Re: [fossil-users] Right way to try something new and save/revert?
On Mon, Jan 28, 2013 at 7:20 AM, Richard Hipp d...@sqlite.org wrote: On Sun, Jan 27, 2013 at 9:18 AM, Gilles gilles.gana...@free.fr wrote: So it looks like fossil ls branch name doesn't actually list the files in the branch but rather the files in the branch currently active (and its files checked out in the work directory). The command is actually: fossil ls $VERSION If you specify a branch-name for $VERSION, then it takes the most recent check-in on that branch. See http://www.fossil-scm.org/fossil/doc/trunk/www/checkin_names.wiki for additional information on $VERSION names. The output is all files under management for the particular $VERSION that you specify. That's what I tough, but it seems that specifying VERSION doesn't work as expected, it have no effect and always list files on the branch corresponding to the current checkout. Here a simple test that demonstrate it on a new repository created from scratch... - $ fossil init testrepo.fossil project-id: d0b945d10191d8df07ddc6a7b0e860f72819f41d server-id: 63e9f2921e6eff72fd7423aabec7a424de2b82a9 ... $ fossil open testrepo.fossil $ echo bla file1.txt $ fossil add file1.txt ADDED file1.txt $ fossil ci -m added file1.txt New_Version: 3973f4d3d4b5d3a2bdd59fe27f3fcab22efece3c $ echo bla bla file2.txt $ fossil add file2.txt ADDED file2.txt $ fossil ci -m add file2 on branch-A -b branch-A New_Version: e9a764634e288fb490a717e3b302443ee3556f0c $ fossil ls file1.txt file2.txt $ fossil ls trunk file1.txt file2.txt $ fossil up trunk REMOVE file2.txt --- updated-to: 3973f4d3d4b5d3a2bdd59fe27f3fcab22efece3c 2013-01-28 13:20:56 UTC tags: trunk comment: added file1.txt (user: mgagnon) changes: 1 file modified. fossil undo is available to undo changes to the working checkout. $ fossil ls file1.txt $ fossil ls branch-A file1.txt - 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] Right way to try something new and save/revert?
On Fri, 25 Jan 2013 13:19:09 +0100, Stephan Beal sgb...@googlemail.com wrote: FWIW: it appears to me that it does not react intuitively when given a branch name as a version: [stephan@host:~/cvs/fossil/fossil]$ fossil ls encoding-glob --age | sort -r | head 2013-01-25 08:53:54 win/Makefile.msc 2013-01-25 08:53:54 src/makemake.tcl 2013-01-24 15:19:08 win/Makefile.mingw 2013-01-24 15:19:08 src/main.mk 2013-01-24 14:38:19 src/stat.c those changes were all made in the trunk branch, not the encoding-glob branch, and have not been merged into that branch since then. So it looks like fossil ls branch name doesn't actually list the files in the branch but rather the files in the branch currently active (and its files checked out in the work directory). ___ 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] Right way to try something new and save/revert?
On Fri, 25 Jan 2013 18:33:20 +0100, Stephan Beal sgb...@googlemail.com wrote: The json command doesn't exist in the Fossil I have (1.24 [0c65916136]): That's an optional feature and getting it to build on Windows in 1.24 apparently requires modifying the makefile (that was fixed yesterday while waiting on a large database import). If you have the latest sources and nmake is in your %PATH, you can try (from fossil's top-most source directory): nmake -f win/Makefile.mfc FOSSIL_ENABLE_JSON=1 Thanks. I'll try compiling the latest source on my Windows host. What about switching to another branch (eg. going from trunk to experimental), and simply running fossil ls? Will that list the revisions without touching the files in my work directory? No - switching branches checks out the files from that branch, so it changes the local files. Good to know. Thank you. ___ 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] Right way to try something new and save/revert?
On Fri, 25 Jan 2013 10:43:03 -0800, Themba Fletcher themba.fletc...@gmail.com wrote: I have another question: fossil branch ls lists branches available in the repo, but is there a command to list all the files/revisions that have been commited to the experimental branch? I think 'fossil diff --brief --branch experimental' might be what you're looking for. Thanks for the command. However, after editing dummy.c twice and commiting the changes, it just displays this: CHANGED dummy.c The ultimate goal is really to be able to locate code I know I tried at some point in the Experimental branch. Grep would really help. Then a branch could be used as some kind of code vault. ___ 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] Right way to try something new and save/revert?
