fix: Align xrpld RPM packaging with DEB package by legleux · Pull Request #7529 · XRPLF/rippled

legleux · 2026-06-10T21:41:06Z

The DEB package already skips the dwz debug optimization step because it takes a long time. This change brings RPM package validation into the same rough size/time class as Debian package generation.

Changes:

Disable RPM dwz debuginfo optimization.
Write generated RPM payloads uncompressed, intentionally trading larger RPM artifacts for much faster package build/validation time.
Clarify package version/release naming so the raw xrpld software version (XRPLD_VERSION) stays distinct from the normalized package metadata version (pkg_version).
Align RPM upgrade behavior with DEB by avoiding an automatic xrpld service restart on package upgrade; operators pick up the new binary on the next service restart.

For context from CI validation:

old RPM: ~370 MB, ~12 min
Debian: ~566 MB, <3 min
new RPM: ~607 MB, <2 min

The old RPM artifact was smaller, but it spent about 10 extra minutes optimizing/compressing debug-package payloads for a short-lived CI artifact. Debian is already the practical comparison point here: DWZ optimization is not viable there for xrpld today (Too many DIEs, not optimizing), and the Debian debug package still builds quickly.

Operationally, debug-symbol packages are mostly useful when analyzing coredumps. The download data supports optimizing this path for CI latency rather than RPM debug artifact size. For 3.1.3, the stats were:

package type	release downloads	debug-symbol downloads
deb	2,610	61
rpm	383	2

Some of the Debian debug downloads were us too.

github-actions · 2026-06-11T14:18:37Z

This PR has conflicts, please resolve them in order for the PR to be reviewed.

github-actions · 2026-06-11T17:30:18Z

All conflicts have been resolved. Assigned reviewers can now start or resume their review.

codecov · 2026-06-11T18:25:15Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 81.9%. Comparing base (7e0ff53) to head (034e465).
⚠️ Report is 5 commits behind head on develop.

Additional details and impacted files

@@            Coverage Diff            @@
##           develop   #7529     +/-   ##
=========================================
- Coverage     82.0%   81.9%   -0.0%     
=========================================
  Files         1007    1007             
  Lines        76876   76854     -22     
  Branches      8982    8984      +2     
=========================================
- Hits         63003   62979     -24     
- Misses       13864   13866      +2     
  Partials         9       9

see 12 files with indirect coverage changes

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

xrplf-ai-reviewer

No issues.

Review by Claude Sonnet 4.6 · Prompt: V15

mathbunnyru

Approved, if it will take less time in CI (please, check, when it's done)

mathbunnyru

Unfortunately, doesn't work:
PR / package / xrpld-rhel-gcc-release-amd64 (pull_request)Successful in 12m

legleux · 2026-06-11T22:08:39Z

Unfortunately, doesn't work: PR / package / xrpld-rhel-gcc-release-amd64 (pull_request)Successful in 12m

I saw that. I removed the knobs but I removed the thing that made it work also. Re-running.

xrplf-ai-reviewer

macOS/Windows CI coverage dropped — see inline.

Review by Claude Sonnet 4.6 · Prompt: V15

xrplf-ai-reviewer

Took a pass through this

Unrelated scope creep flagged inline — macOS/Windows CI removal is bundled into an RPM debuginfo-only change.

Review by ReviewBot 🤖

Review by Claude Sonnet 4.6 · Prompt: V15

xrplf-ai-reviewer

Had a read through

Uncompressed payload flag applied unconditionally — affects shipped package size. See inline.

Review by ReviewBot 🤖

Review by Claude Sonnet 4.6 · Prompt: V15

legleux · 2026-06-12T00:40:26Z

This brings RPM package validation into the same rough size/time class as Debian.

