[00:07] mwhudson: the puller has to write to submit requests :) [00:07] oh truw [00:07] i was thinking on the wrong level [00:20] mwhudson: that's it [00:21] i think i can safely promise that if the direction is read or write it will only mean that we're actually reading or writing [00:23] jelmer: is that new dulwich branch pushed yet? [00:24] mwhudson: dankje [00:24] poolie: you can? do you mean reading or writing to a socket? [00:24] mwhudson: argh, wrong channel [00:24] :) [00:24] mwhudson: I meant yes [00:24] wrong language too [00:24] ok [00:24] lifeless: can't I? yes. [00:24] fun and games with hardware: [00:24] https://bugs.edge.launchpad.net/ubuntu/+source/dmraid/+bug/372170 [00:24] Ubuntu bug 372170 in dmraid "intel isw raid metadata at odd offset" [Undecided,In progress] [00:25] my new machine :( [00:25] jelmer: r296 is good now? [00:25] mwhudson: yep [00:25] jelmer: thanks [00:29] lifeless: ping [00:29] kingos: hi [00:29] lifeless: hi. sorry to keep bugging you ... [00:30] lifeless: any update on that bug of mine? I have some colleagues who would massively appreciate an increase in performance of our bzr setup! [00:30] kingos: no - swamped with network issues - three critical bugs [00:30] kingos: I do want to get to it [00:30] lifeless, igc, jml, others: spiv is still away today [00:30] lifeless: ah ok thanks. are you the only person who could help? [00:31] poolie: thanks. [00:31] kingos: I don't think I am [00:31] kingos: vila might be in a position to help sooner though [00:31] lifeless: I imagine it will be quite tricky, something to do with *ghosts* I imagine from my infrequent reading of email lists (not that I know what they are)! [00:34] I would have a go myself if I had the first idea where to start. [00:38] which bug is this? [00:39] bug #356028 [00:39] Launchpad bug 356028 in bzr "bzr check fails with KeyError" [Undecided,New] https://launchpad.net/bugs/356028 [00:40] we have a problem with our repository that is stopping us from upgrading. [00:40] jelmer: I mailed you a consistency issue the other day [00:40] jelmer: is it the same one as inkscape triggers? [00:46] lifeless: no, I'm not sure what that one is exactly [00:46] lifeless: did you see my reply? [00:47] no [00:47] * lifeless looks [00:47] how do we move the inkscape one forward [00:48] lifeless: somebody with enough knowledge of the internals needs to have a look at why bzr thinks it needs the lhs parent even if that's not necessary for the delta [00:49] jelmer: can I help you learn that knowledge? [00:49] lifeless: lhs parent of a text [00:54] lifeless: I'd rather focus on bzr-git for now, not sure I want to dive into pack internals === dereine is now known as dereine[OFF] [00:58] brb, fooding [01:13] morning [01:41] abentley: is BB down? [01:53] hello igc [01:53] hi poolie [02:10] poolie: I can't skype yet on this new machine [02:11] poolie: and as we haven't stood up; let me say I'm working on https://bugs.edge.launchpad.net/bzr/+bug/360791, and after that I plan to go and redo check; check is part of the upgrade process and its way to slow [02:11] Ubuntu bug 360791 in bzr/1.14 "get_stream on stacked branch causes "Error received from smart server: ('NoSuchRevision',)"" [Critical,In progress] [02:12] I don't know that I can make it hugely faster, but I want to stab at it [02:23] that makes sense [02:24] last time i looked it was very inefficient, reading things several times [02:24] it's pretty old code [02:24] so i think you'll find some tasty fruit [02:24] i'm reading the subtree related threads [02:24] i mean nested tree [02:39] igc: restarted. [02:39] abentley: thanks [02:41] poolie: did I tell you, my new machine is 4-core on die? [02:41] yes [02:45] still buzzing about that bit :) [02:45] hyperthreading on top of that [02:47] lifeless: *8* cpu whatsists!?!? [02:47] spm: yup [02:47] *awesome*! [02:47] spm: i7 920 [02:48] now I just need to get eucalpytus up locally, then I can use ec2test @ home :P [02:48] ha! [02:50] BasicOSX: uhm [02:50] BasicOSX: 1.13.2 is looking rather close to bzr.dev ? [02:52] BasicOSX: rev 4113 looks to me like you merged _all_ of bzr.dev, or bzr1.14 or something [03:02] * poolie tries to avoid being distracted to read about i7s [03:04] :> [03:04] I have a fix for bug 360971 [03:04] Launchpad bug 360971 