On Wed, 09 Jan 2013 12:10:35 +0100, Gilles gilles.gana...@free.fr wrote: Am I correct in understanding that this is the right way to proceed to try some new code, and either save it (whether it works or not, just as a track-record) or discard it? I have another question: fossil branch ls lists branches available in the repo, but is there a command to list all the files/revisions that have been commited to the experimental branch? Thank you. ___ 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] Right way to try something new and save/revert?
On Fri, Jan 25, 2013 at 11:59 AM, Gilles gilles.gana...@free.fr wrote: I have another question: fossil branch ls lists branches available in the repo, but is there a command to list all the files/revisions that have been commited to the experimental branch? There are two ways i know of to do that, but both show the files _in_ the branch, whether or not they have been modified _in that branch_ since the branch was created: 1) The web interface (/dir) allows one to browse a specific branch. http://fossil-scm.org/index.html/dir?ci=encoding-glob 2) The JSON dir command with the checkin option: [stephan@host:~/cvs/fossil/fossil]$ f json dir -checkin encoding-glob { fossil:bb63588e1bd65f1556bcf778a49833eda55c4e14, timestamp:1359113099, command:dir, procTimeMs:36, payload:{ name:/, checkin:f480173856136f6ad30686edc875de969a6f1a91, entries:[ ... { name:COPYRIGHT-BSD2.txt, uuid:c7dd1de9f9539a5a859c2b41fe4560604a774476, size:1528, timestamp:1274036925, downloadPath:/raw/COPYRIGHT-BSD2.txt?name=c7dd1de9f9539a5a859c2b41fe4560604a774476 }, ... } but note that the output format from /json/dir is a bit dumb (IMO). i would be very glad to get input about how to improve it. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ 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] Right way to try something new and save/revert?
On Fri, 25 Jan 2013 12:28:25 +0100, Stephan Beal sgb...@googlemail.com wrote: There are two ways i know of to do that, but both show the files _in_ the branch, whether or not they have been modified _in that branch_ since the branch was created: Thanks for the info. I don't understand what you mean by both show the files _in_ the branch, whether or not they have been modified _in that branch_ since the branch was created. Do you meant that they'll list all files in a branch, even those that were only commited once since the branch was created? I prefer to use the CLI because it's faster than the web interface. Hopefully there's an easier way than running json dir and reading its output. What about using tags to list files, eg. trunk vs. experimental? ___ 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] Right way to try something new and save/revert?
On Fri, Jan 25, 2013 at 12:41 PM, Gilles gilles.gana...@free.fr wrote: I don't understand what you mean by both show the files _in_ the branch, whether or not they have been modified _in that branch_ since the branch was created. Do you meant that they'll list all files in a branch, even those that were only commited once since the branch was created? By that i meant: your question was a way to see what has been committed to that branch. The /dir and /json/dir commands show what is in the branch, regardless of whether it was committed in that branch or imported via the parent branch when the branch was created. I prefer to use the CLI because it's faster than the web interface. Hopefully there's an easier way than running json dir and reading its output. fossil co branchname ls -la There is no equivalent of the /dir page for the CLI. What about using tags to list files, eg. trunk vs. experimental? That works with the /dir page and /json/dir command. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ 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] Right way to try something new and save/revert?
On Fri, Jan 25, 2013 at 1:14 PM, Stephan Beal sgb...@googlemail.com wrote: There is no equivalent of the /dir page for the CLI. It appears that i lied, at least partially: Martin shows us the 'ls' command (which is new to me!). -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ 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] Right way to try something new and save/revert?
On Fri, Jan 25, 2013 at 1:11 PM, Martin Gagnon eme...@gmail.com wrote: Is this a bug or I mis understand meaning of ?VERSION? for the ls command ? FWIW: it appears to me that it does not react intuitively when given a branch name as a version: [stephan@host:~/cvs/fossil/fossil]$ fossil ls encoding-glob --age | sort -r | head 2013-01-25 08:53:54 win/Makefile.msc 2013-01-25 08:53:54 src/makemake.tcl 2013-01-24 15:19:08 win/Makefile.mingw 2013-01-24 15:19:08 src/main.mk 2013-01-24 14:38:19 src/stat.c those changes were all made in the trunk branch, not the encoding-glob branch, and have not been merged into that branch since then. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ 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] Right way to try something new and save/revert?
