Re: [fossil-users] Merge two repositories... not.
Maybe consider moving everything on repo B to a separate branch, so that A is on trunk, B is on a branch. Then you can issue a merge command on the home computer, I imagine this would be easier that doing push / merge at the same time. Just a thought, but I haven't actually merged two fossil repos myself. Tomek On Wed, Aug 24, 2011 at 4:55 PM, Martin S. Weber martin.we...@nist.govwrote: So, I'm aware of drh's message from 2010-03-06 [1]. Is merging repositories this way still supposed to work? If so, I'll open a ticket with the information below. -Martin [1] http://www.mail-archive.com/fossil-users@lists.fossil-scm.org/msg01641.html # background: two teams working on some stuff, the fossil-using # team working on something that ended up being a subset of the # cvs-using team. Joerg's version of cvs2fossil made a fossil out # of the unitsdb team (see below). # The goal was to merge the resulting fossil with the subset that # the prefixes_derived team was working on, resulting in a single # fossil that contains both... $ mkdir merge $ cd merge fossil open ../unitsdb.fossil # ... snip, output deleted $ fossil info project-name: unnamed repository: /home/www/tmp/m/unitsdb.fossil local-root: /home/www/tmp/m/merge/ project-code: eeb7e06236b08dc4b57b6ab3b957fe5756c64f5b server-code: dc2df780602e6d8561d1a7280acbe4aabfe15ad3 checkout: df5c154920608f11c26a4a98f3d179b792b1ef94 2011-08-24 20:20:09 UTC parent: 443c51cf173da58579e5457c7a634e3e51d85773 2011-08-18 15:58:06 UTC tags: trunk comment: (Martin) Checkin before I munge everything.. (user: wwwpl) $ cd .. $ fossil sqlite3 -R prefixes_derived.fossil SQLite version 3.7.7 2011-06-24 11:29:51 Enter .help for instructions Enter SQL statements terminated with a ; sqlite update config set value='eeb7e06236b08dc4b57b6ab3b957fe5756c64f5b' where name='project-code'; sqlite .quit $ cd merge $ fossil ui -P 12345 $ cd .. $ mkdir pd $ cd pd $ fossil open ../prefixes_derived.fossil # snip and then project-name: Prefixed and Derived units repository: /home/www/tmp/m/prefixes_derived.fossil local-root: /home/www/tmp/m/pd/ project-code: eeb7e06236b08dc4b57b6ab3b957fe5756c64f5b server-code: 67266cf692b50dd078ffee601caf7f345fa3d243 checkout: cccd2619d9fb3724f96a806f55f443837dbc9bbe 2011-08-16 23:33:50 UTC parent: f3876370ba605de6706717018ee0f41dde9d4df0 2011-08-11 21:13:35 UTC child:53c88f11df96c0156e9cbf3e0ba28905da261500 2011-08-16 23:38:07 UTC tags: trunk # project codes agree$ pwd /home/www/tmp/m/pd $ fossil push http://localhost:12345 Bytes Cards Artifacts Deltas Sent:7858169 0 0 Received:7928169 0 0 Sent: 295775322 14 1 Received:7270155 0 0 Sent: 274809307 16 0 Received:6518139 0 0 Sent: 273684292 15 0 Received:5813124 0 0 Sent: 286499289 3 0 Received:5672121 0 0 Sent: 338281281 8 0 Received:5296113 0 0 Sent: 325753276 5 0 Received:5061108 0 0 Sent: 294428269 6 1 Received:4732101 0 0 Sent: 267993199 25 45 Received:7129152 0 0 Sent: 251184169 14137 waiting for server... Now. The server sits here eating 100% CPU and increasing its memory usage -- and not doing anything else. right now this is how it looks from top(1): (pid user RR NI VIRT RES SHR %CPU %MEM TIME) 12375 wwwpl 33 8 2567m 1.3g 972 R 99.8 34.2 7:00.64 fossil This is fossil version 1.18 [df9da91ba8] 2011-07-13 23:03:41 UTC ___ 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] Merge two repositories... not.
