Re: [fossil-users] RFC before trunking 'usage' command

2013-09-29 Thread Chad Perrin
On Sun, Sep 15, 2013 at 01:52:49PM +0200, Stephan Beal wrote:
 On Sun, Sep 15, 2013 at 12:56 PM, BohwaZ boh...@bohwaz.net wrote:
 
  provides that, but it's ok as long as it's not turned on by default and
  we can clear the history, like before copying a repository file to
  publish it online.
 
 
 It's stored in the checkout db, so it's _never_ synced. The '-c' flag
 clears the history.

My understanding is that the concern raised was about copying (i.e.
shell command cp, or maybe scp), not syncing.

-- 
Chad Perrin [ original content licensed OWL: http://owl.apotheon.org ]
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


Re: [fossil-users] RFC before trunking 'usage' command

2013-09-29 Thread Stephan Beal
On Sun, Sep 29, 2013 at 6:14 PM, Chad Perrin c...@apotheon.net wrote:

 My understanding is that the concern raised was about copying (i.e.
 shell command cp, or maybe scp), not syncing.


It would only be copied if someone copies their checkout db file, which
would be a highly unusual thing to do. Of course, when copying your whole
home dir or some such it would be copied along with anything else, but
there is _certainly_ more sensitive info under one's home dir than the list
of fossil commands one called while under a given checkout. i.e. i don't
personally consider this to be any sort of security thread, but... there
are those with a better eye for this sort of thing. For example, i'm not
going to add saving of command arguments (only the command name) because
BohwaZ pointed out the case of passing a password to the commands which
take a URL, and i do not want to special-case any of the arguments to
handle that (there's always another corner case which will slip by...).

In any case, this feature is very low on the prio list, and it's still not
clear whether it should really be added or allowed to die off quietly.
While i find it an interesting feature, it's certainly not vital and it
does have a number of touchy questions associated with it. Opinions are of
course welcomed.

-- 
- stephan beal
http://wanderinghorse.net/home/stephan/
http://gplus.to/sgbeal
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


Re: [fossil-users] RFC before trunking 'usage' command

2013-09-15 Thread BohwaZ
Le Wed, 11 Sep 2013 22:58:56 +0200, Stephan Beal
sgb...@googlemail.com a écrit :

 If you don't find the idea offensive/abusive, you needn't say
 anything. If you do, please voice your concerns. If there is
 sufficient outcry i'll drop that branch instead of integrating it. If
 we do decide to add it, i will add a config option to toggle it on
 and off (disabled by default).

I don't think I have any use for that as my shell history already
provides that, but it's ok as long as it's not turned on by default and
we can clear the history, like before copying a repository file to
publish it online.

Logging arguments may pose some risk as if you do:

$ fossil remote-url https://bohwaz:secretpassw...@fossil.dev.org/

(yes you can omit the password and fossil will ask you for it, I know)

The password will be logged in cleartext fossil in that case, and that
may be a problem.

 PS: feel free to suggest a better command name than 'usage'. 'history'
 would be the obvious choice (due to it's analog in Unix shells), but
 that command name potentially has better uses in a future fossil
 version, so i don't want to steal that one.

'cmdhist' seems fine for me, as I would think that 'history' provides
the history of last checkins, and 'usage' would provide me with some
help on fossil command line.

-- 
BohwaZ
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


Re: [fossil-users] RFC before trunking 'usage' command

2013-09-15 Thread Stephan Beal
On Sun, Sep 15, 2013 at 12:56 PM, BohwaZ boh...@bohwaz.net wrote:

 provides that, but it's ok as long as it's not turned on by default and
 we can clear the history, like before copying a repository file to
 publish it online.


It's stored in the checkout db, so it's _never_ synced. The '-c' flag
clears the history.


 Logging arguments may pose some risk as if you do:

 $ fossil remote-url https://bohwaz:secretpassw...@fossil.dev.org/


VERY good point. So far that's not implemented, though.


 (yes you can omit the password and fossil will ask you for it, I know)

 The password will be logged in cleartext fossil in that case, and that
 may be a problem.


Definitely. Thanks for pointing that out. (That said, i've only ever used
the password that way when cloning.)

'cmdhist' seems fine for me, as I would think that 'history' provides
 the history of last checkins, and 'usage' would provide me with some
 help on fossil command line.


i agree on all points, i just find cmdhist clumsy to type ;). cmdhist.
Fossil would, however, allow an abbreviation of 'cm', so it wouldn't be
that bad.