in gnome-app-install "gnome-app-install crashed with ValueError in _refilter() (dup-of: 339148)" [Undecided,New] https://launchpad.net/bugs/360971 [03:04] Launchpad bug 339148 in gnome-app-install "gnome-app-install crashed with ValueError in _refilter()" [High,Incomplete] https://launchpad.net/bugs/339148 [03:04] erm [03:04] not that [03:05] I have a fix for bug 360791 [03:05] Launchpad bug 360791 in bzr/1.14 "get_stream on stacked branch causes "Error received from smart server: ('NoSuchRevision',)"" [Critical,In progress] https://launchpad.net/bugs/360791 [03:05] yes, that [03:11] yay way to go [03:17] lifeless: i got approval for the patch recommending use of lp reviews so if you don't yell i'm going to merge it [03:17] if there are questions or problems we can improve it later on [03:17] thumper: ^^ [03:18] dogfood tastes good [03:19] poolie: I saw the patch :) [03:20] poolie: I would like more docs myself; its fine to say we shouldn't document it in that file, but we _should_ link to appropriate docs [03:20] well, i did link to 2 or 3 existing docs [03:20] poolie: oh, I may have missed that [03:20] if there are more that should be linked, or written and linked, let's do it [03:20] i also filed a bug asking for more detail, [03:21] bug 372057 [03:21] Launchpad bug 372057 in launchpad-code "code review help doesn't mention making proposals by mail" [Undecided,New] https://launchpad.net/bugs/372057 [03:21] np [03:25] jml: is there a 'reviews I need to do' page in lp? [03:25] lifeless: not cross-project. [03:26] a) is there a bug asking for one? Is there one per-project that lists 'reviews I need to do' then? [03:26] lifeless: there's https://code.launchpad.net/bzr/+active-reviews, for example [03:26] (maybe without the hyphen) [03:26] lifeless: yes, there's definitely a bug asking for one. [03:26] rockstar was working on it [03:26] I might well have filed it :) [03:27] lifeless: https://code.edge.launchpad.net/bzr/+activereviews [03:27] ok [03:27] jml: if you could toss me the bug asking for it, I'll me-too it [03:28] mwhudson, I'm working on it right now. :) [03:28] lifeless: https://bugs.edge.launchpad.net/launchpad-code/+bug/325985 [03:28] Ubuntu bug 325985 in launchpad-code "Page that shows reviews I *must* do and reviews I *can* do" [Medium,Triaged] [03:45] lunchish stuff then check() [03:51] * igc lunch [04:04] how can there get to be reviews one *must* do ??? [04:05] * SamB wonders if this involves a phb:employer-of RDF relation [04:06] * SamB wonders why you people eat lunch at such odd times [04:06] SamB: reviews someone has specifically asked *you* to do, as opposed to asked a team you are in to do [04:07] mwhudson: OHHH [04:07] I see [04:07] that sounds an awful lot simpler to implement than what I was thinking ;-) [04:07] heh [04:08] "if you haven't done this review by the end of next week we will kidnap your first born child" ? [04:08] so it's "reviews wanted of *me*" vs "reviews where I'll *do*" [04:09] push vs pull [04:10] yar [04:17] igc, hi? [04:18] hi poolie [04:18] lifeless, mwhudson: lp-needs is ok by me [04:20] igc, can we have a chat? [04:21] poolie: of course [04:22] poolie: http://paste.ubuntu.com/165270/ [04:22] poolie: just call me on pots or skype when you're ready [04:22] poolie: could you very quickly eyeball that [04:22] ok [04:22] poolie: I'm seeking a 'sounds plausible' or 'eww' [04:22] lifeless, others: btw [rfc] how about deleting the very old pre-brisbane-core performance documents from the tree? [04:23] poolie: many are still relevant; I do agree we should houseclean a little [04:23] yes, of course keeping the relevant ones [04:23] that pastebin seems plausible [04:24] the other things i have in mind are: [04:24] there are some checks that may be format-specific, some that may be done more efficiently in a format-specific way, but some should be done regardless of format (this is vague) [04:25] and also that it may be useful for it to not abort on the first error if possible but rather tell us if there's more than one [04:25] as far as ui i'd like it to be more clear by default if everything is ok [04:25] at the moment normal output sometimes disturbs people [04:26] so something like "checked %d revisions in %d branches; everything's ok" [04:26] and then otoh if there is a problem