kbulgrien | $ bzr status --no-recurse | 00:54 |
---|---|---|
kbulgrien | bzr: ERROR: no such option: --no-recurse | 00:54 |
kbulgrien | $ bzr status --no-recurse | 00:54 |
kbulgrien | bzr: ERROR: no such option: --no-recurse | 00:54 |
kbulgrien | jelmer: no | 00:55 |
kbulgrien | $ bzr status --no-recurse | 00:55 |
kbulgrien | bzr: ERROR: no such option: --no-recurse | 00:55 |
jelmer | kbulgrien: I get the point, no need to repeat it three times :) | 00:55 |
kbulgrien | mistake | 00:55 |
kbulgrien | sorry | 00:55 |
jelmer | kbulgrien: I must be mistaken, I thought bzr status had a --no-recurse option too | 00:55 |
kbulgrien | was scrolled up and thought middle click wasn't pasting. | 00:56 |
kbulgrien | well, it gets annoyed when there is read-only stuff... haven't figured out if it bails or simply reports problem, but --no-recurse would avoid that. | 00:57 |
jelmer | kbulgrien: what kind of read-only stuff? bzr status shouldn't write to any files in the working tree | 00:58 |
kbulgrien | weigon sure comes and goes a lot | 00:58 |
jelmer | yeah | 00:58 |
jelmer | weigon: ayt? | 00:58 |
kbulgrien | well, again, regarding ro stuff, its stuff that is probably atypical of what most people do. | 00:59 |
kbulgrien | I'm just changing what I used to do with another vcs over to bzr. | 01:00 |
jelmer | kbulgrien: how is bzr complaining though/ | 01:00 |
kbulgrien | $ bzr status * | 01:00 |
kbulgrien | bzr: ERROR: [Errno 13] Permission denied: '/etc/shorewall/hosts' | 01:00 |
jelmer | kbulgrien: ah, not readable at all | 01:01 |
jelmer | that's different from readonly :) | 01:01 |
kbulgrien | probably another reason I need to look at etc keeper if I am going to use bzr for this | 01:01 |
kbulgrien | yeah, punchy. sorry. been working in the yard all day. am beat. | 01:02 |
jelmer | kbulgrien: and .bzr/ is readable? | 01:02 |
kbulgrien | yes | 01:02 |
jelmer | kbulgrien: why isn't /etc/shorewall/hosts not openable in tha tcase? | 01:03 |
jelmer | presumably it has the same data as /etc/shorewall/hosts | 01:03 |
kbulgrien | because I am not root | 01:03 |
jelmer | kbulgrien: but why is .bzr/ readable for non-root in that case? | 01:04 |
jelmer | kbulgrien: presumably that would allow you to get at data that is otherwise only accessible for root | 01:04 |
kbulgrien | sudo | 01:04 |
kbulgrien | I only use it when absolutely necessary | 01:04 |
kbulgrien | so for controlling those files I sudo, otherwise I work as sysadmin user. | 01:05 |
kbulgrien | That's fine for everything else, but a pain when checking status. | 01:05 |
kbulgrien | It's ok... as long as it isn' | 01:05 |
kbulgrien | t bailing out. | 01:05 |
kbulgrien | But then I'd guess if it wasn't bailing out it would have showed all the stuff. | 01:06 |
jelmer | kbulgrien: why is that file not readable then though? | 01:06 |
kbulgrien | all /etc/shorewall is rw root. | 01:07 |
kbulgrien | I was wanting status in /etc, not etc/shorewall. | 01:07 |
kbulgrien | I didn't need sudo in /etc | 01:07 |
kbulgrien | So I didn't use it, but then it recursed and apparently is bailing out on the unreadable. | 01:08 |
jelmer | kbulgrien: why is it not readable for regular users I mean? Presumably they can get at its contents from .bzr/ anyway. | 01:08 |
jelmer | lifeless: ping | 01:08 |
lifeless | hmm? | 01:08 |
kbulgrien | bzr is operator.root 770, so no, regular users can't get to it. | 01:09 |
jelmer | lifeless: can you perhaps ban weigon, he isn't responding and has been reconnecting every two or three minutes for the last day or so | 01:09 |
jelmer | kbulgrien: then how do you expect 'bzr status' to work as regular user? It needs to access .bzr/ | 01:09 |
kbulgrien | I'm tempted to say forget it, but, I expect to have to do some magic to check status of /etc/shorewall. | 01:10 |
kbulgrien | I have a little bitty front-end script that does the magic - when I need it. | 01:11 |
kbulgrien | My point was that all I wanted to do was check status of /etc, not everything underneath etc. | 01:11 |
kbulgrien | There is no way to limit bzr status to what I want to do so I don't have to use the magic unnecessasrily | 01:11 |
jelmer | kbulgrien: that's bug 287880 | 01:12 |
ubot5 | Launchpad bug 287880 in Bazaar "bzr status: add non-recursive option" [Medium,Confirmed] https://launchpad.net/bugs/287880 | 01:12 |
