Hello, I've tried the trick with marking explicitly parent of incrementally imported changes line, but this does not work fully. First of all, it definitely helps with the timeline, I can see both or all independent lines of incremental imports nicely connected together. I can also see just one resulting leaf, but on update of checkout from such repository fossil fails and removes all files which were not touched by the last incremental import line.
I'm attaching tarball of 3 scripts which clearly shows the behavior and may be used to duplicate/debug it. Put it into some directory and run ./test.sh -- also please modify reconnect.sh first and if your GNU grep is "grep" then please rename "ggrep" to "grep". The scripts were developed on my Solaris workstation where GNU grep is ggrep. Thanks! Karel On Sat, Sep 3, 2016 at 8:26 PM, Karel Gardas <gard...@gmail.com> wrote: > On Fri, Sep 2, 2016 at 8:28 PM, Svyatoslav Mishyn <j...@openmailbox.org> > wrote: >> Maybe `fossil reparent` and then `fossil leaves --recompute` will help.. >> > > Great! This is indeed working as long as there is acceptable to have > additional artifacts -- are those tags? My testing timeline shows > this: > > $ fossil timeline > === 2016-09-03 === > 17:06:00 [4250222b71] Edit [36f4b3ef8d4a245d|36f4b3ef8d]: Add "parent" > with value "71ffe7ef09c0c3d6e05cd6d31b43e2850586323c". (user: karel) > 17:05:12 [533d224a5c] *CURRENT* change 15 (user: > karel.gar...@centrum.cz tags: trunk) > 17:05:11 [5be7e9c3a3] change 14 (user: karel.gar...@centrum.cz tags: trunk) > 17:05:09 [9ffa7c82c9] change 13 (user: karel.gar...@centrum.cz tags: trunk) > 17:05:08 [1f98a7fe09] change 12 (user: karel.gar...@centrum.cz tags: trunk) > 17:05:07 [36f4b3ef8d] change 11 (user: karel.gar...@centrum.cz tags: trunk) > 17:04:08 [fb6393d303] Edit [6abd71f3a3f2dba4|6abd71f3a3]: Add "parent" > with value "2c3bd124e4733718b448835fe9a8767b0bce9067". (user: karel) > 14:30:14 [71ffe7ef09] change 10 (user: karel.gar...@centrum.cz tags: trunk) > 14:30:13 [d8a49f0142] change 9 (user: karel.gar...@centrum.cz tags: trunk) > 14:30:12 [e17627bb9e] change 8 (user: karel.gar...@centrum.cz tags: trunk) > 14:30:11 [5d3ddbd363] change 7 (user: karel.gar...@centrum.cz tags: trunk) > 14:30:10 [6abd71f3a3] change 6 (user: karel.gar...@centrum.cz tags: trunk) > 14:30:09 [2c3bd124e4] change 5 (user: karel.gar...@centrum.cz tags: trunk) > 14:30:08 [a139dda296] change 4 (user: karel.gar...@centrum.cz tags: trunk) > 14:30:07 [da50998d5f] change 3 (user: karel.gar...@centrum.cz tags: trunk) > 14:30:06 [7f46e6e37d] change 2 (user: karel.gar...@centrum.cz tags: trunk) > 14:30:05 [acfb6b6234] change 1 (user: karel.gar...@centrum.cz tags: trunk) > +++ no more data (17) +++ > > > The question is if the process may be somehow automated or even part > of incremental git import? I'm asking since usually the situation > looks as: > > karel@silence:~/tmp/fossil-test/git-incremental-import/workspace-fossil$ > fossil timeline > === 2016-09-03 === > 18:15:39 [30a3929a48] change 10 (user: karel.gar...@centrum.cz tags: trunk) > 18:15:38 [db0ddbb9ba] change 9 (user: karel.gar...@centrum.cz tags: trunk) > 18:15:37 [aa40441d2d] change 8 (user: karel.gar...@centrum.cz tags: trunk) > 18:15:36 [302c071b45] change 7 (user: karel.gar...@centrum.cz tags: trunk) > 18:15:35 [2b36d21f54] change 6 (user: karel.gar...