giving data that will help with a bug report with few roundtrips wbn [04:26] naturally ;) [04:26] i have all those things in mind [04:27] at the moment I'm trying to architect a flow that will be easy to code and debug, but scale well etc [04:27] which is what that doc trys to get at [04:27] I will add those other things to the doc though, because explicit better than implicit [04:27] lifeless: should check become an alias for reconcile --dry-run one day? [04:29] igc: maybe; reconcile was written with the idea that there are 'small issues' and 'big issues' and reconcile is the 'big issue hammer' [04:30] but perhaps this is wrong [04:33] poolie: I believe http://code.aaronbentley.com/bzr/bzrrepo/nested-trees-loom is the public branch for nested trees === verterok_ is now known as verterok [04:50] poolie: you seem to be repeatedly mailng the same mail [04:51] poolie: I'm up to 4 copies, letting you know now ;P [05:06] sorry [05:07] i kept getting "failed to send the message" [05:58] poolie: when you get a chance, can you land http://bundlebuggy.aaronbentley.com/project/bzr/request/%3Ce01316480903232138n34d2e189w71e7e8ee1f7e255a%40mail.gmail.com%3E please? [06:13] How do you people have the user accounts on their server configured for a central bzr repo accessible via ssh? [06:13] I'm curious how file permissions work in this case [06:14] My goal is to setup an AWS EC2 machine with an EBS to house a central repo for my company's code. There is just two of us with SSH access to the EC2 machine and we want to setup central (master) bzr branches there [06:40] igc, oh thanks [07:08] hi all [07:09] lifeless: 3 things [07:09] vila: yo! [07:09] eferraiuolo: hi; uhm I use lp :) [07:09] 1) Most important first: have fun with your new i7 :-) [07:09] vila: 6 minutes for parallel test [07:10] vila: I suspect your machine is faster or something [07:10] lifeless: 3 here :) i7 965 @3.2 GHz [07:10] ah, i got the 920 [07:10] 2.6G [07:11] hmm, doesn't explain the whole difference then, may be the SSD helps too [07:11] oh, if youhave an ssd that is likely it [07:11] eferraiuolo: what he said; otherwise add accounts for all the other people, make a group-owned directory for the repository and off you go [07:11] I'll try on a tmpfs [07:11] and 12GB RAM... [07:12] but 4 should be far enough :) [07:12] lifeless: 2) Testing and lock noise stipple, can you give a bit of feedback ? [07:12] 3) PQM and unicode [07:13] vila: I got 3GB, for trichannel mode [07:13] for 2- what you and john said sounds fine [07:14] I'm fine with a method to tell the test suite we're about to get rid of a lock [07:14] as for break-lock, I think it should fire a release hook event [07:14] lifeless: it seems that pqm gives ANSI_X3.4-1968 as fs encoding, that means all UnicodeFilenameFeature depending tests are skipped, I thought you filed an RT ticket about that ? [07:15] lifeless: ok [07:15] firing a release hook event in break-lock would make tests that do a break-lock line up, I think? [07:16] vila: I did, let me check [07:16] I should look into that but, a priori, we don't try to balance lock/unlocks precisely (as in cheking the actual locks), we just count the locks [07:16] * * * * * LANG=en_GB.utf8 PQM_CONFIG=/home/pqm/pqm-config/bzr-pqm.conf USER=pqm HOME=/home/pqm PYTHONPATH=/home/pqm/pqm:/home/pqm/source/bzr.dev:/home/pqm/lib/python /home/pqm/pqm/bin/pqm --run --cron [07:17] vila: counting is a rough approximation [07:17] vila: its hard to have the count line up and the physical locks not match [07:17] lifeless: I understand that, on the other hand.. yeah, what you said :) [07:17] specifically, to get a release we have to have locked it first [07:18] lifeless: break lock firing a release sounds appealing, I'll try that first [07:19] so all the releases can be trusted to have had to have a matching acquire; the question then becomes can we have an acquire that doesn't hook properly; or similar [07:19] lifeless: setting LANG doesn't seem to be enough or is not taken into account then: [07:19] yeah [07:19] I suspect dchroot-run or something is mangling it [07:19] Unable to represent path u'\u05e9\u05dc\u05d5\u05dd' in filesystem encoding "ANSI_X3.4-1968" [07:19] ...non_ascii.TestNonAscii.test_ignored(iso-8859-1) SKIP 1ms [07:19] spm: ^ [07:20] and that's from the "normal" run, not the ascii one [07:20] yah [07:20] whee. fun. [07:20] spm: there was a ticket about this, so I'm just going to ask ;) [07:20] I'll tend to consider that encoding problem rather critical as it means we can have some serious regressions go unnoticed [07:21] lifeless: :-) So... what's the problem per-se? as I'm coming a bit cold at this. I gather is a LANG/dchroot issue? [07:21] spm: we set lang around pqm [07:22] it doesn't seem to propogate down to the running tests [07:22] spiv: ... and python think the file system encoding is ANSI_X3.4-1968 (similar to ascii AIUI) [07:22] spm: what we need is LANG set to something present in the chroot, when make is invoked in the chroot [07:22] eerk, s/spiv/spm/ [07:23] vila: is cool - we're used to it - well I am, can't speak for Andrew :-) [07:23] lifeless: hmmmm..... [08:09] lifeless: how about modifying the "precommit_hook=/home/pqm/bin/kill-chroot-processes.sh /srv/pqm.ubuntu.com/chroot-amd64 && /home/pqm/bin/dchroot-run make check" to add the LANG ==> dchroot-run LANG=blah make check ??? [08:11] spm: lets do that [08:12] spm: but please do it with vila [08:12] so that he can send a patch through immediately; cause if it fails it will need to be rolled back [08:13] lifeless: sounds like a plan. [08:13] vila: still around? and able to submit a patch on demand? :-) [08:14] spiv: what kind of patch ? [08:16] AAArgh [08:16] vila: whatever it was that caused the UTF fail earlier. I'll just modify the pqm submit [08:16] spiv: what kind of patch ? [08:16] I wasn't going to say anything :-) [08:16] shudder 'sp' still matches both 'spiv' and 'spm' vila.... [08:16] :-D [08:17] Obviously, the alphabet is conspiring against you. [08:17] vila: you're submitting to pqm against bzr trunk? [08:17] right, so a patch that fails if the fs encoding doesn't support unicode [08:18] I can submit against anything, if you can set up a test branch I can submit against that (but it needs some public access>) [08:18] * vila tries rewiring visual control *before* press-enter control [08:20] vila: cool. ok have modifed for sftp://escudero.canonical.com/srv/www.bazaar-ng.org/rsync/bzr/bzr.dev - so submit to pqm against that whenever you're ready? [08:21] vila: anything really; just want to check that trunk isn't *already* broken [08:21] vila: because if it is it will reject all merges until its fixed. [08:22] lifeless: Oh, I see, I was able to reproduce the problem locally (at least I hope so) and fixed 3 failures where pqm failed on the first, so I'm pretty sure we are safe [08:22] vila: pretty sure != safe :) [08:22] lifeless: indeed [08:22] lifeless: Just wanted to let you know I received your irc message and yes, I was working 1.13.2 and 1.14.1 at the same time so I could have messed something up. I had to work late tonight, just got home (2:22am) and have to be back at work at 8am so I won't be able to look at it until to night. [08:24] BasicOSX: thanks; thats fine. I'm not sure it is safe to do anything anyway. [08:24] spm: submission sent with commit message 'testing pqm unicode support' [08:25] BasicOSX: but it might be something we want to add a step to into the guide ;P [08:25] 1. Do not work on 2 releases at the same time, BasicOSX did this is borked the 1.13.2 release. [08:26] That seems direct and short-n-sweet [08:27] vila: ta [08:27] What I hate about commercial develop, product goes out the door when marketing says it's time to go. /blah... off to bed [08:28] BasicOSX: s/commercial// and s/marketing/sponsor/ :-) This is not always a bad thing but the consequences are rarely well understood [08:32] spm: pqm.bazaar-vcs.org says: Coming up [08:32] 1. [08:32] Wed May 6 08:24:24 2009 UTC: Vincent Ladeuil , Request for non-PQM managed branch. [08:34] spm: if it's simpler I can send the patch against bzr.dev, it's a one-liner, if it fails, your solution is wrong, if it succeeds, your solution is right and I'll send a submission to revert my patch [08:38] vila: that sounds fair. I can easily remove items in the queue... [08:39] spm: ok, resubmitted [08:44] vila: and old submission removed [08:44] vila: how long does pqm take to test bzr these days? [08:45] pfew, no idea, it's fire and forget for me :) [08:45] heh [08:46] oki we'll it's dinner for me. I'll keep checking. if it really