Re: [PATCH 1/3] backfill: reject rev-list arguments that do not make sense
- From
Derrick Stolee <stolee@gmail.com>
- Date
- Apr 16, 2026, 14:11 UTC
- Message-ID
- <9885d0e1-2d0f-4cde-b4ed-671a3ba173e5@gmail.com>
- In-Reply-To
- <6c76f1e862790858297ff9b1debc1a38756b1913.1776297482.git.gitgitgadget@gmail.com>
On 4/15/2026 7:58 PM, Elijah Newren via GitGitGadget wrote:
Show 22 quoted lines
> From: Elijah Newren <newren@gmail.com> > > Some rev-list options accepted by setup_revisions() are silently > ignored or actively counterproductive when used with 'git backfill', > because the path-walk API has its own tree-walking logic that bypasses > the mechanisms these options rely on: > > * -S/-G (pickaxe) and --diff-filter work by computing per-commit > diffs in get_revision_1() and filtering commits whose diffs don't > match. Since backfill's goal is to download all blobs reachable > from commits in the range, filtering out commits based on diff > content would silently skip blobs -- the opposite of what users > want. > > * --follow disables path pruning (revs->prune) and only makes > sense for tracking a single file through renames in log output. > It has no useful interaction with backfill. > > * -L (line-log) computes line-level diffs to track the evolution > of a function or line range. Like pickaxe, it filters commits > based on diff content, which would cause blobs to be silently > skipped.
I think these make a lot of sense, especially because these computations require downloading missing blobs in order to find the diffs that justify some of the choices of commit filtering.
> * --diff-merges controls how merge commit diffs are displayed. > The path-walk API walks trees directly and never computes > per-commit diffs, so this option would be silently ignored.
I think there are a few other "format" based options that were silently ignored on purpose, because there's no output. Perhaps we should change the use of options like this to a warning instead of a failure?
> * --filter (object filtering, e.g. --filter=blob:none) is used by > the list-objects traversal but is completely ignored by the > path-walk API, so it would silently do nothing.
This is correct to remove because while it doesn't work with path-walk right now, it might in the future. We don't want the filter to mess with the functionality of 'git backfill' that sets its own scope for which blobs to download.
> Rather than letting users think these options are being honored, > reject them with a clear error message.
I agree that the majority of these should be hard failures. As mentioned, some could be soft warnings. That could be an adjustment to make in the future, so is not blocking for this patch.
Show 22 quoted lines
> +static void reject_unsupported_rev_list_options(struct rev_info *revs)
> +{
> + if (revs->diffopt.pickaxe)
> + die(_("'%s' cannot be used with 'git backfill'"),
> + (revs->diffopt.pickaxe_opts & DIFF_PICKAXE_REGEX) ? "-G" : "-S");
> + if (revs->diffopt.filter || revs->diffopt.filter_not)
> + die(_("'%s' cannot be used with 'git backfill'"),
> + "--diff-filter");
> + if (revs->diffopt.flags.follow_renames)
> + die(_("'%s' cannot be used with 'git backfill'"),
> + "--follow");
> + if (revs->line_level_traverse)
> + die(_("'%s' cannot be used with 'git backfill'"),
> + "-L");
> + if (revs->explicit_diff_merges)
> + die(_("'%s' cannot be used with 'git backfill'"),
> + "--diff-merges");
> + if (revs->filter.choice)
> + die(_("'%s' cannot be used with 'git backfill'"),
> + "--filter");
> +}
> +My only nit-pick suggestion is to make the translated string a macro so it can be more obvious that it is repeated exactly.
Thanks, -Stolee