Re: [fossil-users] meld-support
On Tue, Mar 8, 2011 at 12:44 PM, Jan Danielsson jan.m.daniels...@gmail.comwrote: Hello, Anyone here using meld[1]? When merging with many conflicts I find meld to be of invaluable help, though I noticed that meld is missing explicit fossil-support. Is meld more than just a visual 3-way merge tool, like xxdiff or kdiff3? Can meld be plugged into Fossil using the gmerge-command setting? To make a pretty short story even shorter, I'm new to fossil and I never opened a meld vc-plugin in an editor before today. But I put together the attached file (heavily based on the monotone-plugin (don't mind the left-overs)). I tried it with meld 1.5.0, and it appears to be working[2] (though my tests are extremely limited). If anyone has any suggestions for improvements, let me know. I'm going to clean it up (specifically the _get_dirsandfiles() method), and then send it to the meld-folks, but if there's any feedback to be heard from fossil users first, I'd like to hear them. Specifically I'd like to know: - What states (which could be relevant to meld) are missing from the state_map? - Any obvious missing *_command()'s? (I don't know how one would translate resolved_command() to fossil, so I simply disabled it). On a slight tangent; my reading and testing of meld-plugins did yield a question concerning the states in fossil: I noticed that some VC's have support for a added yet missing state. I tried adding a file to a fossil repository, then removed the file from the file system (using regular rm), and fossil ls -l showed the file as ADDED rather than MISSING. Is this by design, or is it an oversight? (I realize it's a situation which should happen, but let's say someone does something bizarre..). The fossil-plugin for meld could rescan all added files to see if they actually exist, and set them to missing if they have disappeared. But I can't shake the feeling that fossil ls -l should be picking it up. Thoughts? [1] http://meld.sourceforge.net/ [2] Just drop fossil.py into meld's vc/ directory -- Kind regards, Jan Danielsson ___ 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] meld-support
On Tue, Mar 8, 2011 at 1:12 PM, Richard Hipp d...@sqlite.org wrote: On Tue, Mar 8, 2011 at 12:44 PM, Jan Danielsson jan.m.daniels...@gmail.com wrote: Hello, Anyone here using meld[1]? When merging with many conflicts I find meld to be of invaluable help, though I noticed that meld is missing explicit fossil-support. Is meld more than just a visual 3-way merge tool, like xxdiff or kdiff3? Can meld be plugged into Fossil using the gmerge-command setting? While we're on the subject, should the gmerge-command setting be extended to be smart about commonly used graphical merges. Should you be able to say just: fossil setting gmerge-command kdiff3 and fossil would know to add the %baseline %original %merge -o %output arguments onto the end automatically? Similar automatic arguments would apply to xxdiff and meld, perhaps? Or is that a case of the software trying to be too clever? To make a pretty short story even shorter, I'm new to fossil and I never opened a meld vc-plugin in an editor before today. But I put together the attached file (heavily based on the monotone-plugin (don't mind the left-overs)). I tried it with meld 1.5.0, and it appears to be working[2] (though my tests are extremely limited). If anyone has any suggestions for improvements, let me know. I'm going to clean it up (specifically the _get_dirsandfiles() method), and then send it to the meld-folks, but if there's any feedback to be heard from fossil users first, I'd like to hear them. Specifically I'd like to know: - What states (which could be relevant to meld) are missing from the state_map? - Any obvious missing *_command()'s? (I don't know how one would translate resolved_command() to fossil, so I simply disabled it). On a slight tangent; my reading and testing of meld-plugins did yield a question concerning the states in fossil: I noticed that some VC's have support for a added yet missing state. I tried adding a file to a fossil repository, then removed the file from the file system (using regular rm), and fossil ls -l showed the file as ADDED rather than MISSING. Is this by design, or is it an oversight? (I realize it's a situation which should happen, but let's say someone does something bizarre..). The fossil-plugin for meld could rescan all added files to see if they actually exist, and set them to missing if they have disappeared. But I can't shake the feeling that fossil ls -l should be picking it up. Thoughts? [1] http://meld.sourceforge.net/ [2] Just drop fossil.py into meld's vc/ directory -- Kind regards, Jan Danielsson ___ 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 -- 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] meld-support
On Tue, Mar 08, 2011 at 01:12:22PM -0500, Richard Hipp wrote: On Tue, Mar 8, 2011 at 12:44 PM, Jan Danielsson jan.m.daniels...@gmail.comwrote: Hello, Anyone here using meld[1]? When merging with many conflicts I find meld to be of invaluable help, though I noticed that meld is missing explicit fossil-support. Is meld more than just a visual 3-way merge tool, like xxdiff or kdiff3? I've never used it for 3-way merge; for me it's like an editor with two files asside and its diff graphically. Something like a more graphical vimdiff, with mouse control and all that. I use it as it has a comfortable directory diff, quick merging by mouse, and an acceptable editor for any new text I may have to type. In fact, I use to review branches using meld, but setting one directory with a fossil checkout of the branch leaf, and then another directory with the pivot to the branch I want to merge to. Then I run meld between the two directories. ___ 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] meld-support