-- 
- stephan beal
http://wanderinghorse.net/home/stephan/
http://gplus.to/sgbeal
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


Re: [fossil-users] RFC before trunking 'usage' command

2013-09-12 Thread Chad Perrin
On Thu, Sep 12, 2013 at 12:14:01AM +0200, Stephan Beal wrote:
 On Wed, Sep 11, 2013 at 11:58 PM, B Harder brad.har...@gmail.com wrote:
 
  fossil set tracker [on|off] ?
 
 IIRC Trac was Fossil's predecessor for ticket tracking, and tracker
 always implies tickets to me (as a by-day/by-night developer).
 
 How about...
 
 fossil set nsa-mode on|off

If anyone asks, just say NSA stands for Normal Shell Access.  Then,
to solve the cmdhist doesn't roll off the fingers naturally problem,
we can use nsahist, which I think flows a little better when typing.

-- 
Chad Perrin [ original content licensed OWL: http://owl.apotheon.org ]
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


Re: [fossil-users] RFC before trunking 'usage' command

2013-09-11 Thread Kees Nuyt
[Default] On Wed, 11 Sep 2013 22:58:56 +0200, Stephan Beal
sgb...@googlemail.com wrote:

Hi, all,

Once 1.27 officially goes out the door, i plan on pulling the new 'usage'
command into the trunk (for 1.28). My question to you is: is there anyone
who finds the idea of fossil recording which commands (not their arguments)
and the times of their usage in the local checkout database offensive or
insecure? These data are _only_ local to the current checkout and are never
synchronized. The 'usage' command is only intended for informal personal
statistics gathering, and not for spying on developer's activity within a
given tree (that would be possible if devs have access to each others'
checkout directories).

I wouldn't mind logging the arguments as well, especially for commands
that modify the repository or a stash.
I wouldn't be interested in logging query-class commands, like 
fossil ls, fossil status, fossil timeline

If you don't find the idea offensive/abusive, you needn't say anything. If
you do, please voice your concerns. If there is sufficient outcry i'll drop
that branch instead of integrating it. If we do decide to add it, i will
add a config option to toggle it on and off (disabled by default).

PS: feel free to suggest a better command name than 'usage'. 'history'
would be the obvious choice (due to it's analog in Unix shells), but that
command name potentially has better uses in a future fossil version, so i
don't want to steal that one.

usage - commandtrace or commandhistory or cmdtrace or cmdhistory

Note: on Linux/unix, it doesn't add much, 
  history [-OPTIONS] | grep fossil
will do fine, on MS Windows, I think it really adds value.

-- 
Groet, Cordialement, Pozdrawiam, Regards,

Kees Nuyt

___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


[fossil-users] RFC before trunking 'usage' command

2013-09-11 Thread Stephan Beal
Hi, all,

Once 1.27 officially goes out the door, i plan on pulling the new 'usage'
command into the trunk (for 1.28). My question to you is: is there anyone
who finds the idea of fossil recording which commands (not their arguments)
and the times of their usage in the local checkout database offensive or
insecure? These data are _only_ local to the current checkout and are never
synchronized. The 'usage' command is only intended for informal personal
statistics gathering, and not for spying on developer's activity within a
given tree (that would be possible if devs have access to each others'
checkout directories).

If you don't find the idea offensive/abusive, you needn't say anything. If
you do, please voice your concerns. If there is sufficient outcry i'll drop
that branch instead of integrating it. If we do decide to add it, i will
add a config option to toggle it on and off (disabled by default).

PS: feel free to suggest a better command name than 'usage'. 'history'
would be the obvious choice (due to it's analog in Unix shells), but that
command name potentially has better uses in a future fossil version, so i
don't want to steal that one.

-- 
- stephan beal
http://wanderinghorse.net/home/stephan/
http://gplus.to/sgbeal
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


Re: [fossil-users] RFC before trunking 'usage' command

2013-09-11 Thread Stephan Beal
On Wed, Sep 11, 2013 at 11:26 PM, Kees Nuyt k.n...@zonnet.nl wrote:

 I wouldn't mind logging the arguments as well, especially for commands
 that modify the repository or a stash.


i considered it, but the code overhead just wasn't worth it at that time of
night. One of these days i'll miss having that, add it, and add a -verbose
option which also shows the args.


 I wouldn't be interested in logging query-class commands, like
 fossil ls, fossil status, fossil timeline


Do you mean the SQL code itself? If so, try:

f time -n 1 -sqltrace

One can learn a lot about how fossil works by using that.

usage - commandtrace or commandhistory or cmdtrace or cmdhistory


i think 3 of those might mess up the column alignment in 'help' ;).


 Note: on Linux/unix, it doesn't add much,
   history [-OPTIONS] | grep fossil
 will do fine, on MS Windows, I think it really adds value.