kbulgrien | ok | 01:12 |
kbulgrien | I need to get more acquainted with launchpad | 01:12 |
jelmer | I think the permissions is a different issue though, but I don't really understand what you're trying to do | 01:12 |
kbulgrien | I'm version controlling "my" edits in an OS. someone told me I should use etckeeper for that, but I've been doing things this way for years with CVS. | 01:13 |
kbulgrien | So that's what I tried doing when I decided to jump over to bzr. | 01:14 |
kbulgrien | The frontend for bzr is tiny compared to the frontend for cvs. | 01:14 |
kbulgrien | It was also giving me a practical at-home experience with bzr to learn it. | 01:14 |
kbulgrien | I have to get comfortable with it on my own before I push it up at work. | 01:15 |
kbulgrien | I don' t have an bzr projects at home at the moment, other than minor server stuff. | 01:15 |
kbulgrien | I realize my usecase is very odd. | 01:16 |
jelmer | kbulgrien: I don't think your usecase is very odd - I'm using etckeeper too | 01:16 |
kbulgrien | I just haven't tried etckeeper. | 01:16 |
jelmer | kbulgrien: my /etc/.bzr/ is (intentionally) not accessible by non-root, since it versions some files that I don't want non-root users on my system to be able to access | 01:17 |
kbulgrien | I fell into doing things the way I had always done them with cvs. | 01:17 |
kbulgrien | I didn't even realize etckeeper was a vcs frontend until someone on here told me to look at it. | 01:17 |
jelmer | kbulgrien: I think one of the differences with CVS is that CVS has a single file with the data on each versioned file. Bazaar stores information on multiple files together in single files (a pack) | 01:17 |
kbulgrien | yes. when I am doing os stuff, I have user is special. | 01:18 |
kbulgrien | I never do os config as my regular user. | 01:18 |
kbulgrien | I cannot even sudo with my regular user. | 01:18 |
jelmer | kbulgrien: can that os config user access /etc/shorewall/hosts ? | 01:19 |
kbulgrien | So this user is non-root, so his repo and sandboxes are not accessible to anyone else. | 01:19 |
kbulgrien | only by sudo | 01:19 |
kbulgrien | my front end script stats the file, gets perms, saves them, tweaks perms so bzr can do what it wants, then puts everything back, so bzr can run non-root too. | 01:20 |
jelmer | kbulgrien: that sounds like a race condition waiting to happen? | 01:20 |
kbulgrien | not when I am the only admin, eh? | 01:20 |
fullermd | Waiting? :p | 01:20 |
kbulgrien | sure, its technically an exposure as it lowers perm tightness, but again, this user is not normal. | 01:22 |
kbulgrien | all it needs is group root accessible. other users aren't group root. | 01:22 |
jelmer | kbulgrien: in that case, why doesn't your script tweak the permissions of /etc/shorewall/hosts? | 01:22 |
kbulgrien | If I was working on etc/shorewall I would have used it. | 01:22 |
kbulgrien | but I was trying to get status of say /etc/* for changes, so didn't need the frontend. | 01:23 |
kbulgrien | It would get very messy to recurse and tweak perms for stuff like that when I didn't want to do it. | 01:23 |
kbulgrien | That would be a disaster waiting to happen. | 01:23 |
kbulgrien | The only problem is that I cannot avoid recurse. | 01:24 |
jelmer | kbulgrien: unlike CVS, bzr versions directories too though. so if you run 'bzr status shorewall' it will recurse into that directory | 01:24 |
jelmer | kbulgrien: I think (but am not sure) if you 'bzr status FILE' it won't access the other files in the directory | 01:25 |
kbulgrien | sure. | 01:25 |
kbulgrien | agreed | 01:25 |
kbulgrien | so I tried bzr status . | 01:25 |
kbulgrien | I guess I could go find . -type f -maxdepth 1 -exec bzr status, but that seems unreasonable. | 01:26 |
fullermd | Well, since bzr versions directories, I'd half expect a non-recursive status of '.' to have to touch shorewall too (not shorewall/foo necessary, but the inode of the dir at least) | 01:26 |
jelmer | kbulgrien: if you give it a directory it will report the status for that directory recursively | 01:26 |
kbulgrien | yes | 01:27 |
kbulgrien | I was grasping at straws and am new to bzr | 01:27 |
kbulgrien | I was amazed that there was no limit recursion option. | 01:28 |
