Re: [fossil-users] Is fossil export known to be broken?
Thanks for clarifying that. So I guess it would be either release note advise or schema change notification when running maybe even any command, if it was up to me. ;) Pozdrawiam / With best regards, Orzech W dniu 13.12.2015 o 22:58, Joerg Sonnenberger pisze: > I wouldn't advise to run it after every update, especially for large > repos. On something like NetBSD's src.fossil, it can take hours even on > a fast machine. > > Joerg > ___ > fossil-users mailing list > fossil-users@lists.fossil-scm.org > http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users signature.asc Description: OpenPGP digital signature ___ 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] Is fossil export known to be broken?
On Sun, Dec 13, 2015 at 09:57:44PM +0100, Piotr Orzechowski wrote: > I guess it would be better to change help message about rebuild then. > As of now, it says "Run this command after updating the fossil > executable in a way that changes the database schema." So I would > expect to see some information about such need in release notes. > Perhaps it would be better to add to help that it is a good practice > to run this command after each update. But then, why wouldn't rebuild > happen automatically? I wouldn't advise to run it after every update, especially for large repos. On something like NetBSD's src.fossil, it can take hours even on a fast machine. Joerg ___ 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] Is fossil export known to be broken?
I guess it would be better to change help message about rebuild then. As of now, it says "Run this command after updating the fossil executable in a way that changes the database schema." So I would expect to see some information about such need in release notes. Perhaps it would be better to add to help that it is a good practice to run this command after each update. But then, why wouldn't rebuild happen automatically? Pozdrawiam / With best regards, Orzech W dniu 13.12.2015 o 18:00, jungle Boogie pisze: > On 13 December 2015 at 05:12, Piotr Orzechowski wrote: >> I've just looked at release notes for 1.34 and 1.33 and there was no >> straightforward note about need to rebuild repositories. > > Admittedly I don't do it every time, but I thought that was something > to be done for every update? > > > Here's openBSD's port advising it's a good idea to run rebuild: > http://openports.se/devel/fossil > > After upgrading to a newer version of fossil, it is always a good idea > to run: "fossil all rebuild". Running "rebuild" this way is not always > necessary, but it never hurts. > > > signature.asc Description: OpenPGP digital signature ___ 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] Is fossil export known to be broken?
On 13 December 2015 at 05:12, Piotr Orzechowski wrote: > I've just looked at release notes for 1.34 and 1.33 and there was no > straightforward note about need to rebuild repositories. Admittedly I don't do it every time, but I thought that was something to be done for every update? Here's openBSD's port advising it's a good idea to run rebuild: http://openports.se/devel/fossil After upgrading to a newer version of fossil, it is always a good idea to run: "fossil all rebuild". Running "rebuild" this way is not always necessary, but it never hurts. -- --- inum: 883510009027723 sip: jungleboo...@sip2sip.info xmpp: jungle-boo...@jit.si ___ 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] Is fossil export known to be broken?
On Sun, Dec 13, 2015 at 02:12:01PM +0100, Piotr Orzechowski wrote: > I've just looked at release notes for 1.34 and 1.33 and there was no > straightforward note about need to rebuild repositories. Maybe it would > be better to add such information to requiring release notes, or maybe > even fossil would be able to suggest repo rebuilding itself? > After bisecting, I found that this commit: http://www.fossil-scm.org/index.html/info/8e44cf6f4df4f9f0 Fix the problem. Since this fix is about mlink table and not on the export command directly, updating fossil is not enought to solve the problem, but it require a rebuild. Probably a note could be added to the 1.34 release notes. regards, -- Martin G. ___ 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] Is fossil export known to be broken?