breaks and I don't respond soon - ring me via the staff directory or something :-) [08:47] spm: enjoy your dinner [08:47] hello [08:49] I'd like to share my logsearch plugin (in early infancy) [08:49] for some review and suggestions for better API [08:52] zyga: send a annoucement to bazaar@lists.canonical.com and may be bazaar-announce@lists.canonical.com [08:52] zarq: you'll get a wider audience :) [09:00] vila, i will? [09:00] Ok, I'll push the code to a launchpad branch first [09:00] zarq: ghaa, sorry, that was for zyga of course :-/ [09:00] heh ok [09:00] * vila kills a chicken or two [09:02] Hey gang [09:04] hi the tip of a revision are e.g. the 2 revisions which were merged together to one? [09:08] limor: the tip is the most recent revision in a branch, a revision can have several parents, one for usual commits, at least two for merges [09:09] ah ok [09:20] zarq: ocean's wide! [09:20] without the apostrophe even [09:21] spm: pqm is now running the ascii part so your fix seems good [09:21] vila: oh sweet! [09:22] vila: I suspect you killed the correct type of chickens. [09:22] LOL [09:25] spm: 40 minutes or so [09:25] lifeless: ta [09:30] hey gang, quick question: I'm a python and bzr noob, so bear with me: [09:30] bzr-svn isn't working for me, so I'm running "bzr selftest svn" [09:30] But it complains "bzr: ERROR: No module named tests [09:30] Is this a Python module or a bzr module? [09:32] ah ha "bzr: ERROR: No module named tests [09:32] pardon me [09:32] found it: bzrlib.tests module [09:34] vila: that appears to have finished - all good at your end? [09:50] vila: sorry I didn't get to review your http auth patch today [09:50] vila: I'll do it tomorrow if no-one beats me to it [09:50] * igc dinner [09:50] spm: yup, I'll submit a patch to revert [09:50] igc: np, enjoy your dinner [09:51] vila: thanks! [09:51] Magneus: you'd better try 'bzr selftest' or only 'bzr selftest -s bp.svn' [09:54] spm: reverting patch sent [09:54] lifeless: now pretty sure == safe ;-P [09:54] Will do, thanks vila. It seems that bzr can't find my selftest module, however. [09:55] Is there an environmental variable I need to set? [09:55] Magneus: oh, I misread, what does bzr version says ? [09:55] Oh, wait, you're on windows ? [09:55] Gentoo [09:55] 1.14 [09:56] bzr 1.14.1 on python 2.6.1 (linux2) [09:57] Magneus: if that is a gentoo bzr, maybe they've stripped the tests out [09:57] ok, 'bzr version' should tell you where bzrlib is installed, there you should find a test directory unless the tests.. what LarstiQ said :)( [09:58] well, I ls'ed the bzrlib folder [09:58] and found that there is a healthy-looking tests folder [09:58] full of .py files [09:58] Magneus: good, good :) [09:58] which is what has me thinking it's something like an env variable [09:59] Magneus: there usually aren't any env variables involved with running the tests [09:59] Magneus: nope, nothing needs to be set 'selftest' is a first-class command :) [09:59] blah [09:59] Magneus: your ~/.bzr.log might have some clues as to what is going on [09:59] I wonder why it's not finding it, then [09:59] Ah! log files. good stuff [09:59] ty [10:02] ah ha, looks like it's wanting pycurl [10:02] is that requisite for tests? [10:03] Magneus: only for https, it is not strictly required [10:03] or well, not even then per se [10:03] Magneus: so no [10:03] understood [10:04] also getting an error "ImportError: cannot import name ForeignBranch [10:04] Magneus: that sounds like bzr-svn [10:04] hello folks.. [10:05] is there any way to avoid giving the pass in plain text in bazaar.conf when using 'smtplib' ? [10:05] best, can I make it to prompt me ? [10:05] LarstiQ: right you are. I realize I had left the .6 version in my plugins folder just now >< [10:06] much better. back to "no module named tests" [10:07] Magneus: is that now bzr-svn complaining it can't find its tests? [10:07] Nah. simply a "bzr selftest" call [10:07] Let me know if you want to see the traceback [10:07] pastebin it please :) [10:07] http://pastebin.com/d123100bf [10:08] ok, so it's the xmloutput plugin [10:08] ah [10:08] I can axe it if need be [10:08] so xmloutput is trying to import its own tests folder? [10:09] my branch is now online with the exception of exposing some launchpad bug/feature [10:11] Looks like