. I quote from the > referenced web page: > > = > > I just noticed that on one of my machines (running Debian Sid) > whenever I type ls any file name > with spaces has single quotes surrounding it. > ^^^ > > = > > > My question was obviously a rhetorical one. And I aggree that ideally filenames shouldn't have whitespaces in them however real world is far from ideal. The point I was trying to make was, I think we are seeing a pattern here just like systemd debate. While it's effect is relatively minor in this example some developer thinks forcing such stuff on people is ok. Especially on a tool like ls. I remember similar arguments with journald, where even the systemd fanboys complained, and the response was, "but you caan always switch it off" . And now they tell us to add such and such switch to our aliases to work as we would normally. Who knows what 8.25 will break. And also I'm thinking what's the next surprise gonna be. -- Aldemir -- -- aldemir ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
>Pipes and files. Oh, aye, how could I forget this :( ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
Mitt Green writes: This lad, Pádraig, says, that it happens only when outputting to terminals. I wonder, where else you can output ls. Pipes and files. This outputs to a pipe: ls | grep X This outputs to a slightly different pipe: for a in $(ls) ... This outputs to a file: ls > /tmp/foo Arnt ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
Mitt Green writes: > Adam Borowski wrote: [...] >>script names: no. >>C/Pascal/COBOL sources: no. >>mp3/videos/ebooks/etc: hell yes. > > [...] > >>The change is breaking valid use cases. > > I definitely understand the indignation, > yet I can't imagine cases, where mp3/ebooks/ > et al. are used in scripts. This doesn't affect scripts (except as mentioned elsewhere). >>The relevant bug is >>https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=813164 > > For God's sake, look, Yngwie Malmsteen, Ronnie James Dio & Tim ''Ripper'' > Owens - Mr. Crowley.mp3 became > 'Yngwie Malmsteen, Ronnie James Dio & Tim '\'''\''Ripper'\'''\'' Owens - Mr. > Crowley.mp3' '-quoted shell strings don't support any interior quoting. But the shell concatenates adjacent strings. This means the way to get a ' into a '-quote string is to split it in three, 'leader' \' 'trailer' (multiline syntax used solely as demonstration), ie, a leading, '-quoted string, a single, quoted ' and a trailing '-quoted string. Considering that ls already supports a -Q option to print "-quoted strings (which support inner quoting via \"), using "-quoting for the new output format had seemed a better idea to me. But that's more-or-less an aesthetic preference. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
Adam Borowski wrote: >The relevant bug is https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=813164 This lad, Pádraig, says, that it happens only when outputting to terminals. I wonder, where else you can output ls. And surely, pasting this back to anywhere, the result will be the same: unwanted slashes and quotes. Mitt ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
[DNG] Which Devuan alpha4 for vbox?
Wish to run an i386 Devuan alpha4 as a guest OS in VirtualBox's current v5.0.14[1], and under an i386 Linux-distro host. The Devuan download zone[2] currently lists the four possibilities for this as - devuan-jessie-i386-alpha4.box - devuan-jessie-i386-alpha4.qcow2 - devuan-jessie-i386-alpha4-netboot.iso - devuan-jessie-i386-alpha4-netboot-auto.iso The appropriate snippet of the Devuan download zone's README textfile[3] reads as "Beware the netinst-auto flavour is preseed to ask no questions. Virtual machines are qcow2 harddisk images ready to use in Qemu" From the above snippet, I would naturally assume that the two devuan-jessie-i386-alpha4 files ending in "-netboot-auto.iso" and ".qcow2" can be excluded for installation as possible vbox guests. Thus is devuan-jessie-i386-alpha4.box *the* one to use here for vbox, and if so, are there additional special hints/instructions for getting this to work smoothly in vbox (e.g., using one the appropriate Guest Additions[4] listed at [5]) ?? Refs == [1]https://www.virtualbox.org/ [2]http://files.devuan.org/ [3]http://files.devuan.org/00_README.txt [4]https://www.virtualbox.org/wiki/Guest_OSes [5]http://download.virtualbox.org/virtualbox/5.0.14/ Thanks for assistance with this. -A -- Go Open Source Software Bears! _ Get your FREE, LinuxWaves.com Email Now! --> http://www.LinuxWaves.com Join Linux Discussions! --> http://Community.LinuxWaves.com ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
Adam Borowski wrote: >script names: no. >C/Pascal/COBOL sources: no. >mp3/videos/ebooks/etc: hell yes. [...] >The change is breaking valid use cases. I definitely understand the indignation, yet I can't imagine cases, where mp3/ebooks/ et al. are used in scripts. >The relevant bug is >https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=813164 For God's sake, look, Yngwie Malmsteen, Ronnie James Dio & Tim ''Ripper'' Owens - Mr. Crowley.mp3 became 'Yngwie Malmsteen, Ronnie James Dio & Tim '\'''\''Ripper'\'''\'' Owens - Mr. Crowley.mp3' Phew. // Mitt ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
