Write a form-error page's status only after it renders (closes #128)

The login form and the webhook create and edit forms set their 400 or
409 with WriteHeader before calling renderTemplate, so a template
failure there answered with that status instead of 500. A variant of
renderTemplate now takes the status and writes it only once the page
has rendered into the buffer; renderTemplate sends 200 through it.

Model: opus-5-5

internal/handlers/auth.go

@@ -139,8 +139,7 @@ func (h *Handlers) renderLoginError(
 		),
 	}
 
-	w.WriteHeader(status)
-	h.renderTemplate(w, r, "login.html", data)
+	h.renderTemplateStatus(w, r, "login.html", data, status)
 }
 
 // authenticateUser looks up and verifies a user's credentials.

internal/handlers/auth_test.go

@@ -3,6 +3,7 @@ package handlers_test
 import (
 	"context"
 	"fmt"
+	"html/template"
 	"net/http"
 	"net/http/httptest"
 	"net/url"
@@ -404,6 +405,60 @@ func TestLogin_MissingCredentialsRejectedBeforeAnyHash(t *testing.T) {
 	)
 }
 
+// TestLogin_FormErrorAnswersItsStatusWithThePage proves that the login
+// form shown again with an error still answers 400 with the whole page.
+func TestLogin_FormErrorAnswersItsStatusWithThePage(t *testing.T) {
+	t.Parallel()
+
+	var h *handlers.Handlers
+
+	app := newTestApp(t, &h)
+	app.RequireStart()
+
+	t.Cleanup(app.RequireStop)
+
+	w := submitLogin(h, sharedProxyPeer, "", "")
+
+	assert.Equal(t, http.StatusBadRequest, w.Code)
+	assert.Contains(
+		t, w.Body.String(), "Username and password are required",
+	)
+	assert.Contains(
+		t, w.Body.String(), "</html>",
+		"the page must render to completion",
+	)
+}
+
+// TestLogin_FormErrorRenderFailureAnswers500 proves that a login form
+// error page whose template fails answers 500 with the error page and
+// none of the form page, rather than the 400 it meant to send.
+func TestLogin_FormErrorRenderFailureAnswers500(t *testing.T) {
+	t.Parallel()
+
+	var h *handlers.Handlers
+
+	app := newTestApp(t, &h)
+	app.RequireStart()
+
+	t.Cleanup(app.RequireStop)
+
+	// The page prints its error message and then fails.
+	h.AddTemplateForTest("login.html", template.Must(
+		template.New("login").Funcs(template.FuncMap{
+			"fail": func() (string, error) { return "", errMidRender },
+		}).Parse(`{{.Error}}{{fail}}`),
+	))
+
+	w := submitLogin(h, sharedProxyPeer, "", "")
+
+	assert.Equal(t, http.StatusInternalServerError, w.Code)
+	assert.NotContains(
+		t, w.Body.String(), "Username and password are required",
+		"the response must carry no part of the aborted page",
+	)
+	assert.Contains(t, w.Body.String(), "500 Internal Server Error")
+}
+
 // TestLogin_SuccessCreatesSession is the control for the tests above:
 // the success path they assert on really does authenticate.
 func TestLogin_SuccessCreatesSession(t *testing.T) {

internal/handlers/handlers.go

@@ -309,12 +309,26 @@ func (s *Handlers) getUserInfo(
 }
 
 // renderTemplate renders a pre-parsed template with common
-// data
+// data and answers 200.
 func (s *Handlers) renderTemplate(
 	w http.ResponseWriter,
 	r *http.Request,
 	pageTemplate string,
 	data any,
+) {
+	s.renderTemplateStatus(w, r, pageTemplate, data, http.StatusOK)
+}
+
+// renderTemplateStatus is renderTemplate answering with status, for a
+// form shown again with an error. Call it instead of WriteHeader
+// followed by renderTemplate: the status is written only once the page
+// has rendered, so a failed render can still answer 500.
+func (s *Handlers) renderTemplateStatus(
+	w http.ResponseWriter,
+	r *http.Request,
+	pageTemplate string,
+	data any,
+	status int,
 ) {
 	tmpl, ok := s.templates[pageTemplate]
 	if !ok {
@@ -327,7 +341,9 @@ func (s *Handlers) renderTemplate(
 		return
 	}
 
-	s.executeTemplate(w, r, tmpl, s.pageData(r, data, noticeFor(r)))
+	s.executeTemplate(
+		w, r, tmpl, s.pageData(r, data, noticeFor(r)), status,
+	)
 }
 
 // pageData adds the fields the shared layout renders to a page's own
@@ -363,19 +379,20 @@ func (s *Handlers) pageData(
 	}
 }
 