removing xmloutput did the trick [10:12] bzr selftest is running now [10:12] igc: https://code.launchpad.net/~zkrynicki/+junk/bzr-logsearch (if you remember yesterday's chat) [10:13] Magneus: good. If that is the most recent version of xmloutput, it might need filing a bug about that [10:14] Got it. Thanks, LarstiQ. Another question: [10:14] Selftest is finding a lot of little errors, 'str' object has no attribute 'get' - etc etc [10:14] cause for concern? [10:16] Magneus: there aren't supposed to be any errors, I wouldn't get concerned just yet though. [10:18] alrighty. I'll poke around at bzr-svn, then [10:18] I appreciate the help! [10:18] Ha. Segfault on 'bzr selftest svn' [10:24] vila: cool. I'll leave that LANG fix on bzr.dev (only) overnight. I've emailed the other losas so they know of it's existance if you do need a reversion. G'night! :-) [10:25] spm: I think it's fine to deploy it, unless lifeless objects, I see no reason to *not* have it everywhere [10:25] vila: blame it on me being a paranoid sysadmin - doing a "major" change and then buggering off is... not good :-) [10:26] spm: no pb, take your time and enjoy your night :) [10:26] question: If I use a shared repository the revisions of the branches inside the shared repository will be found in the shared repository .bzr and not in the .bzr of the branche as it would be normally? [10:26] limor: yes [10:32] folks: is there any way to avoid giving the pass in plain text in bazaar.conf when using 'smtplib' ? [10:51] the "bzr commit" I started yesterday, committing all of jaunty's source code in one operation, is still running: it's not using much CPU, it's not doing system calls, and it's virtual memory size is 9+ gigabytes, which is more than the physical RAM of the machine [10:52] it doesn't show any obvious progress going on, either [10:52] why would you do that? [10:52] luks, for fun, to see if bzr can handle it [10:52] well, if you wait a day, it obviously can't [10:52] but I don't think I want to wait much longer [10:53] you could find a smaller project to commit to find that out :) [10:53] any bzr devs want me to do something to see if they can learn something useful from this? [10:55] liw: is that with the development format? [10:56] james_w, --development-rich-root [10:59] does this code look good for a post commit hook: http://pastebin.com/d366a9b34 ? [11:00] liw: I suspect its thrashing [11:01] cornucopic: branch.Branch.hooks.install_named_hook should not be indented [11:02] luks, you mean outside the function ? [11:02] Can one define bzr aliases that involve pipes to external programs? [11:02] (diffstat comes to mind) [11:03] cornucopic: yes, it should be executed when the code is imported [11:03] Lo-lan-do: using the bzr-extcommand plugin [11:03] ok. trying.. [11:04] liw: iotop [11:04] luks: Trying that, thanks [11:04] Lo-lan-do: http://bzr.oxygene.sk/bzr-plugins/extcommand/__init__.py [11:04] there are also some examples [11:07] luks, Thanks it works. Here is what I want to do: Since, I don't want my SMTP password in a plain text file, what I want to do is, post commit, I want to call my own code to send me the commit message using the smtp_connection.py.. Is that a good line of thinking? [11:08] cornucopic: is the plain text password in your plugin a better solution? [11:08] I don't see how are you going to avoid storing the password in plain text [11:09] cornucopic: I send mail via a local mta, then it has [root accessible only] the password for forwarding it outwards [11:09] luks, I can prompt the user for the password and have it stored in smtp_password. [11:10] luks: Yay, it works :-) [11:11] cornucopic: you could just set right file permissions on the config file, it will be accessible only by root or you [11:11] unless you are on windows [11:12] Lo-lan-do: of course it does :P [11:12] luks, No I am on Ubuntu :) file permissions is a easy/indirect solution.. [11:12] luks, however I have another idea now.. [11:13] luks, the file which parses the contents of the smtp credentials..if it comes across a smtp_password, field, then it prompts the user for the password? [11:15] luks, i.e. if the smtp_password field is blank. [11:15] luks, Does it sound reasonable? [11:16] cornucopic: I don't know much about hooking into the config system [11:16]