I am curious how you arrived at 707443 . Can you remember how you did it? Code-wise it is the same as a9c4423 , but log-wise it is the same as a8afe54 . Code-wise they differ by some re-indentation and some additional comments. It is almost as if you have arrived at a8afe54 then rebase and squash a diff and drop the additional entry, or start with a9c4423 and edited the top part of the log.
About ccfd2c, please either push it out as a branch, or posting a log/whatchanged listing among the documentation. I see in ccfd2c, you have squashed it down from 178 commits to 155, and filled out some of the messages with more details. The latter is good, though I am not sure about the former. I think some of the earlier commits really should have been split into a few, not combined. I believe I describe how it can be done briefly, but let me give you a more detailed procedure. Let's say you have make changes to file uvw and file pqr, and committed the two changes as commit A. However, they are unrelated changes and logically should be separate. So you make commit -B, which undo the change in uvw only, and you do commit B which redo the change in uvw. Then you can rebase and squash A and -B together, and you end up with commit A-B which contains changes in pqr only, and B which contains changes in uvw only. You can split changes like this, after you have done A, and X, Y, Z... and sometimes later, as long as none of X, Y, Z changes either uvw or pqr closer than a few lines where A changes them. This mode of working is often necessary when summitting patches to projects with very strict policies about "one patch per only one type/purpose/reason of change'. To be honest, it is often necessary when working on large colaborative projects, and a good habit even for solo ones, just to not get confused by one's own work 6 months later. Anyway, rebasing public branches are almost never done without a good reason, and also usually accompanied by *advanced* warnings to anybody else who might be working simultaneously on the same code: "Your work will break!" *before* doing so. OTOH, it is acceptable to post "watch only" work-in-progress private branches and rebase/rewrite the hell of them, so you could and you should post your own 'better annotation' branches, as long as it is for reference only and it is generally understood that others should not be basing their own work on those. -------------------------------------------- On Sun, 17/8/14, Hin-Tak Leung <[email protected]> wrote: okay. The additional work leading up to ccfd2c and 707443 shouldn't go to waste - you should either push them as 'better-written-branches, or just including the listing, in 'git log' form, or better still, in 'git whatchanged' form, among the documentation. I hope you learned a few interesting things about git on the way - being able to recover 'lost commits' could be a useful skill; also gain some understanding of the use of sha1 checksum to identify commits and their integrity/security implications. ------------------------------ On Sun, Aug 17, 2014 09:22 BST ahi wrote: >Head should now be at a8afe54 > > >On Sun, Aug 17, 2014 at 11:08 AM, Hin-Tak Leung <[email protected] >> wrote: > >> >> You have made a mistake some where - as well as tried to change things >> while you restore. You seem to have restored the code to a9c4423 instead of >> a8afe54, as well as tried to modify a9c4423, so your current head is >> 707443... your head should be restored to a8afe54, or decendent of it, not >> anything else. What additional change did you try to make while restoring? >> a restore is a restore, do not try to make restore=somethingelse. If you >> need to make further changes, then that's further commits. >> >> =================== >> okay, as far as I see, you haven't been using rebase much until recently, >> so there are only two possible restoring points -a8afe54 or a9c4423 . Both >> of them satisfy "git log a8afe54..dc1796075ba452c8e92df5fa252aaaaf91bd >> 0f9b" and "git log a9c4423..dc1796075ba452c8e92df5fa252aaaaf91b d0f9b" >> shows nothing. >> ========================= >> >> >> >> ------------------------------ >> On Sun, Aug 17, 2014 08:31 BST ahi wrote: >> >> >The repo should now be restored to it's previous state. >> >git log HEAD..dc1796075ba452c8e92df5fa252aaaaf91bd0f9b >> >shows nothing. >> > >> > >> >-- >> >Best Regards, >> >Ahmed >> >> > > >-- >Best Regards, >Ahmed _______________________________________________ Spdx-tech mailing list [email protected] https://lists.spdx.org/mailman/listinfo/spdx-tech