On Fri, 25 Jan 2013 13:16:46 +0100, Stephan Beal sgb...@googlemail.com wrote: It appears that i lied, at least partially: Martin shows us the 'ls' command (which is new to me!). Thanks guys. The json command doesn't exist in the Fossil I have (1.24 [0c65916136]): === [C:\]fossil json dir -checkin experimental fossil: fossil: unknown command: json fossil: use help for more information === It seems a bit involved just to know what revisions are available in a given branch. What about switching to another branch (eg. going from trunk to experimental), and simply running fossil ls? Will that list the revisions without touching the files in my work directory? ___ 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] Right way to try something new and save/revert?
On Fri, 25 Jan 2013 14:14:20 +0100, Gilles gilles.gana...@free.fr wrote: What about switching to another branch (eg. going from trunk to experimental), and simply running fossil ls? Will that list the revisions without touching the files in my work directory? I notice that fossil update experimental -n doesn't say which work files will be replaced by what's in the experimental branch in the repo (1 file modified.): === UPDATE dummy.c --- updated-to: 80ebe786997e4fc943bbc2d9564f8bb727e81675 2013-01-11 16:36:56 UTC tags: experimental comment: Testing branches (user: Fred) changes: 1 file modified. === Generally speaking, I find running update followed by commands to list which files/revisions are available and their contents (ls, timeline, finfo, etc.) a bit dangerous. It'd be coold if there were a simple way to list files/revisions in a branch without actually checking out anything into the work directory. ___ 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] Right way to try something new and save/revert?
On Fri, 25 Jan 2013 14:38:25 +0100, Gilles gilles.gana...@free.fr wrote: I notice that fossil update experimental -n doesn't say which work files will be replaced by what's in the experimental branch in the repo (1 file modified.): === UPDATE dummy.c --- updated-to: 80ebe786997e4fc943bbc2d9564f8bb727e81675 2013-01-11 16:36:56 UTC tags: experimental comment: Testing branches (user: Fred) changes: 1 file modified. === My mistake: I was looking lower and didn't see the UPDATE line. Not big enough :-D ___ 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] Right way to try something new and save/revert?
On Fri, Jan 25, 2013 at 2:59 AM, Gilles gilles.gana...@free.fr wrote: On Wed, 09 Jan 2013 12:10:35 +0100, Gilles gilles.gana...@free.fr wrote: Am I correct in understanding that this is the right way to proceed to try some new code, and either save it (whether it works or not, just as a track-record) or discard it? I have another question: fossil branch ls lists branches available in the repo, but is there a command to list all the files/revisions that have been commited to the experimental branch? I think 'fossil diff --brief --branch experimental' might be what you're looking for. Thank you. ___ 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] Right way to try something new and save/revert?
On Fri, 11 Jan 2013 20:55:23 +0400, Konstantin Khomoutov flatw...@users.sourceforge.net wrote: The logic behind `fossil update` is like follows. Thanks for the infos + suggestion in the other post. ___ 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] Right way to try something new and save/revert?