For context:

old RPM: ~370 MB, ~12 min
Debian: ~566 MB, <3 min
new RPM: ~607 MB, <2 min

The old RPM artifact was smaller, but it got there by spending about 10 extra minutes optimizing/compressing debug-package payloads for a disposable CI artifact. Debian is already the practical comparison point here: DWZ optimization is not viable there for xrpld today (Too many DIEs, not optimizing), and the Debian debug package still builds quickly at ~566 MB.

Operationally, the debug-symbol packages are mostly useful when we need to analyze coredumps. The download data supports optimizing this path for CI latency rather than RPM debug artifact size.
For 3.1.3, the stats were:

package type	release downloads	debug-symbol downloads
deb	2,610	61
rpm	383	2

Some of the debian debug downloads were us also.

So for PR validation, I think it is reasonable to optimize for fast proof that the package builds rather than spending another ~10 minutes minimizing a short-lived RPM debug artifact that is rarely consumed.

legleux · 2026-06-12T04:07:12Z

Almost as fast as deb now

rpm 1m 54s
deb 1m 47s

Copilot

Pull request overview

Copilot reviewed 6 out of 6 changed files in this pull request and generated 4 comments.

 Name:     xrpld
-Version:  %{xrpld_version}
-Release:  %{xrpld_release}%{?dist}
+Version:  %{pkg_version}
+Release:  %{pkg_release}%{?dist}


xrplf-ai-reviewer

No issues.

Review by Claude Sonnet 4.6 · Prompt: V15

mathbunnyru · 2026-06-18T19:00:05Z

+if [[ "${XRPLD_VERSION}" == -* || "${XRPLD_VERSION}" == *- ]]; then
+    echo "build_pkg.sh: invalid XRPLD_VERSION='${XRPLD_VERSION}'." >&2
+    echo "Version cannot start or end with '-'." >&2
+    exit 1
+fi


Why do we check these specific conditions?

Good catch. If we got here we've gotten past BuildInfo's constraints, shouldn't be possible by the time packaging sees it.

mathbunnyru · 2026-06-18T19:02:07Z

+    # changelog's third field "distribution", but the suite/codename (noble,
+    # bookworm, ...) is assigned by the publishing infra at ingest; this


Just in case - we don't and won't build for different codenames.
It's because we build with a custom century-old libc

Yep, this was a wording issue. I removed the implication that packaging is selecting or supporting Debian/Ubuntu codenames. The changelog distribution value is just the XRPLF release channel we publish to (stable, develop, unstable), not the build host codename.

mathbunnyru · 2026-06-18T19:02:28Z

+    # changelog's third field "distribution", but the suite/codename (noble,
+    # bookworm, ...) is assigned by the publishing infra at ingest; this
+    # suite-agnostic build only sets the component. RPM has no equivalent.
+    #   3.2.0 -> stable, *-b0[+metadata] -> develop, other pre-release -> unstable.


Maybe let's mention RCs here as well?
They are far more important than betas

Added RCs explicitly. The docs/comment now call out that non-b0 prereleases, including beta and RC builds, go to unstable.

mathbunnyru · 2026-06-18T19:04:11Z

    xrpld.sysusers      sysusers.d config (used by both RPM and DEB)
    xrpld.tmpfiles      tmpfiles.d config (used by both RPM and DEB)
    xrpld.logrotate     logrotate config (installed to /etc/logrotate.d/xrpld)
+    50-xrpld.preset     systemd preset (RPM only; DEB enables the unit via dh_installsystemd)


Should be in rpm dir if it's only used in rpm

Let's just remove it altogether since it's one line!

mathbunnyru · 2026-06-18T19:04:52Z

+| Package type | Image (`package_configs.<distro>[].image` in `linux.json`) | Tools required                                 |
+| ------------ | ---------------------------------------------------------- | ---------------------------------------------- |
+| RPM          | `ghcr.io/xrplf/xrpld/packaging-rhel:sha-<sha>`             | `rpmbuild`                                     |
+| DEB          | `ghcr.io/xrplf/xrpld/packaging-debian:sha-<sha>`           | `dpkg-buildpackage`, `debhelper-compat (= 13)` |


Probably anything >= 13 would work.

