[fossil-users] VCS Theory

[Scott Doctor]

I am looking for information about the theory of VCS that is 
being used for systems such as Fossil, Git... Not so much the 
how-to-use, but the concepts and issues.


Any suggestions of either links to something like wikipedia 
pages or a well written book I can find at the university library.


--

Scott Doctor
sc...@scottdoctor.com
--

___
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] server returned an error - clone aborted

[Warren Young]

On Nov 18, 2015, at 9:59 PM, fossil@9ox.net wrote:
> 
> $ fossil version
> This is fossil version 1.33

1.34 is out.  That said, I tested with an even older Fossil on Cygwin, and it 
works for me.

> $ fossil clone ssh://nrb@101.102.103.104//var/repos/test C:\test.fossil

I see a bunch of potential problems here.  

First, you say you’re using stunnel but then you give an ssh URL.  Shouldn’t 
that be https instead?  Can you in fact ssh into that server from your Cygwin 
installation?  Can you “clone” the repo via scp, or does only Fossil fail?

Second, you’re using DOS paths with Cygwin.  You should be saying something 
like /cygdrive/c/test.fossil here.  Not only are you making Cygwin process the 
drive letter for you in a non-POSIX way, the backslash will be interpreted by 
the shell, so that if your clone *did* work, it would end up in a file called 
C:test.fossil in the current directory, not in the root of your C drive.

Third, writing a SQLite DB to the root of the Windows system drive via Cygwin 
may cause locking errors due to the way Windows permissions and Cygwin/POSIX 
file locking interact.  You should be writing it to some user-owned directory.

If you work through all of that and it’s still failing, add something like this 
to ~/.ssh/config:

Host 101.102.103.104
   LogLevel VERBOSE

You should get a couple of screenfuls of debug noise from ssh when you retry 
the clone.  (Be sure the config file is mode 600 and the .ssh directory mode 
700!)

> Windows:
> C:\Users\nrb>"c:\Program Files (x86)\Fuel\fossil.exe" version
> This is fossil version 1.34 [62dcb00e68] 2015-11-02 17:35:44 UTC
> 
> nrb@Lenovo-PC ~
> $ fossil clone ssh://nrb@101.102.103.104//var/repos/test C:\test.fossil

I don’t see how you’re forcing the Cygwin bash instance to run the Fuel version 
of Fossil here.  You should be passing a fully-qualified path:

$ '/cygdrive/c/Program Files (x86)/Fuel/fossil.exe' clone ...


___
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] VCS Theory

[Warren Young]

On Nov 19, 2015, at 12:51 PM, Scott Doctor  wrote:
> 
> I am looking for information about the theory of VCS that is being used for 
> systems such as Fossil, Git... Not so much the how-to-use, but the concepts 
> and issues.

Hmmm.  VCSes are pretty strongly on the practical side of software development, 
so I’m not sure that the ivory tower types have done a lot of formal study of 
them.  You frequently hear of new CS grads showing up at their first job having 
never used any VCS.

Still, some digging in the ACM Digital Library turned up these papers, which 
look to be the sort of thing you’re after:

  http://dl.acm.org/citation.cfm?id=1500879.1500959
  http://dl.acm.org/citation.cfm?id=1761628.1761684
  http://dl.acm.org/citation.cfm?id=1789074.1789084
  http://dl.acm.org/citation.cfm?id=1267680.1267708

I rejected a bunch of hits along the lines of “why use a VCS” or analyses of 
the practice and customs associated with a particular VCS, because you’ve 
already drunk the kool-aid and don’t seem to be interested in anthropology.
___
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] VCS Theory

[Richard Hipp]

On 11/19/15, Scott Doctor  wrote:
> I am looking for information about the theory of VCS that is
> being used for systems such as Fossil, Git... Not so much the
> how-to-use, but the concepts and issues.
>
> Any suggestions of either links to something like wikipedia
> pages or a well written book I can find at the university library.

I'm working on a list of (a) problems that VCS solves for me and/or
(b) capabilities that VCS provides for me.  This is what I have so
far:

(1) The ability to exactly reproduce any historical version of the project.

(2) Provide an audit trail for all changes - the ability to trace
every byte of the project back to the date and time when it was first
entered, together with a description of why it was entered.

(2a) Security - the ability to detect unauthorized changes to the project

(3) Disconnected operation - the ability to continue doing productive
work even when off network.

(4) The ability to work simultaneously on multiple branches.

(5) Situational awareness - reporting tools that quickly and clearly
show the status and history of the project.

(6) Continuity - no single point of failure - all project history is
preserved redundantly in geographically distributed locations.

-- 
D. Richard Hipp
d...@sqlite.org
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

[fossil-users] allow downloading wiki page attachments for users without Read permission

[Eric Rubin-Smith]

A user with the following permission flags:

bcfhjkmnprtw

for a site in which the virtual users Reader, Developer, Anonymous and
Nobody have no default privileges, cannot download an attachment directly
from a wiki page attachments list.  The list points to URIs such as this
one:

https://www.example.com/artifact/66a9b2083b1ad09a

This is because the Read permission is needed to download a raw artifact.

However, the same user may access the attachment via a different URI:

https://www.example.com/attachdownload?file=file_name.exe=Wiki+Page+Name

This bounces the user through a code path that does this (attach.c):

  g.perm.Read = 1;
  cgi_replace_parameter("name",zUUID);
  if( fossil_strcmp(g.zPath,"attachview")==0 ){
artifact_page();
  }else{
cgi_replace_parameter("m", mimetype_from_name(zFile));
rawartifact_page();
  }

i.e. pretend the user has read permission for this request, and press on to
download the thing.

So it would be more convenient for the wiki page to link to the permissive
URI rather than the restrictive one.

Note that a similar issue exists with tickets.  A user permitted to view
tickets should be able to conveniently download the objects attached to the
ticket.
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users