Me, too, but i'm aware that many other people have much greater concerns
regarding any sort of usage protocols. For me it's harmless, but it's not
just about me.

-- 
- stephan beal
http://wanderinghorse.net/home/stephan/
http://gplus.to/sgbeal
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


Re: [fossil-users] RFC before trunking 'usage' command

2013-09-11 Thread Stephan Beal
On Thu, Sep 12, 2013 at 12:08 AM, Ron Wilson ronw.m...@gmail.com wrote:

 On Wed, Sep 11, 2013 at 5:58 PM, B Harder brad.har...@gmail.com wrote:

 fossil set tracker [on|off] ?

 Or 'fossil set cmdhist [on|off]'


i like cmdhist but my fingers don't like typing it :/. cmdhist. Just feels
awkward to type. Kind of like the word awkward. cmdhist. cmdhist. Maybe
it'll grow on me after a while. cmdhist. It's a bit better now, anyway.

-- 
- stephan beal
http://wanderinghorse.net/home/stephan/
http://gplus.to/sgbeal
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


Re: [fossil-users] RFC before trunking 'usage' command

2013-09-11 Thread Ron Wilson
On Wed, Sep 11, 2013 at 5:58 PM, B Harder brad.har...@gmail.com wrote:

 fossil set tracker [on|off] ?

Or 'fossil set cmdhist [on|off]'

Good idea.
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


Re: [fossil-users] RFC before trunking 'usage' command

2013-09-11 Thread Ron Wilson
On Wed, Sep 11, 2013 at 5:56 PM, Kees Nuyt k.n...@zonnet.nl wrote:

 [Default] On Wed, 11 Sep 2013 23:32:10 +0200, Stephan Beal
 sgb...@googlemail.com wrote:
  usage - commandtrace or commandhistory or cmdtrace or cmdhistory
 
  i think 3 of those might mess up the column alignment in 'help' ;).

 Yeah, I'm happy with 'cmdhist' or 'chghist' or 'modhist' or 'shhist' ;)


I'd say 'cmdhist'. Is both short and clear what it does. While 'usage' is
technically correct as well, in Unix/Linux/POSIX, usage generally refers to
how-to-use, not history if use.
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


Re: [fossil-users] RFC before trunking 'usage' command

2013-09-11 Thread B Harder
fossil set tracker [on|off] ?
On Sep 11, 2013 2:32 PM, Stephan Beal sgb...@googlemail.com wrote:

 On Wed, Sep 11, 2013 at 11:26 PM, Kees Nuyt k.n...@zonnet.nl wrote:

 I wouldn't mind logging the arguments as well, especially for commands
 that modify the repository or a stash.


 i considered it, but the code overhead just wasn't worth it at that time
 of night. One of these days i'll miss having that, add it, and add a
 -verbose option which also shows the args.


 I wouldn't be interested in logging query-class commands, like
 fossil ls, fossil status, fossil timeline


 Do you mean the SQL code itself? If so, try:

 f time -n 1 -sqltrace

 One can learn a lot about how fossil works by using that.

 usage - commandtrace or commandhistory or cmdtrace or cmdhistory


 i think 3 of those might mess up the column alignment in 'help' ;).


 Note: on Linux/unix, it doesn't add much,
   history [-OPTIONS] | grep fossil
 will do fine, on MS Windows, I think it really adds value.


 Me, too, but i'm aware that many other people have much greater concerns
 regarding any sort of usage protocols. For me it's harmless, but it's not
 just about me.

 --
 - stephan beal
 http://wanderinghorse.net/home/stephan/
 http://gplus.to/sgbeal

 ___
 fossil-users mailing list
 fossil-users@lists.fossil-scm.org
 http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users


Re: [fossil-users] RFC before trunking 'usage' command

2013-09-11 Thread Kees Nuyt
[Default] On Wed, 11 Sep 2013 23:32:10 +0200, Stephan Beal
sgb...@googlemail.com wrote:


 I wouldn't be interested in logging query-class commands, like
 fossil ls, fossil status, fossil timeline


Do you mean the SQL code itself? 

No, I now realize query is a confusing word.

I mean the class of commands that doesn't change anything to a
repository or a checkout is not interesting for logging.

So,
fossil {ls|status|timeline|changes|extra| ... }
don't have to be logged,

fossil {pull|push|sync|open|commit|stash| ... }
are interesting though.


 usage - commandtrace or commandhistory or cmdtrace or cmdhistory

 i think 3 of those might mess up the column alignment in 'help' ;).

Yeah, I'm happy with 'cmdhist' or 'chghist' or 'modhist' or 'shhist' ;)

-- 
Groet, Cordialement, Pozdrawiam, Regards,

Kees Nuyt

___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users