Changed the wording.
Probably this CMake could work right? 😸

set(CMAKE_CXX_STANDARD 20) set(CMAKE_CXX_STANDARD_REQUIRED OFF)

mathbunnyru · 2026-06-18T19:07:51Z

+The version is not a CMake input on this path: `cmake/XrplVersion.cmake` reads
+it from `src/libxrpl/protocol/BuildInfo.cpp` into `xrpld_version`, and


You have more than 1 way of determining xrpld version.
Can we leave just one, please?
I think ./xrpld --version with a bit of parsing should be fine - you have the binary in places where you need it.

Love it.
Packaging derives the version from the binary, just fail early if ${BUILD_DIR}/xrpld is missing or not executable, before trying to parse version. That makes the "you must already have a binary" contract obvious.

mathbunnyru · 2026-06-18T19:53:07Z

I think this stopped being trivial, to be honest, so requires 2 approvals, and it's not ready to merge yet

xrplf-ai-reviewer

No issues.

Review by Claude Sonnet 4.6 · Prompt: V15

xrplf-ai-reviewer

No issues.

Review by Claude Sonnet 4.6 · Prompt: V15

xrplf-ai-reviewer

No issues.

Review by Claude Sonnet 4.6 · Prompt: V15

xrplf-ai-reviewer

No issues.

Review by Claude Sonnet 4.6 · Prompt: V15

xrplf-ai-reviewer

No issues.

Review by Claude Sonnet 4.6 · Prompt: V15

Copilot

Pull request overview

Copilot reviewed 7 out of 7 changed files in this pull request and generated 1 comment.

xrplf-ai-reviewer

No issues.

Review by Claude Sonnet 4.6 · Prompt: V15

Copilot

Pull request overview

Copilot reviewed 7 out of 7 changed files in this pull request and generated 1 comment.

legleux · 2026-06-18T21:38:02Z

I think this stopped being trivial, to be honest, so requires 2 approvals, and it's not ready to merge yet

I think so too. 🤕

legleux added the DraftRunCI Normally CI does not run on draft PRs. This opts in. label Jun 10, 2026

legleux force-pushed the legleux/rpm-compression branch from cfc80fb to 497fb2a Compare June 10, 2026 21:42

github-actions Bot added the PR: has conflicts label Jun 11, 2026

legleux force-pushed the legleux/rpm-compression branch from 497fb2a to a61534e Compare June 11, 2026 17:30

github-actions Bot removed the PR: has conflicts label Jun 11, 2026

legleux changed the title ~~ci: speed up disposable package checks~~ ci: speed up rpm debuginfo package generation Jun 11, 2026

legleux changed the title ~~ci: speed up rpm debuginfo package generation~~ ci: Speed up rpm debuginfo package generation Jun 11, 2026

legleux marked this pull request as ready for review June 11, 2026 19:47

legleux requested review from bthomee and mathbunnyru June 11, 2026 19:47

xrplf-ai-reviewer Bot approved these changes Jun 11, 2026

View reviewed changes

mathbunnyru approved these changes Jun 11, 2026

View reviewed changes

mathbunnyru requested changes Jun 11, 2026

View reviewed changes

legleux force-pushed the legleux/rpm-compression branch from a61534e to d1fd10e Compare June 11, 2026 22:07

legleux removed the DraftRunCI Normally CI does not run on draft PRs. This opts in. label Jun 11, 2026

xrplf-ai-reviewer Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread .github/workflows/on-pr.yml Outdated

legleux force-pushed the legleux/rpm-compression branch from d1fd10e to 3961412 Compare June 11, 2026 22:33

xrplf-ai-reviewer Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread .github/workflows/on-pr.yml Outdated

legleux force-pushed the legleux/rpm-compression branch from 3961412 to 7740ee6 Compare June 11, 2026 23:31

xrplf-ai-reviewer Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread package/rpm/xrpld.spec

legleux added the Trivial Simple change with minimal effect, or already tested. Only needs one approval. label Jun 12, 2026

legleux requested a review from mathbunnyru June 12, 2026 04:07

bthomee requested review from Copilot and removed request for bthomee June 12, 2026 11:10