On Fri, 11 Jan 2013 13:06:17 +0100 Gilles gilles.gana...@free.fr wrote: Am I correct in understanding that this is the right way to proceed to try some new code, and either save it (whether it works or not, just as a track-record) or discard it? So the right way to experiment and keep tried code for later reference is to save those trials in a branch, distinct from the trunk. But since I've never used branching, I'm not clear about how to see which files have been saved in the branch vs. those in the trunk, You don't need this: a branch diverges from the trunk (in your case), rather than appearing out of the void, so initially it has the same files in the same state the trunk does. Moreover, the Fossil's philosophy is that a new branch is only created when the first commit on it is made, so you start with the trunk, do some changes then do fossil commit --branch crazy_idea and only then your new branch comes into existence. Getting the diff between the starting point of the branch (or an earlier state) is quite rarely needed -- most of the time you're interested in the diff between your work directory (local changes) and its base version (the last recorded commit on the current branch). If you still need the diff between two arbitrary commits, use fossil diff --from A --to B where A and B are either branch/tag names or (possibly abbreviated) hashes of the relevan commits, extracted from the timeline view. You can omit either --from or --to, so, to diff your current checkout to the trunk, just use fossil diff --from trunk and how to check the changes made in a branch to such and such file, Fire up Fossil web UI and click on the links marked patch and diff in the commit view. and how to go back and forth between the trunk and the branch before applying commit. Technically, just by doing `fossil update branchname` as this tries to merge your uncommitted local changes with the branchname you're switching your open checkout to. But, again, what you're asking for is not needed as this is not how VCSes work, so don't do that. Instead, implement your feature on your branch, and if you're OK with it and want it in your mainline (the trunk), go back to the trunk and then *merge* your feature branch into it -- this will bring all the changes implemented on that feature branch and will record that your new commit has two parents (the trunk and the feature branch) so you have understanable history record. ___ 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] Right way to try something new and save/revert?
On Fri, 11 Jan 2013 17:19:40 +0400, Konstantin Khomoutov flatw...@users.sourceforge.net wrote: and how to check the changes made in a branch to such and such file, Fire up Fossil web UI and click on the links marked patch and diff in the commit view. This is really what I want to do: Being able to see all the things I tried on a file in the branch. Most of the time, I want to keep track of things I tried just in case instead of just forgetting about them with fossil revert. Can this be done with calling a Windows application, either a regular editor or a differ instead of the Fossil web server? I prefer working with dedicated applications. Technically, just by doing `fossil update branchname` as this tries to merge your uncommitted local changes with the branchname you're switching your open checkout to. I was under the impression that update simply told Fossil than any subsequent commit should be done on that thread/trunk. Good to know that update will perform a merge, so it has consequences. ___ 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] Right way to try something new and save/revert?
On Fri, 11 Jan 2013 17:11:44 +0100, Gilles gilles.gana...@free.fr wrote: This is really what I want to do: Being able to see all the things I tried on a file in the branch. Most of the time, I want to keep track of things I tried just in case instead of just forgetting about them with fossil revert. What I'm driving at: 1. Keep tried but NOK algos in a branch called eg. experimental 2. Find a simple way to locate old algo's I know I tried before by searching Fossil, regardless of which branch they are (trunk or experimental). I know Fossil doesn't support grep yet, but I find it useful to being able to just commit code and know for sure that I can find anything in the repo. ___ 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] Right way to try something new and save/revert?
On Fri, 11 Jan 2013 17:11:44 +0100 Gilles gilles.gana...@free.fr wrote: Fire up Fossil web UI and click on the links marked patch and diff in the commit view. This is really what I want to do: Being able to see all the things I tried on a file in the branch. Most of the time, I want to keep track of things I tried just in case instead of just forgetting about them with fossil revert. `fossil revert` is there for special whoops! cases, not for organising some sort of serious workflow around it. Can this be done with calling a Windows application, either a regular editor or a differ instead of the Fossil web server? I prefer working with dedicated applications. There's no official GUI front-end for fossil, but there are several attempts at creating such a thing. Google mail archives for it, using something like http://www.google.com/search?q=GUI+site%3Amail-archive.com%2Ffossil-users%40lists.fossil-scm.org to get the relevant discussions/links. Using GUI diffing programs was also discussed several times. Technically, just by doing `fossil update branchname` as this tries to merge your uncommitted local changes with the branchname you're switching your open checkout to. I was under the impression that update simply told Fossil than any subsequent commit should be done on that thread/trunk. Good to know that update will perform a merge, so it has consequences. The logic behind `fossil update` is like follows. Your work tree (working directory, current checkout and whatever else it's being named) is always based on some state stored in the repository (except for the special case when the repository has no commits at all yet). So the main task of `fossil update` is to bring your work tree to the state held by the commit you supplied to this command (directly, or via the name of a branch or a tag). Of course, when fossil does this, it has to be non-destructive: it should not stomp over untracked files (files present in the work tree but not being managed by fossil) and it should preserve your local modifications, if any, which it does by merging. After than, when you modify something and record a commit, that new commit inherits certain properties from the base commit of your work tree, and if that commit is on a branch, then your new commit will be on that same branch, too. ___ 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] Right way to try something new and save/revert?