I've just looked at release notes for 1.34 and 1.33 and there was no straightforward note about need to rebuild repositories. Maybe it would be better to add such information to requiring release notes, or maybe even fossil would be able to suggest repo rebuilding itself? Pozdrawiam / With best regards, Orzech W dniu 13.12.2015 o 12:23, org.fossil-scm.fossil-us...@io7m.com pisze: > On 2015-12-12T16:06:20 -0800 > jungle Boogie wrote: >> Great news! >> >> Did you happen to change your git version at all during either test? > Nope, same version all along (2.6.2). > > M > ___ > fossil-users mailing list > fossil-users@lists.fossil-scm.org > http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users signature.asc Description: OpenPGP digital signature ___ 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] Is fossil export known to be broken?
On 2015-12-12T16:06:20 -0800 jungle Boogie wrote: > > Great news! > > Did you happen to change your git version at all during either test? Nope, same version all along (2.6.2). M ___ 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] Is fossil export known to be broken?
On 12 December 2015 at 14:48, wrote: > Ah! > > I've just discovered that if I use the trunk version of fossil and run > it on a freshly created repository, then it does export correctly. If I > use the trunk version of fossil on the repository created by the older > version, it fails. If I 'fossil rebuild' the old repository, exporting > works correctly! Great news! Did you happen to change your git version at all during either test? -- --- inum: 883510009027723 sip: jungleboo...@sip2sip.info xmpp: jungle-boo...@jit.si ___ 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] Is fossil export known to be broken?
On 2015-12-12T17:14:56 -0500 Martin Gagnon wrote: > > For some reason, mine have an extra line at the end that look like this: > > M 100644 :10 README-b0.txt > > For the rest, it's pretty similar. Ah! I've just discovered that if I use the trunk version of fossil and run it on a freshly created repository, then it does export correctly. If I use the trunk version of fossil on the repository created by the older version, it fails. If I 'fossil rebuild' the old repository, exporting works correctly! I have no idea as to the cause, but it seems that fixing it requires using a new version of Fossil AND rebuilding the repository before exporting. I have now correctly exported my original repository and all files appear to be intact! M ___ 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] Is fossil export known to be broken?
On Sat, Dec 12, 2015 at 10:09:26PM +, org.fossil-scm.fossil-us...@io7m.com wrote: > On 2015-12-12T21:43:49 + > wrote: > > > On 2015-12-12T16:21:14 -0500 > > Martin Gagnon wrote: > > > > > > > > The script you've attached is exactly what I used to create the test > > > > case. Which versions are you using? > > > > > > This is fossil version 1.34 [a4889252f1] 2015-12-07 18:19:28 UTC > > > > > > git version 1.9.1 > > Fossil trunk produces an identical fast-import file to the older > version. Could you verify that the fast-import file that your repo case > produces is the same as: > > http://waste.io7m.com/2015/12/12/fossilexport/test.export > > It probably won't be byte-for-byte identical, but I'd assume that the > "mark" numbers would be the same, in addition to the order and number > of commits. For some reason, mine have an extra line at the end that look like this: M 100644 :10 README-b0.txt For the rest, it's pretty similar. Regards, -- Martin G. ___ 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] Is fossil export known to be broken?
On 2015-12-12T21:43:49 + wrote: > On 2015-12-12T16:21:14 -0500 > Martin Gagnon wrote: > > > > > > The script you've attached is exactly what I used to create the test > > > case. Which versions are you using? > > > > This is fossil version 1.34 [a4889252f1] 2015-12-07 18:19:28 UTC > > > > git version 1.9.1 Fossil trunk produces an identical fast-import file to the older version. Could you verify that the fast-import file that your repo case produces is the same as: http://waste.io7m.com/2015/12/12/fossilexport/test.export It probably won't be byte-for-byte identical, but I'd assume that the "mark" numbers would be the same, in addition to the order and number of commits. M ___ 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] Is fossil export known to be broken?
On 2015-12-12T16:21:14 -0500 Martin Gagnon wrote: > > > > The script you've attached is exactly what I used to create the test > > case. Which versions are you using? > > This is fossil version 1.34 [a4889252f1] 2015-12-07 18:19:28 UTC > > git version 1.9.1 Hrm, fossil is quite a bit newer and git is quite a bit older. I'll try the current fossil trunk and see if that eliminates the issue. Is also possible that fossil generates fast-import data that the older git can work with but newer versions won't. M ___ 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] Is fossil export known to be broken?