May be fossil export of one piped on a fossil import of the other one will work!! I never tried but it might work I don't know how would look like the timeline when doing that... -- Martin Le 2011-08-24 à 16:55, Martin S. Weber martin.we...@nist.gov a écrit : So, I'm aware of drh's message from 2010-03-06 [1]. Is merging repositories this way still supposed to work? If so, I'll open a ticket with the information below. -Martin [1] http://www.mail-archive.com/fossil-users@lists.fossil-scm.org/msg01641.html # background: two teams working on some stuff, the fossil-using # team working on something that ended up being a subset of the # cvs-using team. Joerg's version of cvs2fossil made a fossil out # of the unitsdb team (see below). # The goal was to merge the resulting fossil with the subset that # the prefixes_derived team was working on, resulting in a single # fossil that contains both... $ mkdir merge $ cd merge fossil open ../unitsdb.fossil # ... snip, output deleted $ fossil info project-name: unnamed repository: /home/www/tmp/m/unitsdb.fossil local-root: /home/www/tmp/m/merge/ project-code: eeb7e06236b08dc4b57b6ab3b957fe5756c64f5b server-code: dc2df780602e6d8561d1a7280acbe4aabfe15ad3 checkout: df5c154920608f11c26a4a98f3d179b792b1ef94 2011-08-24 20:20:09 UTC parent: 443c51cf173da58579e5457c7a634e3e51d85773 2011-08-18 15:58:06 UTC tags: trunk comment: (Martin) Checkin before I munge everything.. (user: wwwpl) $ cd .. $ fossil sqlite3 -R prefixes_derived.fossil SQLite version 3.7.7 2011-06-24 11:29:51 Enter .help for instructions Enter SQL statements terminated with a ; sqlite update config set value='eeb7e06236b08dc4b57b6ab3b957fe5756c64f5b' where name='project-code'; sqlite .quit $ cd merge $ fossil ui -P 12345 $ cd .. $ mkdir pd $ cd pd $ fossil open ../prefixes_derived.fossil # snip and then project-name: Prefixed and Derived units repository: /home/www/tmp/m/prefixes_derived.fossil local-root: /home/www/tmp/m/pd/ project-code: eeb7e06236b08dc4b57b6ab3b957fe5756c64f5b server-code: 67266cf692b50dd078ffee601caf7f345fa3d243 checkout: cccd2619d9fb3724f96a806f55f443837dbc9bbe 2011-08-16 23:33:50 UTC parent: f3876370ba605de6706717018ee0f41dde9d4df0 2011-08-11 21:13:35 UTC child:53c88f11df96c0156e9cbf3e0ba28905da261500 2011-08-16 23:38:07 UTC tags: trunk # project codes agree$ pwd /home/www/tmp/m/pd $ fossil push http://localhost:12345 Bytes Cards Artifacts Deltas Sent:7858169 0 0 Received:7928169 0 0 Sent: 295775322 14 1 Received:7270155 0 0 Sent: 274809307 16 0 Received:6518139 0 0 Sent: 273684292 15 0 Received:5813124 0 0 Sent: 286499289 3 0 Received:5672121 0 0 Sent: 338281281 8 0 Received:5296113 0 0 Sent: 325753276 5 0 Received:5061108 0 0 Sent: 294428269 6 1 Received:4732101 0 0 Sent: 267993199 25 45 Received:7129152 0 0 Sent: 251184169 14137 waiting for server... Now. The server sits here eating 100% CPU and increasing its memory usage -- and not doing anything else. right now this is how it looks from top(1): (pid user RR NI VIRT RES SHR %CPU %MEM TIME) 12375 wwwpl 33 8 2567m 1.3g 972 R 99.8 34.2 7:00.64 fossil This is fossil version 1.18 [df9da91ba8] 2011-07-13 23:03:41 UTC ___ 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] Merge two repositories... not.
On 08/24/11 18:14, Martin Gagnon wrote: May be fossil export of one piped on a fossil import of the other one will work!! I never tried but it might work I don't know how would look like the timeline when doing that... Actually I did try fossil deconstruct'ing both repositories into the same directory with reconstructing it afterwards. The progress reaching 102.1% made me think it picked up the rest, but it didn't. Note though that I did not adjust the project code before I did the deconstruct/reconstruct. Also I assume that a clash of hashes is unlikely, but not impossible this way (but, then again, I assume the same problem (or let's call it probability) exists for pushing content like drh suggested 1.5 years ago). -Martin ___ 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] Merge two repositories
On Mar 6, 2010, at 6:29 AM, Twylite wrote: Hi, Is it possible to merge two repositories, or to import one repository into another? Suppose your two projects are A and B. You want to import B into A. First find the project-code for A using the fossil info command. Then change the project code of B to match the project code of A using the sqlite3 command-line client: sqlite3 B.fossil UPDATE config SET value='project-code-for-A' WHERE name='project-code'; Then you will be able to push all the content of B into A. I would like to retain the identity of one repository (and if possible its artifact IDs), but I don't really care about the artifact IDs of the repository being imported -- mainly I want to bring all history of two different projects into one repository. I seem to have managed something along these lines by deconstructing multiple repos into one folder, and then reconstructing a new repository from that folder. One problem I have encountered is that the new repository is just that: new. It has a different project-id and server-id. This seems to be a feature of reconstruct even when a single repo is involved. The effect is that I get C = A + B, rather than A += B, which means that existing clients of my large repository A will need to clone a new repository (in this case C) if when I do such merges, rather than just pulling down updates. Any suggestions? Regards, Twylite ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users D. Richard Hipp d...@hwaci.com ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users