Backtestshugorun-20260930-140149#15078
New content command thinks there's a path conflict for a file that doesn't exist yet
Our patch was judged better than the merged fix.
In this judgment, Patch B = our agent, Patch A = the merged fix.
Both patches make the same production change: they add `draft: false` to the placeholder front matter in `CreateContentPlaceHolder`. This stops the section's `cascade: draft: true` from marking the temporary placeholder as a draft, which is what excluded it from the build and caused the 'no page found / conflicts with existing content' error. Correctness is therefore identical, and the comparison comes down to tests. Patch A adds a Go unit test in `create/content_test.go` that reproduces the drafts section with the cascade. It then asserts on `title: "Trip to Puebla"` in YAML form. Whether that holds depends on the archetypes that `initFs` sets up, so the assertion is fragile and may not match the actual archetype output format. It covers only the leaf-bundle `index.md` case. Patch B adds a testscript that runs the real `hugo new content` command end to end in a fresh project. It uses the user's exact `_index.md` front matter, including `weight`. It covers both scenarios the issue mentions: the `trip-to-puebla/index.md` bundle and a plain `another-post.md` file. It also checks explicitly that the output is not the 'Content Placeholder' and that the archetype-derived title is present. This matches the issue's CLI reproduction more closely, tests the specific symptom (the placeholder being left behind), and would clearly fail before the fix. Scope and craft are comparable. Patch B's test is broader and closer to the reported scenario, so B is the stronger patch.
The judge saw only the two patches, labelled A and B in random order, with issue links and version numbers removed. The reasoning above is shown verbatim.
Solve
33m 22s17 turns
Wall-clock time in the container
Full test suite
No new failures
3,671 passed · 93 skipped · 3,773 total
Contamination audit
Clean
No sign the agent saw the real fix
Patch size
29vs 32 merged
Changed lines, ours vs merged
The issue
View on GitHubI run my personal blog using Hugo for about two years ago, and it has been great!, but there have been issues with how I create pages.
I have section called posts/drafts/ where I create drafts for posts before I move them to a posts/YEAR/MONTH directory where I can post them. The bug happens when I create the page for a post, like this:
<pre>
hugo new content posts/drafts/trip-to-puebla/index.md
</pre>
Running this command with the debug flag returns the following:
<pre>
INFO build: step process substep collect files 138 files_total 138 pagesources_total 71 resourcesources_total 67 duration 37.1674ms
INFO build: step process duration 38.2639ms
INFO build: step assemble duration 2.064063s
INFO build: step render pages 0 content 0 duration 548.6µs
INFO build: step render deferred count 0 duration 0s
INFO build: step postProcess duration 0s
INFO build: duration 2.1061069s
ERROR no page found for "C:\\...\\Documents\\projects\\kfcmanwebsite2\\content\\posts\\drafts\\trip-to-puebla\\index.md"; the target path conflicts with existing conten
</pre>
What's funny is that a page _is_ created, but it's a placeholder page that looks like this:
---
title: "Content Placeholder"
build:
render: never
list: never
publishResources: false
---
After consulting with ChatGPT (yes you can stone me now) and testing some things here's another stuff that cursed AI and I found:
- The name of the page still causes some issues, doesn't matter if its
trip-to-puebla/index.mdortrip-to-puebla. - Renaming the
draftsdirectory and rerunning the command with the new name still throws the error - This issue does not happen if I run it when creating something like
hugo new content posts/2026/07/foo/index.md
One final thing, I have a branch bundle on posts/drafts/_index.md that has the following front matter:
title: Drafts
build:
render: never
cascade:
draft: true
build:
render: link
weight: 9999
draft: true
This is so any of the pages inside posts/drafts don't get published.
What version of Hugo are you using (hugo version)?
<pre>
hugo v0.163.3-4d22555aebf458d5d150500c9ac4bee5b24cf0d3+extended windows/amd64 BuildDate=2026-06-18T16:18:24Z VendorInfo=gohugoio
</pre>
Does this issue reproduce with the latest release?
I'm using the latest version, so it _should_ be reproducible. I think. The repo for my hugo project can be found here, though it may be a little heavy.
