Hello,

Recent discussion of private branches led me to discover one interesting
detail about private branches.

When merging in the private branch, Fossil does not add a P-card for the
private  branch, which  seems sensible  to me,  however, it  does add  a
T-card to close the private branch. Here's an example manifest:

C done
D 2016-05-12T03:04:36.959
F file 2a48c062c41d247e6b0411d24208ecec768c16b9
P dbb9906b61d2a92c9a011ce574a4ba2314ad486d
R 1b389beb678c047e0fc5b2ae93d53165
T +closed 5b29c5390e1023c2c0a56b5d1fd259d843618616
U amb
Z c5370df5ec7bc8d8eaa51e2eb36f7796

If I'm not  mistaken, the extra P-card is not  included to avoid phantom
artifacts  from showing  up in  repositories  to which  this checkin  is
synchronized, but I suspect that the presence of the T-card will obviate
this.

This leads  me to the conclusion  that if one has  private branches, one
can never close them using --integrate,  because to do so will introduce
artifacts  which cannot  be resolved  and which  will eternally  clutter
repositories  with phantom  artifacts.  Notice that  there  is only  one
phantom,  and it's  the  one  from using  --integrate  (in the  manifest
above):

sqlite> SELECT * FROM phantom, blob WHERE phantom.rid = blob.rid;
14|14|0|-1|5b29c5390e1023c2c0a56b5d1fd259d843618616|

The same is not true when simply  adding a control artifact to close the
private branch:

$ fossil art dd982a405c21c2900449f0c4f63042f44f2c78c6
D 2016-05-12T03:17:20.743
T +closed 4c6005c80b40790f379385fb8b4a848f194ccb2d
U amb
Z 93f4010059e478f88d4d7d78f33a1d26

Which is private:

sqlite> SELECT * FROM private, blob USING (rid) WHERE uuid = 
'dd982a405c21c2900449f0c4f63042f44f2c78c6';
21|12|118|dd982a405c21c2900449f0c4f63042f44f2c78c6|

Is this  something that should  be documented? Or should  --integrate be
modified to add  a separate control artifact when private  content is in
use?

Thanks,

Andy
-- 
TAI64 timestamp: 400000005733fa3a


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

Reply via email to