On Fri, 11 Jan 2013 17:25:24 +0100 Gilles gilles.gana...@free.fr wrote: [...] What I'm driving at: 1. Keep tried but NOK algos in a branch called eg. experimental 2. Find a simple way to locate old algo's I know I tried before by searching Fossil, regardless of which branch they are (trunk or experimental). I know Fossil doesn't support grep yet, but I find it useful to being able to just commit code and know for sure that I can find anything in the repo. You could go for documenting your experiments then. First of all, you can assign arbitrary tags to the tip commits of your experimental lines of history. A tag view is there to select the one you need. Another easy possibilty is the wiki: you can refer to any commit using its SHA-1 name in square brackets, so you can create a page titled, say, experiments and roll something like this: h2My experiments/h2 * [44c9e93a9187a07340da8b2984b2118990637d5a|A crazy idea] * [8a428d5c2a0097418170509ab03faed0a335f796|Using bars to style foos] ... where you get those SHA-1 commit names from the commit views. The end result is that you have a nice-looking list of your experimental lines of history, decorated by arbitrary comments. ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
[fossil-users] Right way to try something new and save/revert?
Hello Am I correct in understanding that this is the right way to proceed to try some new code, and either save it (whether it works or not, just as a track-record) or discard it? To try some new code: 1. Commit current code 2. Try new code 3. a. if OK, commit new code : fossil commit -m New stuff b. if NOK and don't care to save it, just go back to previous code: fossil revert myfile.c c. if NOK but want to keep track of attempt, commit and go back to n-1 : fossil commit -m Failed attempt fossil finfo myfile.c : write down UUID (first hash) of n-1 revision fossil revert -r UUID myfile.c Thank you. ___ 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] Right way to try something new and save/revert?
On Wed, 09 Jan 2013 12:10:35 +0100 Gilles gilles.gana...@free.fr wrote: Hello Am I correct in understanding that this is the right way to proceed to try some new code, and either save it (whether it works or not, just as a track-record) or discard it? To try some new code: 1. Commit current code 2. Try new code 3. a. if OK, commit new code : fossil commit -m New stuff b. if NOK and don't care to save it, just go back to previous code: fossil revert myfile.c c. if NOK but want to keep track of attempt, commit and go back to n-1 : fossil commit -m Failed attempt fossil finfo myfile.c : write down UUID (first hash) of n-1 revision fossil revert -r UUID myfile.c If in your step 2 you add/remove/move files you should fossil addremove before the commit. Remember to open the repository before any change to your code, if not, when you do fossil open rep.fossil, it will note it's different from fossilized source tree and undo all your changes. Thank you. ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users --- --- Eduardo Morras emorr...@yahoo.es ___ 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] Right way to try something new and save/revert?
On Wed, 9 Jan 2013 12:23:23 +0100, Eduardo Morras emorr...@yahoo.es wrote: 2. Try new code If in your step 2 you add/remove/move files you should fossil addremove before the commit. Thanks for the tip. I usually only make changes to existing files, usually just localized changes in a single file, but it's indeed important to keep in mind. Remember to open the repository before any change to your code, if not, when you do fossil open rep.fossil, it will note it's different from fossilized source tree and undo all your changes. I always leave Fossil open, so it's not an issue, but I'll keep that in mind. Thank you. ___ 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] Right way to try something new and save/revert?
On Wed, Jan 9, 2013 at 6:10 AM, Gilles gilles.gana...@free.fr wrote: Hello Am I correct in understanding that this is the right way to proceed to try some new code, and either save it (whether it works or not, just as a track-record) or discard it? To try some new code: 1. Commit current code 2. Try new code 3. a. if OK, commit new code : fossil commit -m New stuff b. if NOK and don't care to save it, just go back to previous code: fossil revert myfile.c c. if NOK but want to keep track of attempt, commit and go back to n-1 : fossil commit -m Failed attempt fossil finfo myfile.c : write down UUID (first hash) of n-1 revision fossil revert -r UUID myfile.c I'd do it this way: fossil commit -m Failed Attempt --branch dead-end fossil up prev That will save your failed attempt in your history forever. The failed attempt will be on a branch, however, out of the way. I avoid putting non-working code on the trunk, since that will cause problems for future bisects. If you don't what to save your failed attempt forever, but just want to keep it around for a few days (or weeks) in case you later decide to go back it it, I'd save it in the stash instead: fossil stash save The previous command will save all your changes in a stash entry, then revert your check-out to its pristine state. -- 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] Right way to try something new and save/revert?
