On 6/28/2022 12:54 PM, Junio C Hamano wrote: > "Derrick Stolee via GitGitGadget" <gitgitgadget@xxxxxxxxx> writes: > >> 3. Focus on the edits that most-recently edited these lines. Doing some >> scripting, I was able to construct this date-sorted list of previous >> edits (by diffing the git blame output before and after this >> change). The most-recent changes before this are: >> >> 2005-08-26: 52a22d1e726 ([PATCH] Subject: [PATCH] Add some documentation., 2005-08-26) >> ... >> 2022-04-20: 9e5ebe9668a (rebase: use correct base for --keep-base when a branch is given, 2022-04-20) > > That is a new concept ;-) > > It is an interesting exercise to see which previous changes had > these mark-up mistakes, but it is not immediately obvious to me how > we can take advantage of the information. I just wanted to see how much of these edits were due to "really old documentation from when we had different habits" and how many were related to recent updates, so I should give more pause. I think the discussion about the "apply" versus "merge" backends should have given me more pause because those sections were written recently and are more likely to have correct punctuation. >> I look forward to feedback on how to do this better (if it is indeed a >> good idea to do in the first place). > > Correcting mark-up to result in an easier-to-read documentation is a > good idea, of course. I wonder if we can also help the developers > mark-up correctly in their first attempt (e.g. do we have clear and > concise guidelines that are well publicized?) I'm not sure we have them well publicized. It's worth thinking about. Hopefully creating a more standard use within at least one (long) file will help establish some patterns organically. >> --strategy-option=<strategy-option>:: >> Pass the <strategy-option> through to the merge strategy. >> This implies `--merge` and, if no strategy has been >> - specified, `-s ort`. Note the reversal of 'ours' and >> - 'theirs' as noted above for the `-m` option. >> + specified, `-s ort`. Note the reversal of `ours` and >> + `theirs` as noted above for the `-m` option. > > These references to "ours" and "theirs" is what we called out > earlier in the "swapped" description (hunk -348,10), i.e. > > when a merge conflict happens, the side reported as 'ours' > is the so-far rebased series ... and 'theirs' is the working > branch. > > which the patch left in 'emphasis' not `verbatim`. I think this > section should do the same. > > The 'ours' (but not 'theirs' because there is no such thing) that is > explained as useless as a strategy in the intervening paragraph > (hunk -360,9) refers to the name of a strategy, and it is correct to > mark it as `verbatim`. Thanks! >> --ignore-whitespace:: >> Ignore whitespace differences when trying to reconcile >> -differences. Currently, each backend implements an approximation of >> -this behavior: >> + differences. Currently, each backend implements an approximation of >> + this behavior: >> + >> -apply backend: When applying a patch, ignore changes in whitespace in >> +'apply backend:' When applying a patch, ignore changes in whitespace in >> ... >> -merge backend: Treat lines with only whitespace changes as unchanged >> +'merge backend:' Treat lines with only whitespace changes as unchanged > > It somehow looks curious (at the source level---I haven't seen the > formatted HTML output) to have the punctuation colon as part of the > phrase marked up. I wonder if these were meant to be more like so: > > apply backend;; > When applying a patch, ... > > merge backend;; > Treat lines with ... This appears to work well. The doc-diff shows this: - apply backend: When applying a patch, ignore changes in whitespace - in context lines. Unfortunately, this means that if the "old" lines - being replaced by the patch differ only in whitespace from the - existing file, you will get a merge conflict instead of a - successful patch application. + apply backend: + When applying a patch, ignore changes in whitespace in context + lines. Unfortunately, this means that if the "old" lines being + replaced by the patch differ only in whitespace from the + existing file, you will get a merge conflict instead of a + successful patch application. - merge backend: Treat lines with only whitespace changes as - unchanged when merging. Unfortunately, this means that any patch - hunks that were intended to modify whitespace and nothing else will - be dropped, even if the other side had no changes that conflicted. + merge backend: + Treat lines with only whitespace changes as unchanged when + merging. Unfortunately, this means that any patch hunks that + were intended to modify whitespace and nothing else will be + dropped, even if the other side had no changes that conflicted. >> @@ -536,8 +536,8 @@ See also REBASING MERGES and INCOMPATIBLE OPTIONS below. >> >> -x <cmd>:: >> --exec <cmd>:: >> - Append "exec <cmd>" after each line creating a commit in the >> - final history. <cmd> will be interpreted as one or more shell >> + Append `exec <cmd>` after each line creating a commit in the >> + final history. `<cmd>` will be interpreted as one or more shell >> commands. Any command that fails will interrupt the rebase, >> with exit code 1. > > As noticed by others, "git help -m rebase" is somewhat harmed with > this change when rendered to plain text without any attributes. The > roff output actually is > > .RS 4 > Append > \fBexec <cmd>\fR > after each line creating a commit in the final history\&. > > and even on plain text tty, "exec <cmd>" part is now shown in bold > (as opposed to be in plain text inside double quotes, which was how > the original got rendered). So I think this change is an > improvement. Thanks for digging into the details here. It would be interesting if doc-diff could make this clearer, but that's a project for another time. Thanks, -Stolee