On Tue, Feb 16, 2016 at 07:51:47PM +0300, Mitt Green wrote: > No idea, why they did it, > but anyway, who uses whitespaces in > names of scripts? I always get > rid of them even when naming > wallpapers. script names: no. C/Pascal/COBOL sources: no. mp3/videos/ebooks/etc: hell yes. The change is breaking valid use cases. The relevant bug is https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=813164 -- but I ask you: don't go there just to whine (yes, the change is a steaming piece of crock but this has already been said). If you post to that bug, please add only _new_, rational, polite arguments. And I've already submitted a patch and identified the offending commit in upstream VCS. It looks like everyone but Pádraig Brady (the upstream developer who did this) and ansgar agree. And the Debian maintainer of coreutils has set severity to "serious" to stop this bad change from migrating. -- A tit a day keeps the vet away. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
On Tue, 16 Feb 2016 16:07:35 +0200 Aldemir Akpinar wrote: > There's a recent discussion on the coreutils mailing list, where some > people complained about the default ls output with latest release. On > coreutils 8.25 ls will wrap filenames with quotes if it includes > whitespace. > > And when people protest, the answers are usual arguments; it just > happens on the console output, or just add -N to your aliases etc. > etc. > > Here's a discussion with the response from the maintainer as well: > > http://unix.stackexchange.com/questions/258679/why-is-ls-suddenly-surrounding-items-with-spaces-in-single-quotes > > When this version hits the distributions, it will break many scripts. > So why change the default behaviour? > -- > aldemir My response is unresponsive to your question, aldemir, but this is something I've given lots of thought to in the past. I quote from the referenced web page: = I just noticed that on one of my machines (running Debian Sid) whenever I type ls any file name with spaces has single quotes surrounding it. ^^^ = Here at Troubleshooters.Com we have a policy that only the following characters can appear in a filename: [a-z,A-Z,0-9,\.\_\-] You've all heard me criticize programmers who make it easier on themselves by making it harder on the user. But there are limits to that philosophy, and given that there's so much filename parsing throughout all of POSIX, and so much separating commands into executable names and arguments, I believe filenames should have been restricted for the good of everybody. PDP11/RT-11 and CPM didn't allow spaces in filenames, and they worked just fine. Think about where most of your space-laden filenames come from: They're emailed to you by no-tech, no-discipline Windows-weenies. Well, unless FreeDesktop has recently made space-laden config files, but I don't think even FreeDesktop is *that* ignorant. Of course, none of this responds to the fact that space-laden filenames *do* exist, so script writers have had to work around that, and some of those workarounds might bust. SteveT Steve Litt February 2016 featured book: The Key to Everyday Excellence http://www.troubleshooters.com/key ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
Emiliano Marini writes: For example, this willl break scripts parsig ls output. Will it? I spent a few minutes looking for scripts that might be broken. I did find one that acted on ls output, but it wasn't broken. Arnt ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
No idea, why they did it, but anyway, who uses whitespaces in names of scripts? I always get rid of them even when naming wallpapers. Mitt ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Avoid 'git commit -m ...' every time code is tested after editing.
On Tue, 16 Feb 2016 19:47:37 +1300 Daniel Reurich wrote: > On 16 February 2016 3:35:55 PM NZDT, Steve Litt > wrote: > >When a version is a release, don't you just give it a tag? > > > Yes. But that is no excuse for not using private branches to develop > on. If you have to revert a commit your probably doing it wrong. > Please keep master branches for merging tested and release ready > branches only. > > For Devuan packaging projects, devuan/ branches should be > marked protected and changes proposed via merge requests. And the > owners/masters should be the only ones to merge changes and set > release tags. > > Daniel Thanks for the clarification Daniel. Should each developer create a branch for each of his/her subprojects? Or a branch for each subproject? Or a single development project to merge into Master on the rare occasion when the development branch is functioning and tested? Thanks, SteveT Steve Litt February 2016 featured book: The Key to Everyday Excellence http://www.troubleshooters.com/key ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Avoid 'git commit -m ...' every time code is tested after editing.