The two patches
hugolib/content_factory.goindex 36cb29e63..6757c64e3 100644@@ -114,6 +114,7 @@ func (f ContentFactory) CreateContentPlaceHolder(filename string, force bool) (s // the paths correct. placeholder := `--- title: "Content Placeholder"+draft: false build: render: never list: nevertestscripts/commands/new_content.txtindex f432c8ace..3b30520e9 100644@@ -43,6 +43,22 @@ hugo new content P1.md ! hugo new content p1.md stderr 'the target path conflicts with existing content' +# Issue: new content under a draft section with cascade draft: true++cd $WORK+hugo new project drafts-cascade+mkdir drafts-cascade/content/posts/drafts+cp drafts-cascade-index.md drafts-cascade/content/posts/drafts/_index.md+cd drafts-cascade+hugo new content posts/drafts/trip-to-puebla/index.md+exists content/posts/drafts/trip-to-puebla/index.md+! grep 'Content Placeholder' content/posts/drafts/trip-to-puebla/index.md+grep 'title = .Trip to Puebla.' content/posts/drafts/trip-to-puebla/index.md+hugo new content posts/drafts/another-post.md+exists content/posts/drafts/another-post.md+! grep 'Content Placeholder' content/posts/drafts/another-post.md+grep 'title = .Another Post.' content/posts/drafts/another-post.md+ -- hugo.toml -- disableKinds = ['home','rss','section','sitemap','taxonomy','term'] [frontmatter]@@ -52,3 +68,15 @@ publishDate = [':filename', ':default'] DATE _{{ .Date.Format "2006-01-02" }}_ SLUG _{{ .Slug }}_ TITLE _{{ .Title }}_+-- drafts-cascade-index.md --+---+title: Drafts+build:+ render: never+cascade:+ draft: true+ build:+ render: link+weight: 9999+draft: true+---
create/content_test.goindex deafaed09..c8c3c1cdd 100644@@ -159,6 +159,37 @@ site RegularPages: {{ len site.RegularPages }} cContains(c, readFileFromFs(t, fs.Source, filepath.Join("content", "mypage.md")), `draft: true`) } +// See issue 15078.+func TestNewContentWithBuildCascade(t *testing.T) {+ t.Parallel()++ mm := afero.NewMemMapFs()+ c := qt.New(t)++ c.Assert(initFs(mm), qt.IsNil)+ c.Assert(mm.MkdirAll(filepath.Join("content", "posts", "drafts"), 0o755), qt.IsNil)+ c.Assert(afero.WriteFile(mm, filepath.Join("content", "posts", "drafts", "_index.md"), []byte(`---+title: Drafts+build:+ render: never+cascade:+ draft: true+ build:+ render: link+draft: true+---+`), 0o755), qt.IsNil)++ cfg, fs := newTestCfg(c, mm)+ conf := testconfig.GetTestConfigs(fs.Source, cfg)+ h, err := hugolib.NewHugoSites(deps.DepsCfg{Configs: conf, Fs: fs})+ c.Assert(err, qt.IsNil)++ const target = "posts/drafts/trip-to-puebla/index.md"+ c.Assert(create.NewContent(h, "", target, false), qt.IsNil)+ c.Assert(readFileFromFs(c, fs.Source, filepath.Join("content", target)), qt.Contains, `title: "Trip to Puebla"`)+}+ func initFs(fs afero.Fs) error { perm := os.FileMode(0o755) var err errorhugolib/content_factory.goindex 36cb29e63..6757c64e3 100644@@ -114,6 +114,7 @@ func (f ContentFactory) CreateContentPlaceHolder(filename string, force bool) (s // the paths correct. placeholder := `--- title: "Content Placeholder"+draft: false build: render: never list: never
What the agent reported
hugo new content posts/drafts/trip-to-puebla/index.md now works under a section with cascade: draft: true: it writes the real archetype content instead of stopping with the "conflicts with existing content" error. The official ./bt-test gate never actually ran the tests, so all results below come from running go test directly.
What was wrong: Before building the page, hugo new content writes a temporary placeholder file (the "Content Placeholder" page you saw) and then looks that page up. That file said nothing about drafts, so your cascade: draft: true made it a draft. hugo new doesn't build drafts, so the placeholder was left out and the lookup found nothing. The error text calls that a path conflict, though no conflict existed, and the placeholder was left on disk.
What I changed:
hugolib/content_factory.go: the placeholder front matter now includesdraft: false. A value the page sets itself overrides a cascaded one. The placeholder is always replaced by the archetype output, so this never shows up in your file; the final draft setting still comes from your archetype.- I first tried building drafts during
hugo newinstead. That broke an existing test (Issue 12786), where creatingp1/p2.mdnext to an existingp1/index.mdmust be rejected. That check only works because the existing page is a draft and gets left out, so I dropped that approach.
Test evidence:
- New test: I added a case to
testscripts/commands/new_content.txtusing your exact_index.mdfront matter. It runshugo new content posts/drafts/trip-to-puebla/index.mdandhugo new content posts/drafts/another-post.md, then checks that each file exists, no longer says "Content Placeholder", and has the archetype title. - Before the fix it failed with
no page found for ".../trip-to-puebla/index.md"; the target path conflicts with existing content. - After the fix it passes, along with the rest of the
newtest scripts, including the existing conflict cases. - Why not
./bt-test:./bt-test . -run TestCommands/new_contentstopped during setup because it couldn't download Go modules (proxy.golang.org is unreachable here). I ran the tests withgo testagainst the local module cache at/bt/cache/go-modinstead. - Other tests:
./create/...passes.TestCommandsand./hugolib/each have failures, but it's the same module and git-info tests that need network access, with and without my change (checked by stashing it). - Review: an independent review found no blocking issues.
Your observations:
| # | What you reported | Outcome |
|---|---|---|
| 1 | trip-to-puebla/index.md under posts/drafts fails with the conflict error | Fixed; covered by the new test |
| 2 | A "Content Placeholder" file is left behind | Fixed for this case; the test asserts the placeholder text is gone. If hugo new fails for another reason, the placeholder is still left behind (unchanged). |
| 3 | Both the trip-to-puebla/index.md and trip-to-puebla forms fail | Explained, and fixed through the same cause. The test covers the index.md form and a single file (another-post.md); the bare trip-to-puebla form wasn't tested separately. |
| 4 | Renaming drafts doesn't help | Explained: the cause is the cascaded draft: true, not the folder name |
| 5 | posts/2026/07/foo/index.md works | Explained: there's no draft cascade there; draft already defaulted to false, so this path is unaffected |
The same false error could still happen if a section cascades an expiry date that has passed; I left that alone as out of scope. The already-deleted AGENTS.md and CLAUDE.md were that way before I started and are untouched. Everything is left uncommitted.