kbulgrien | ie. cvs recurses. but cvs has a limit recursion option. its not the recursion that I mind. | 01:29 |
kbulgrien | It would be odd for it not to recurse. | 01:29 |
kbulgrien | I limit recursion a lot when I use cvs. | 01:31 |
kbulgrien | Probably because I use partial checkouts a lot. something else that I guess bzr doesn't support. | 01:32 |
kbulgrien | Not being able to do partial checkouts is something I am not sure how to work around at work. | 01:33 |
jelmer | kbulgrien: there are 'views', which give you some of the properties of partial checkouts | 01:34 |
* jelmer gets some sleep.. clock just went from 1:59 to 3 here :) | 01:35 | |
kbulgrien | ok | 01:35 |
kbulgrien | thanks, btw | 01:35 |
* kbulgrien thinks it would be nice to be able to define a view in a "reverse" mode. By specifying things that are NOT wanted. | 04:33 | |
kbulgrien | Some times the list of things you do not want to see is a lot smaller than what you do want to see. | 04:34 |
fullermd | Mmph. Views are totally not a useful substitute for partial checkouts. | 04:36 |
kbulgrien | well, that is a given IMO | 04:37 |
kbulgrien | I am having a hard time thinking about using a vcs without partial checkouts. | 04:37 |
kbulgrien | At work, we use a big repo for a product. S/W guys don't want to see the massive VHDL area, Tool 1 guys don't want to have to deal with Tool x stuff, but we want to be able to tag the whole project easily. | 04:39 |
kbulgrien | Tool n has source files that can be 100's mb XML. one guy works with that. the other 10 shouldn't have to have that checked out and updated, status'd, etc. | 04:40 |
kbulgrien | I wanted to convert over to bzr, but that looks daunting without partial checkouts. | 04:41 |
kbulgrien | I think I heard something about externals... might have to look into that as a possibility | 04:42 |
fullermd | There are not ungood reasons why 80% of the time, you don't really want partial checkouts. And other as-good-or-better solutions to the remaining 10%. | 04:42 |
fullermd | So that completely covers the... uh... wait. It's been a long time since I took math, but.. | 04:43 |
kbulgrien | :-) | 04:43 |
kbulgrien | I think it is a bug that bzr status doesn't return status just because it cannot read a file it tries to return status on. It should report the problem and keep going IMO. | 04:49 |
fullermd | That does open up the possibility of a tidal wave of spam. | 04:53 |
fullermd | Of course, that could probably equally well be labelled as giving a silly answer to a silly question... | 04:53 |
=== sagaci_ is now known as sagaci | ||
* kbulgrien thinks bzr status IS a tidal wave of SPAM. I see no good reason for bzr status to give status for the WHOLE branch by default. | 05:31 | |
lifeless | kbulgrien: what should be the default ? | 05:32 |
kbulgrien | I never (actually I suppose almost never) want status on the whole branch when I ask for status below the top of the branch. | 05:32 |
kbulgrien | default should be status at current dir and below... a NORMAL recurse | 05:33 |
kbulgrien | when I bzr commit a low level, it does not commit the WHOLE branch. | 05:33 |
kbulgrien | (does it?) | 05:34 |
* kbulgrien is a newb. I might not have actually tried that. | 05:34 | |
kbulgrien | I guess bzr status . is a reasonable habit to get into, but I still don't like the default. | 05:37 |
kbulgrien | I guess I can see some people might like the default in some sense. | 05:37 |
kbulgrien | I say normal, like rm -rf doesn't wipe out / unless I am in root. | 05:39 |
kbulgrien | This is a default that differs from the norm of most command line tools. | 05:40 |
j605 | hi | 06:57 |
j605 | i was trying to checkout a project, but i get an error. Permission denied(Public Key). I am running the command as an administrator(using sudo). Is there a way around it? | 06:59 |
j605 | figured it out, had to copy my .ssh directory to /root :) | 07:40 |
fullermd | kbulgrien: Actually, it _does_ commit the whole tree... | 14:46 |
fullermd | kbulgrien: 's a change that took me a while to get used to coming from CVS, but I was happy for it. | 14:47 |
Peng | /1/14 | 14:50 |
=== yofel_ is now known as yofel | ||
poolie | hi all | 22:37 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!