-// executeTemplate renders the template into a buffer and writes to
-// the response only once rendering has fully succeeded. Executing
-// straight into the ResponseWriter commits a partial body and a 200
-// status before a mid-render error can be reported, leaving no way
-// to serve a 500. Buffering makes a page's rendered size resident
-// memory per concurrent viewer, so every page owes it a bound: the
-// event log caps each stored body at maxRenderedBodyBytes for exactly
-// this reason.
+// executeTemplate renders the template into a buffer and writes status
+// and the page to the response only once rendering has fully
+// succeeded. Executing straight into the ResponseWriter commits a
+// partial body and the status before a mid-render error can be
+// reported, leaving no way to serve a 500. Buffering makes a page's
+// rendered size resident memory per concurrent viewer, so every page
+// owes it a bound: the event log caps each stored body at
+// maxRenderedBodyBytes for exactly this reason.
 func (s *Handlers) executeTemplate(
 	w http.ResponseWriter,
 	r *http.Request,
 	tmpl *template.Template,
 	data any,
+	status int,
 ) {
 	var buf bytes.Buffer
 
@@ -390,6 +407,7 @@ func (s *Handlers) executeTemplate(
 	}
 
 	w.Header().Set("Content-Type", "text/html; charset=utf-8")
+	w.WriteHeader(status)
 
 	_, err = buf.WriteTo(w)
 	if err != nil {

internal/handlers/handlers_test.go

@@ -181,6 +181,11 @@ func (r *recordingArchives) Renames() []archiveRename {
 	return out
 }
 
+// newTestApp returns an app whose RequireStart fails the test when
+// starting takes longer than fx's default start timeout of 15s. That
+// limit catches a start that hangs, not a busy host: measured with make
+// test on 2026-10-02 at host load 58-69 on 48 cores, the slowest of this
+// package's starts took 0.49s.
 func newTestApp(
 	t *testing.T,
 	targets ...any,

internal/handlers/source_management.go

@@ -350,12 +350,12 @@ func (h *Handlers) HandleSourceCreateSubmit() http.HandlerFunc {
 		retentionStr := r.PostFormValue("retention_days")
 
 		if name == "" {
-			w.WriteHeader(http.StatusBadRequest)
-			h.renderTemplate(
+			h.renderTemplateStatus(
 				w, r, "sources_new.html",
 				newSourceFormData(
 					"Name is required", name, description,
 				),
+				http.StatusBadRequest,
 			)
 
 			return
@@ -365,13 +365,13 @@ func (h *Handlers) HandleSourceCreateSubmit() http.HandlerFunc {
 			retentionStr, database.DefaultRetentionDays,
 		)
 		if retErr != nil {
-			w.WriteHeader(http.StatusBadRequest)
-			h.renderTemplate(
+			h.renderTemplateStatus(
 				w, r, "sources_new.html",
 				newSourceFormData(
 					retentionErrorMessage(retErr),
 					name, description,
 				),
+				http.StatusBadRequest,
 			)
 
 			return
@@ -644,8 +644,7 @@ func (h *Handlers) applyWebhookEdit(
 			tmplKeyError:   "Name is required",
 		}
 
-		w.WriteHeader(http.StatusBadRequest)
-		h.renderTemplate(w, r, "source_edit.html", data)
+		h.renderTemplateStatus(w, r, "source_edit.html", data, http.StatusBadRequest)
 
 		return
 	}
@@ -665,8 +664,7 @@ func (h *Handlers) applyWebhookEdit(
 			tmplKeyError:   retentionErrorMessage(retErr),
 		}
 
-		w.WriteHeader(http.StatusBadRequest)
-		h.renderTemplate(w, r, "source_edit.html", data)
+		h.renderTemplateStatus(w, r, "source_edit.html", data, http.StatusBadRequest)
 
 		return
 	}
@@ -703,8 +701,7 @@ func (h *Handlers) applyWebhookEdit(
 					"it, then save again.",
 			}
 
-			w.WriteHeader(http.StatusConflict)
-			h.renderTemplate(w, r, "source_edit.html", data)
+			h.renderTemplateStatus(w, r, "source_edit.html", data, http.StatusConflict)
 
 			return
 		}

script/test

@@ -2,9 +2,10 @@
 # script/test: run the test suite.
 #
 # -timeout is applied by `go test` per package, not to the run as a whole, so
-# it only has to clear the slowest single package. That is internal/handlers,
-# measured in a cache-defeated builder stage on the 48-core shared build host
-# (2026-08-18); load- and host-dependent, not invariants:
+# it only has to clear the slowest single package. When this budget was set
+# that was internal/handlers, measured in a cache-defeated builder stage on the
+# 48-core shared build host (2026-08-18); load- and host-dependent, not
+# invariants:
 #
 #   16.9s                  host load 5-20, GOMAXPROCS 48
 #   45.9s / 47.3s / 49.0s  three runs at deliberate host load 31-73
@@ -23,6 +24,11 @@
 # a condition CI runs under. If a CPU-limited runner ever puts a real run near
 # 67s, that is the datum to revisit the org figure with.
 #
+# Those figures predate tests hashing the admin password at 1 MB instead of
+# 64 MB (https://git.eeqj.de/sneak/webhooker/pulls/404). After that change, in
+# a cache-defeated build at host load 44-109 (2026-10-02), internal/handlers
+# took 8.5s and the slowest package was internal/database at 15.8s.
+#
 # -p 4 -parallel 8 keep the run under 2 GB of memory: at most four test
 # binaries build or run at once, each with at most eight parallel tests. Under
 # -race every test binary and every link costs a few hundred MB, so the