On Sat, Dec 12, 2015 at 09:18:12PM +, org.fossil-scm.fossil-us...@io7m.com wrote: > On 2015-12-12T16:07:23 -0500 > Martin Gagnon wrote: > > > > What version of git and fossil are you using ? > > > > I cannot reproduce the problem using your repo case. After the merge, > > the README-b0.txt file is present. > > > > Here a script that reproduce your repo case, does it correspond to your > > test case ? > > Hello! > > $ fossil version > This is fossil version 1.32 [715f88811a] 2015-05-02 21:11:26 UTC > > $ git version > git version 2.6.2 > > The script you've attached is exactly what I used to create the test > case. Which versions are you using? This is fossil version 1.34 [a4889252f1] 2015-12-07 18:19:28 UTC git version 1.9.1 -- Martin G. ___ 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] Is fossil export known to be broken?
On 2015-12-12T23:24:08 +0300 Konstantin Khomoutov wrote: > > Did you report a bug? > > It's just a matter of posting a mail to git at vger.kernel.org; > subscription is not required. I sent the git list the same info and test case a couple of hours ago, but haven't had a response yet. I still don't know if this is a Fossil or a Git issue, so I thought I'd let the git people look at the fast-import file to see if they can spot any obvious mistakes in it. M ___ 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] Is fossil export known to be broken?
On 2015-12-12T16:07:23 -0500 Martin Gagnon wrote: > > What version of git and fossil are you using ? > > I cannot reproduce the problem using your repo case. After the merge, > the README-b0.txt file is present. > > Here a script that reproduce your repo case, does it correspond to your > test case ? Hello! $ fossil version This is fossil version 1.32 [715f88811a] 2015-05-02 21:11:26 UTC $ git version git version 2.6.2 The script you've attached is exactly what I used to create the test case. Which versions are you using? M ___ 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] Is fossil export known to be broken?
On Sat, Dec 12, 2015 at 02:47:59PM +, org.fossil-scm.fossil-us...@io7m.com wrote: > On 2015-12-12T14:21:19 + > wrote: > > > > Ideally, I'd be able to reproduce this with a somewhat smaller > > repository... > > Surprisingly, this turned out to be easier than expected! > > http://waste.io7m.com/2015/12/12/fossilexport/ > > 1. Create fossil repository. > 2. Add README.txt and commit in trunk. > 3. Create branch 'b0' and switch to it. > 4. Add README-b0.txt and commit in b0. > 5. Switch to trunk. > 6. Merge and commit 'b0'. > [snip] > > Importing that results in no README-b0.txt existing in the 'trunk' > branch of the created git repository. > > I don't understand what's going wrong yet, but at least we now have > a repro case. > What version of git and fossil are you using ? I cannot reproduce the problem using your repo case. After the merge, the README-b0.txt file is present. Here a script that reproduce your repo case, does it correspond to your test case ? regards, -- Martin G. ff.sh Description: Bourne shell script ___ 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] Is fossil export known to be broken?
On Sat, 12 Dec 2015 08:42:48 -0500 Richard Hipp wrote: > > It seems that somebody else ran into this at the start of the year: > > http://www.mail-archive.com/fossil-users@lists.fossil-scm.org/msg19238.html > Yeah, that's a bummer. > > Part of the problem stems from the fact that the Git fast-export > format is only sparsely documented, and in some cases incorrectly > documented. And Git is very fussy about the format. So it is > possible to generate a fast-export file that conforms exactly to the > documentation but which Git will not accept. And since Git is the > dominant VCS at the moment, the Git people are under no pressure to > fix this sad state of affairs. Did you report a bug? It's just a matter of posting a mail to git at vger.kernel.org; subscription is not required. ___ 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] Is fossil export known to be broken?