On Tue, Mar 08, 2011 at 01:15:48PM -0500, Richard Hipp wrote: Is meld more than just a visual 3-way merge tool, like xxdiff or kdiff3? Can meld be plugged into Fossil using the gmerge-command setting? While we're on the subject, should the gmerge-command setting be extended to be smart about commonly used graphical merges. Should you be able to say just: fossil setting gmerge-command kdiff3 and fossil would know to add the %baseline %original %merge -o %output arguments onto the end automatically? Similar automatic arguments would apply to xxdiff and meld, perhaps? Or is that a case of the software trying to be too clever? I think that's too clever. I'd keep fossil simpler. Nevertheless, I recall the 'unison' configuration file, and how it deals with it, as that may be an interesting reference: http://www.cis.upenn.edu/~bcpierce/unison/download/releases/stable/unison-manual.html#merge Regards, Lluís. ___ 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] meld-support
On 03/08/11 19:12, Richard Hipp wrote: [---] When merging with many conflicts I find meld to be of invaluable help, though I noticed that meld is missing explicit fossil-support. Is meld more than just a visual 3-way merge tool, like xxdiff or kdiff3? It's mostly a diff/merge-tool (file and directories). But in addition it has some basic interfaces to VCS's. The GUI allows one to revert files, commit files, etc. I don't use those features very frequently, but sometimes it's convenient to have them there when I'm anyway sorting through diffs in meld. Can meld be plugged into Fossil using the gmerge-command setting? Yes, I just tried it, and it works fine. Could someone add this to the list of examples for gmerge-command at http://www.fossil-scm.org/fossil/help/setting: meld %baseline %original %merge %output .. ? Caveats: - The meld --help stipulates that there's an -o or --output to specify output file, but it it appears to be a nop. (Though perhaps I'm misunderstanding how it's supposed to work). Simply passing %output as a fourth argument seems to do the trick. (I used meld HEAD (cloned from their git yesterday) for these tests). - Some Random Web Page stipulates that meld version 1.4 or later is required for the output argument. -- Kind regards, Jan Danielsson 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] meld-support
On 03/08/11 19:15, Richard Hipp wrote: [---] While we're on the subject, should the gmerge-command setting be extended to be smart about commonly used graphical merges. Should you be able to say just: fossil setting gmerge-command kdiff3 and fossil would know to add the %baseline %original %merge -o %output arguments onto the end automatically? Similar automatic arguments would apply to xxdiff and meld, perhaps? Or is that a case of the software trying to be too clever? I think it's being a little bit too clever. If I'm correct in my assumption that there's a problem with the current -o,--output option(s) in meld, and they fix it, then fossil would need to figure out if it's a fixed or unfixed meld being used. It's better to document it with examples (including possible caveats), and make sure the documentation is easy to find. -- Kind regards, Jan Danielsson 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] meld-support
I love meld, but note that fedora 13 14 are currently at v 1.3 Not a problem, but they don't appear to recognize the 4th argument. //Bill On 03/08/2011 02:26 PM, Jan Danielsson wrote: On 03/08/11 19:12, Richard Hipp wrote: [---] When merging with many conflicts I find meld to be of invaluable help, though I noticed that meld is missing explicit fossil-support. Is meld more than just a visual 3-way merge tool, like xxdiff or kdiff3? It's mostly a diff/merge-tool (file and directories). But in addition it has some basic interfaces to VCS's. The GUI allows one to revert files, commit files, etc. I don't use those features very frequently, but sometimes it's convenient to have them there when I'm anyway sorting through diffs in meld. Can meld be plugged into Fossil using the gmerge-command setting? Yes, I just tried it, and it works fine. Could someone add this to the list of examples for gmerge-command at http://www.fossil-scm.org/fossil/help/setting: meld %baseline %original %merge %output .. ? Caveats: - The meld --help stipulates that there's an -o or --output to specify output file, but it it appears to be a nop. (Though perhaps I'm misunderstanding how it's supposed to work). Simply passing %output as a fourth argument seems to do the trick. (I used meld HEAD (cloned from their git yesterday) for these tests). - Some Random Web Page stipulates that meld version 1.4 or later is required for the output argument. ___ 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] meld-support
On 03/08/11 20:45, Bill Whiting wrote: I love meld, but note that fedora 13 14 are currently at v 1.3 Not a problem, but they don't appear to recognize the 4th argument. Just for the record, in case someone finds this thread through Google or something, as long as you have already the appropriate dependencies installed (pygtk, etc), meld is easy to checkout and use (without requiring any additional installation steps). Just do the checkout and then point the gmerge-command to the newly fetched meld/bin/meld using an absolute path. -- Kind regards, Jan Danielsson 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