@centrum.cz tags: trunk) > 18:15:33 [8a8dfe4da6] *CURRENT* change 5 (user: > karel.gar...@centrum.cz tags: trunk) > 18:15:32 [7848d4ba85] change 4 (user: karel.gar...@centrum.cz tags: trunk) > 18:15:31 [ca027158d7] change 3 (user: karel.gar...@centrum.cz tags: trunk) > 18:15:30 [92626573e1] change 2 (user: karel.gar...@centrum.cz tags: trunk) > 18:15:29 [bac45c9453] change 1 (user: karel.gar...@centrum.cz tags: trunk) > +++ no more data (10) +++ > karel@silence:~/tmp/fossil-test/git-incremental-import/workspace-fossil$ > fossil leaves > (1) 2016-09-03 18:15:39 [30a3929a48] change 10 (user: > karel.gar...@centrum.cz tags: trunk) > (2) 2016-09-03 18:15:33 [8a8dfe4da6] change 5 (user: > karel.gar...@centrum.cz tags: trunk) > > > > this is after one import git -> fossil and after additional > incremental import git -> fossil. Now I would need to connect > 8a8dfe4da6 and 2b36d21f54 together by: > > karel@silence:~/tmp/fossil-test/git-incremental-import/workspace-fossil$ > fossil reparent 2b36d21f54 8a8dfe4da6 > karel@silence:~/tmp/fossil-test/git-incremental-import/workspace-fossil$ > fossil leaves --recompute > > and after that I'm able to fossil update and be on the same tree like > in git. However such approach is not too comfortable for running > something like git to fossil mirror on OpenBSD's src repository in > automatic manner since this requires user intervention of finding > changes which need to be reconnected. I can probably write some script > or simple program which would parse fossil timeline output and do > that, but the question is if there is more easier way how to find > "root" revision 2b36d21f54 from the incremental import. If it is, then > script should be fairly easy... > > For the issue duplication I've created simple test case which consist > of two scripts below. Simply put that into empty dir and run > ./test.sh. You can even uncomment 3rd import to see how it looks in > timeline when you do that. > > Thanks! > Karel > > ----->test.sh<-------- > #!/bin/sh > # cleanup and init git repo > set -x > rm -rf workspace > rm -rf workspace-fossil > rm -f workspace.fossil > rm -f *.marks > git init workspace > # first changeset > ./chgen.sh workspace 1 5 > cd workspace > git fast-export --export-marks ../git.marks --all|fossil import --git > ../workspace.fossil > cd ../ > mkdir workspace-fossil > cd workspace-fossil > fossil open ../workspace.fossil > cd .. > # second changeset > ./chgen.sh workspace 6 10 > cd workspace > git fast-export --export-marks ../git.marks --import-marks > ../git.marks --all|fossil import --git ../workspace.fossil -i > --import-marks ../git.marks > cd ../workspace-fossil > fossil timeline > fossil update > fossil timeline > cd .. > # third changeset > # ./chgen.sh workspace 11 15 > # cd workspace > # git fast-export --export-marks ../git.marks --import-marks > ../git.marks --all|fossil import --git ../workspace.fossil -i > --import-marks ../git.marks > # cd ../workspace-fossil > # fossil update > # cd .. > ----->test.sh ends here<----- > > ----->chgen.sh<---- > #!/bin/sh > CWD=`pwd` > cd $1 > for i in `seq $2 $3` > do > echo "File change $i" > file.txt > sleep 1 > git add file.txt > git commit -m "change $i" > done > cd $CWD > > ---->chgen.sh ends here<----
git-incremental-import.tar.bz2
Description: BZip2 compressed data
_______________________________________________ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