Tnx for the repro case! I'm away from office most of today. Will look when I'm able. On 12/12/15, org.fossil-scm.fossil-us...@io7m.com wrote: > On 2015-12-12T14:21:19 + > wrote: >> >> Ideally, I'd be able to reproduce this with a somewhat smaller >> repository... > > Surprisingly, this turned out to be easier than expected! > > http://waste.io7m.com/2015/12/12/fossilexport/ > > 1. Create fossil repository. > 2. Add README.txt and commit in trunk. > 3. Create branch 'b0' and switch to it. > 4. Add README-b0.txt and commit in b0. > 5. Switch to trunk. > 6. Merge and commit 'b0'. > > The test.fossil repository is the repository resulting from the above: > > $ fossil timeline -R test.fossil > === 2015-12-12 === > 14:26:36 [6ee55fff27] *MERGE* Merge b0 (user: someone tags: trunk) > 14:26:14 [5c6acf101e] Add README-b0 (user: someone tags: b0) > 14:25:38 [76a31a2f66] Create new branch named "b0" (user: someone tags: b0) > > 14:25:17 [79f2f13cf7] *BRANCH* Initial (user: someone tags: trunk) > 14:24:46 [fa9a67d8e8] initial empty check-in (user: someone tags: trunk) > +++ no more data (5) +++ > > Exporting the repository results in this: > > --- > blob > mark :4 > data 7 > Hello. > > blob > mark :10 > data 10 > Hello b0. > > commit refs/heads/trunk > mark :3 > committer someone 1449930286 + > data 22 > initial empty check-in > deleteall > > commit refs/heads/trunk > mark :7 > committer someone 1449930317 + > data 7 > Initial > from :3 > M 100644 :4 README.txt > > commit refs/heads/b0 > mark :9 > committer someone 1449930338 + > data 28 > Create new branch named "b0" > from :7 > > commit refs/heads/b0 > mark :13 > committer someone 1449930374 + > data 13 > Add README-b0 > from :9 > M 100644 :10 README-b0.txt > > commit refs/heads/trunk > mark :15 > committer someone 1449930396 + > data 8 > Merge b0 > from :7 > merge :13 > --- > > Importing that results in no README-b0.txt existing in the 'trunk' > branch of the created git repository. > > I don't understand what's going wrong yet, but at least we now have > a repro case. > > M > ___ > 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...@sqlite.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] Is fossil export known to be broken?
On 2015-12-12T14:21:19 + wrote: > > Ideally, I'd be able to reproduce this with a somewhat smaller > repository... Surprisingly, this turned out to be easier than expected! http://waste.io7m.com/2015/12/12/fossilexport/ 1. Create fossil repository. 2. Add README.txt and commit in trunk. 3. Create branch 'b0' and switch to it. 4. Add README-b0.txt and commit in b0. 5. Switch to trunk. 6. Merge and commit 'b0'. The test.fossil repository is the repository resulting from the above: $ fossil timeline -R test.fossil === 2015-12-12 === 14:26:36 [6ee55fff27] *MERGE* Merge b0 (user: someone tags: trunk) 14:26:14 [5c6acf101e] Add README-b0 (user: someone tags: b0) 14:25:38 [76a31a2f66] Create new branch named "b0" (user: someone tags: b0) 14:25:17 [79f2f13cf7] *BRANCH* Initial (user: someone tags: trunk) 14:24:46 [fa9a67d8e8] initial empty check-in (user: someone tags: trunk) +++ no more data (5) +++ Exporting the repository results in this: --- blob mark :4 data 7 Hello. blob mark :10 data 10 Hello b0. commit refs/heads/trunk mark :3 committer someone 1449930286 + data 22 initial empty check-in deleteall commit refs/heads/trunk mark :7 committer someone 1449930317 + data 7 Initial from :3 M 100644 :4 README.txt commit refs/heads/b0 mark :9 committer someone 1449930338 + data 28 Create new branch named "b0" from :7 commit refs/heads/b0 mark :13 committer someone 1449930374 + data 13 Add README-b0 from :9 M 100644 :10 README-b0.txt commit refs/heads/trunk mark :15 committer someone 1449930396 + data 8 Merge b0 from :7 merge :13 --- Importing that results in no README-b0.txt existing in the 'trunk' branch of the created git repository. I don't understand what's going wrong yet, but at least we now have a repro case. M ___ 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] Is fossil export known to be broken?