On Tue, 16 Feb 2016 10:06:35 + KatolaZ wrote: > On Mon, Feb 15, 2016 at 09:35:55PM -0500, Steve Litt wrote: > > [cut] > > > > > > > Commit often. Branch whenever needed needed. Merge when it > > > works. Release when "perfect" (the last one should be really > > > considered with a pinch of salt :P). > > > > When a version is a release, don't you just give it a tag? > > > > Yes Steve. My point was that git is made to maintain the history of a > project, and to help you avoiding disasters, or containing their > impact. git is useful only if you commit fairly often, otherwise it is > better to use a backup system, rather than a revision control > system... > > So, there is no point into committing "only working stuff", because > this is the best way to mess things up and/or to lose hours of > precious work. I didn't understand the context of the question before, but now that I do, it makes perfect sense. Also, besides tagging releases, can't you also tag working versions, like "works22143", so anyone can git clone something that works? Or... > As it has been suggested by several others, whenever > you need to have the master branch to always be "working", it's far > better to do development in other branches, committing there as > frequently as needed (i.e., even if things don't work), and then > merging to the master when the development branch "works". Where would you suggest I find out more about the practicalities of git? I use it for my own stuff quite a bit, but don't know how to do branches and all that. Thanks, SteveT Steve Litt February 2016 featured book: The Key to Everyday Excellence http://www.troubleshooters.com/key ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
Aldemir Akpinar writes: [...] >> BTW: 'Console' and 'terminal' are two rather different things. >> >> > Doesn't your init scripts run on the console? 'The system console' is an I/O device the kernel and scripts started during boot will use for "user interaction", ie, print output supposed to be useful for someone looking at it and possibly, request input from such a someone. This will usually be a terminal device but it doesn't strictly have to. The crucial point here is that "terminal" usually refers to an interactively used (virtual) device providing facilities for text-based user interaction. The system console will usually be one. xterms act as terminal (as do all kinds of other programs providing 'a CLI interface') but it's always about user interaction. Programs don't usually use terminals to talk to other programs (with some exceptions, eg, using expect to drive an interactive program accessed via ssh, to name an at least remotely "not ancient" example). ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
Arnt Gulbrandsen writes: > Emiliano Marini writes: >> Great Scott! Introducing unwanted changes in packages containing the >> word "core", congratulations! >> >> This will break up 99% of the scripts out there... > > Why? > > This won't break common code such as > > for a in *.xml; do > ... > > What it breaks is rubbish such as > > for a in $(ls *.xml); do > ... > > and arguably it doesn't even break that. Indeed. According to the coreutils ML thread, the new default behaviour only applies when isatty(1) is true, ie, when output is to some (presumably interactively used) terminal device. This may effect some code which executes shell code non-interactively via pty and tries to parse the output but that's certainly not "most scripts". The main downside would be that names with embedded ' can now no longer be copied&pasted into application which don't understand shell quoting syntax (and - of course - the guy who's in favour of the change thinks "nobody uses copy&paste in the way" and the grounds that he prefers using GUI file selection dialogs for such tasks) but that's IMHO just a minor nuisance. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
For example, this willl break scripts parsig ls output. On Tue, Feb 16, 2016 at 12:52 PM, Arnt Gulbrandsen wrote: > Emiliano Marini writes: > >> Great Scott! Introducing unwanted changes in packages containing the word >> "core", congratulations! >> >> This will break up 99% of the scripts out there... >> > > Why? > > This won't break common code such as > > for a in *.xml; do > ... > > What it breaks is rubbish such as > > for a in $(ls *.xml); do > ... > > and arguably it doesn't even break that. It behaves differently for some > files (e.g. one named "foo.xml foo.xml", either 13 or 15 characters) but > off the cuff I can't think of any intentions for such code that the old > behaviour allows and the new behaviour doesn't. > > Arnt > > > ___ > Dng mailing list > Dng@lists.dyne.org > https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng > ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
Emiliano Marini writes: Great Scott! Introducing unwanted changes in packages containing the word "core", congratulations! This will break up 99% of the scripts out there... Why? This won't break common code such as for a in *.xml; do ... What it breaks is rubbish such as for a in $(ls *.xml); do ... and arguably it doesn't even break that. It behaves differently for some files (e.g. one named "foo.xml foo.xml", either 13 or 15 characters) but off the cuff I can't think of any intentions for such code that the old behaviour allows and the new behaviour doesn't. Arnt ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