Hello, Much nicer: fossil add kk.cc fossil commit -m v1 (some failed changes) fossil commit -m failed attempt --branch failed_attempt fossil update trunk (some changes) fossil commit -m good Compass Ing. y Sistemas Dr. Ramon Ribo http://www.compassis.com ram...@compassis.com c/ Tuset, 8 7-2 tel. +34 93 218 19 89 08006 Barcelona, Spainfax. +34 93 396 97 46 2013/1/9 Gilles gilles.gana...@free.fr: Hello Am I correct in understanding that this is the right way to proceed to try some new code, and either save it (whether it works or not, just as a track-record) or discard it? To try some new code: 1. Commit current code 2. Try new code 3. a. if OK, commit new code : fossil commit -m New stuff b. if NOK and don't care to save it, just go back to previous code: fossil revert myfile.c c. if NOK but want to keep track of attempt, commit and go back to n-1 : fossil commit -m Failed attempt fossil finfo myfile.c : write down UUID (first hash) of n-1 revision fossil revert -r UUID myfile.c Thank you. ___ 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] Right way to try something new and save/revert?
On Wed, 09 Jan 2013 12:10:35 +0100 Gilles gilles.gana...@free.fr wrote: Am I correct in understanding that this is the right way to proceed to try some new code, and either save it (whether it works or not, just as a track-record) or discard it? To try some new code: 1. Commit current code 2. Try new code 3. a. if OK, commit new code : fossil commit -m New stuff b. if NOK and don't care to save it, just go back to previous code: fossil revert myfile.c c. if NOK but want to keep track of attempt, commit and go back to n-1 : fossil commit -m Failed attempt fossil finfo myfile.c : write down UUID (first hash) of n-1 revision fossil revert -r UUID myfile.c AFAIK, the paradigm to handle messed up commits in Fossil is to move them to a special branch, say, named mistake. You do this from the web interface. So basically you can do a series of commits and then decide that idea was a dead-end. So you can go to the first commit of that dead-end leaf and change its branch to, say, dead-end. You then can `fossil up` to the commit preceding the first commit in the dead-end leaf and start hacking away a new line of history. ___ 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] Right way to try something new and save/revert?
On Wed, 9 Jan 2013 06:37:54 -0500, Richard Hipp d...@sqlite.org wrote: I'd do it this way: fossil commit -m Failed Attempt --branch dead-end fossil up prev That will save your failed attempt in your history forever. The failed attempt will be on a branch, however, out of the way. I avoid putting non-working code on the trunk, since that will cause problems for future bisects. If you don't what to save your failed attempt forever, but just want to keep it around for a few days (or weeks) in case you later decide to go back it it, I'd save it in the stash instead: fossil stash save The previous command will save all your changes in a stash entry, then revert your check-out to its pristine state. Thanks much for the feedback. I guess the time has come to start using branching :-) I didn't know about stash but I'd rather keep all attemps with no limit, so I guess I'll use branching instead. Ramon below mentions fossil update trunk instead of fossil update prev: Does it make a difference? ___ 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] Right way to try something new and save/revert?
On Wed, Jan 9, 2013 at 7:13 AM, Gilles gilles.gana...@free.fr wrote: Ramon below mentions fossil update trunk instead of fossil update prev: Does it make a difference? Not in this case, since the previous check-in will be on trunk - they will be the same thing. -- 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] Right way to try something new and save/revert?
On Wed, 9 Jan 2013 07:23:07 -0500, Richard Hipp d...@sqlite.org wrote: Ramon below mentions fossil update trunk instead of fossil update prev: Does it make a difference? Not in this case, since the previous check-in will be on trunk - they will be the same thing. Thank you. I'll experiment and see how it goes. ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users