On 2015-12-12T08:42:48 -0500 Richard Hipp wrote: > > The import/export functionality of Fossil has been greatly improved in > the past by the work of volunteers tracking down obscure > incompatibilities. If you would like to try to get to the bottom of > the problems you are seeing, and either identify the root cause, or > better to suggest patches, that would be greatly appreciated. Not knowing the fast-import format, I suspect I'm going to be approaching the git people and asking them why git doesn't seem to import the produced file properly. I've checked the exported data and all of the files and commits are present, but it does seem like something isn't carried across merges. Ideally, I'd be able to reproduce this with a somewhat smaller repository... M ___ 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] Is fossil export known to be broken?
On 12/12/15, org.fossil-scm.fossil-us...@io7m.com wrote: > On 2015-12-12T12:26:19 + > wrote: >> >> Is there some way to get more information about what's going wrong here? I >> don't >> know if fossil or git is at fault, and I have no way of knowing how badly >> the >> history has been corrupted by the export or import. > > It seems that somebody else ran into this at the start of the year: > > > http://www.mail-archive.com/fossil-users@lists.fossil-scm.org/msg19238.html > Yeah, that's a bummer. Part of the problem stems from the fact that the Git fast-export format is only sparsely documented, and in some cases incorrectly documented. And Git is very fussy about the format. So it is possible to generate a fast-export file that conforms exactly to the documentation but which Git will not accept. And since Git is the dominant VCS at the moment, the Git people are under no pressure to fix this sad state of affairs. The import/export functionality of Fossil has been greatly improved in the past by the work of volunteers tracking down obscure incompatibilities. If you would like to try to get to the bottom of the problems you are seeing, and either identify the root cause, or better to suggest patches, that would be greatly appreciated. -- 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
Re: [fossil-users] Is fossil export known to be broken?
On 2015-12-12T12:26:19 + wrote: > > Is there some way to get more information about what's going wrong here? I > don't > know if fossil or git is at fault, and I have no way of knowing how badly the > history has been corrupted by the export or import. It seems that somebody else ran into this at the start of the year: http://www.mail-archive.com/fossil-users@lists.fossil-scm.org/msg19238.html M ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
[fossil-users] Is fossil export known to be broken?
Hello! I'm attempting to export a moderately large (241mb) repository to git: http://fossil.io7m.com/repo.cgi/io7m-r1/index The export proceeds without an error: /tmp/r1-fossil$ fossil export > /tmp/r1.bin The import proceeds without an error: /tmp/r1-git$ git init /tmp/r1-git$ git fast-import < /tmp/r1.bin However, the resulting git checkout is missing files that are present in the most recent commit to the fossil repository. /tmp/r1-fossil$ ls -alF io7m-r1-documentation/src/main/resources/com/io7m/r1/documentation/ | wc -l 43 /tmp/r1-git$ ls -alF io7m-r1-documentation/src/main/resources/com/io7m/r1/documentation/ | wc -l 30 In some cases, entire directory hierarchies are missing: /tmp/r1-fossil$ ls -alF io7m-r1-examples/src/main/resources/com/io7m/r1/examples/results/scenes/ | wc -l 80 /tmp/r1-git$ ls -alF io7m-r1-examples/src/main/resources/com/io7m/r1/examples/results/scenes/ | wc -l ls: cannot access io7m-r1-examples/src/main/resources/com/io7m/r1/examples/results/scenes/: No such file or directory 0 Is there some way to get more information about what's going wrong here? I don't know if fossil or git is at fault, and I have no way of knowing how badly the history has been corrupted by the export or import. M ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users