Copilot AI reviewed Jun 18, 2026

View reviewed changes

bthomee reviewed Jun 18, 2026

View reviewed changes

Address RPM packaging review feedback

be0af3c

xrplf-ai-reviewer Bot approved these changes Jun 18, 2026

View reviewed changes

legleux added the Ready to merge *PR author* thinks it's ready to merge. Has passed code review. Perf sign-off may still be required. label Jun 18, 2026

bthomee approved these changes Jun 18, 2026

View reviewed changes

mathbunnyru requested changes Jun 18, 2026

View reviewed changes

mathbunnyru removed Trivial Simple change with minimal effect, or already tested. Only needs one approval. Ready to merge *PR author* thinks it's ready to merge. Has passed code review. Perf sign-off may still be required. labels Jun 18, 2026

Address packaging review feedback

45f98cc

xrplf-ai-reviewer Bot approved these changes Jun 18, 2026

View reviewed changes

legleux added 2 commits June 18, 2026 13:54

Clarify Debian package channel wording

a40abb8

Use Debian component terminology

e3bb424

xrplf-ai-reviewer Bot approved these changes Jun 18, 2026

View reviewed changes

Restrict Debian component prerelease mapping

ff44fcd

xrplf-ai-reviewer Bot approved these changes Jun 18, 2026

View reviewed changes

Reject metadata-only package versions

23a212d

legleux requested a review from Copilot June 18, 2026 21:08

Copilot started reviewing on behalf of legleux June 18, 2026 21:09 View session

xrplf-ai-reviewer Bot approved these changes Jun 18, 2026

View reviewed changes

Copilot AI reviewed Jun 18, 2026

View reviewed changes

Comment thread package/build_pkg.sh

Standardize package version error

034e465

legleux requested a review from Copilot June 18, 2026 21:21

Copilot started reviewing on behalf of legleux June 18, 2026 21:22 View session

xrplf-ai-reviewer Bot approved these changes Jun 18, 2026

View reviewed changes

Copilot AI reviewed Jun 18, 2026

View reviewed changes

Comment thread package/build_pkg.sh

legleux requested a review from mathbunnyru June 18, 2026 22:26

		# changelog's third field "distribution", but the suite/codename (noble,
		# bookworm, ...) is assigned by the publishing infra at ingest; this

		The version is not a CMake input on this path: `cmake/XrplVersion.cmake` reads
		it from `src/libxrpl/protocol/BuildInfo.cpp` into `xrpld_version`, and

Conversation

legleux commented Jun 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions Bot commented Jun 11, 2026

Uh oh!

github-actions Bot commented Jun 11, 2026

Uh oh!

codecov Bot commented Jun 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

xrplf-ai-reviewer Bot left a comment

Choose a reason for hiding this comment

Uh oh!

mathbunnyru left a comment

Choose a reason for hiding this comment

Uh oh!

mathbunnyru left a comment

Choose a reason for hiding this comment

Uh oh!

legleux commented Jun 11, 2026

Uh oh!

xrplf-ai-reviewer Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

xrplf-ai-reviewer Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

xrplf-ai-reviewer Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

legleux commented Jun 12, 2026

Uh oh!

legleux commented Jun 12, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

xrplf-ai-reviewer Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

legleux Jun 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

legleux commented Jun 10, 2026 •

edited

Loading

codecov Bot commented Jun 11, 2026 •

edited

Loading

legleux Jun 18, 2026 •

edited

Loading

legleux Jun 18, 2026 •

edited

Loading