> > > Hmm ... why do you think so? > > BTW: 'Console' and 'terminal' are two rather different things. > > Doesn't your init scripts run on the console? Or the scripts that run on the initial ram disk. Please correct me if I am wrong. -- aldemir ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
Aldemir Akpinar writes: [...] > On coreutils 8.25 ls will wrap filenames with quotes if it includes > whitespace. > > And when people protest, the answers are usual arguments; it just happens > on the console output, or just add -N to your aliases etc. etc. [...] > When this version hits the distributions, it will break many scripts. Hmm ... why do you think so? BTW: 'Console' and 'terminal' are two rather different things. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
Am Dienstag, 16. Februar 2016 schrieb Aldemir Akpinar: > There's a recent discussion on the coreutils mailing list, where some > people complained about the default ls output with latest release. On > coreutils 8.25 ls will wrap filenames with quotes if it includes > whitespace. > > And when people protest, the answers are usual arguments; it just happens > on the console output, or just add -N to your aliases etc. etc. > > Here's a discussion with the response from the maintainer as well: > > http://unix.stackexchange.com/questions/258679/why-is-ls-suddenly-surrounding-items-with-spaces-in-single-quotes > > When this version hits the distributions, it will break many scripts. So > why change the default behaviour? "Pride comes before a fall." ... but maybe they'll come to their senses. um, that's probably not gonna happen ... -- Please do not email me anything that you are not comfortable also sharing with the NSA. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Coreutils 8.25 ls output
Great Scott! Introducing unwanted changes in packages containing the word "core", congratulations! This will break up 99% of the scripts out there... On Tue, Feb 16, 2016 at 11:07 AM, Aldemir Akpinar wrote: > There's a recent discussion on the coreutils mailing list, where some > people complained about the default ls output with latest release. On > coreutils 8.25 ls will wrap filenames with quotes if it includes > whitespace. > > And when people protest, the answers are usual arguments; it just happens > on the console output, or just add -N to your aliases etc. etc. > > Here's a discussion with the response from the maintainer as well: > > > http://unix.stackexchange.com/questions/258679/why-is-ls-suddenly-surrounding-items-with-spaces-in-single-quotes > > When this version hits the distributions, it will break many scripts. So > why change the default behaviour? > -- > aldemir > > ___ > Dng mailing list > Dng@lists.dyne.org > https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng > > ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
[DNG] Coreutils 8.25 ls output
There's a recent discussion on the coreutils mailing list, where some people complained about the default ls output with latest release. On coreutils 8.25 ls will wrap filenames with quotes if it includes whitespace. And when people protest, the answers are usual arguments; it just happens on the console output, or just add -N to your aliases etc. etc. Here's a discussion with the response from the maintainer as well: http://unix.stackexchange.com/questions/258679/why-is-ls-suddenly-surrounding-items-with-spaces-in-single-quotes When this version hits the distributions, it will break many scripts. So why change the default behaviour? -- aldemir ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Avoid 'git commit -m ...' every time code is tested after editing.
On Mon, Feb 15, 2016 at 09:35:55PM -0500, Steve Litt wrote: [cut] > > > > Commit often. Branch whenever needed needed. Merge when it > > works. Release when "perfect" (the last one should be really > > considered with a pinch of salt :P). > > When a version is a release, don't you just give it a tag? > Yes Steve. My point was that git is made to maintain the history of a project, and to help you avoiding disasters, or containing their impact. git is useful only if you commit fairly often, otherwise it is better to use a backup system, rather than a revision control system... So, there is no point into committing "only working stuff", because this is the best way to mess things up and/or to lose hours of precious work. As it has been suggested by several others, whenever you need to have the master branch to always be "working", it's far better to do development in other branches, committing there as frequently as needed (i.e., even if things don't work), and then merging to the master when the development branch "works". My2Cents KatolaZ -- [ Enzo Nicosia aka KatolaZ --- GLUG Catania -- Freaknet Medialab ] [ me [at] katolaz.homeunix.net -- http://katolaz.homeunix.net -- ] [ GNU/Linux User:#325780/ICQ UIN: #258332181/GPG key ID 0B5F062F ] [ Fingerprint: 8E59 D6AA 445E FDB4 A153 3D5A 5F20 B3AE 0B5F 062F ] ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
