Client Code

Client Code describes the technology, size and development history of Rise-Client, the native Windows client for Rise: The Vieneo Province. The client implements atmospheric and spaceflight, ground vehicles, cockpit instrumentation, flight guidance and warnings, terrain and weather rendering, positional audio, networking, a custom GUI, and distribution-platform integrations.

The project began around 2001, according to its principal developer's recollection. The repository preserves approximately 9 years and 8 months of revision history, beginning in January 2017. The project's age, the history retained in Git, and the age of its GitHub hosting are separate measurements.

This analysis is a historical snapshot measured on 14 October 2026. Counts refer to committed source, not uncommitted work or later revisions.

Item	Audited value
Repository	Rise-Client (private; access required)
Active development branch	develop
Audited commit	4e95a9393f4f93eff326476279c153e9fbb9575c
Commit date	13 October 2026, 22:28:09 (America/Chicago)
Application and support code	95,527 code lines
Tests, test tooling and reference implementations	34,044 code lines
Combined project-code subtotal	129,571 code lines across 530 source files
Commits in develop ancestry	4,302
Merged GitHub pull requests	51
Identifiable human contributors in preserved author metadata	2, after consolidating apparent aliases
GUI language files	12

The project-code subtotal excludes comments, blank lines, identified third-party or derived source, generated code, and auxiliary scripts. It includes tests. The counting method below defines these categories.

Technology and scope

• Language and platform: C++17, native Win32, Visual Studio and MSBuild; the audited project uses toolset v145.
• Rendering: Direct3D 11, DXGI and HLSL, with native Windows rendering and input.
• Networking: RakNet.
• Audio: BASS and BASS_FX.
• Platform integrations: Steamworks and the Discord SDK.
• Supporting tools: Python, PowerShell, batch scripts and Windows resource files.
• Separate components: The C# server is maintained outside this repository and is excluded from these counts. The solution also references RiseLauncher, and the build references sibling dependencies such as BASS and RakNet.

The production Win32 migration to native Direct3D 11 is recorded in commit 6b5d818c187c8b3e8147fc1043cf8c52954b50ec, dated 30 September 2026. Legacy Direct3D 9 comparison implementations remain under Tests/Reference9.

The repository contains GUI language files for Chinese, Dutch, English, French, German, Greek, Italian, Japanese, Korean, Portuguese, Russian and Spanish. This analysis counted those assets; it did not validate translation completeness.

Source size

Category	Files	Code lines	Physical lines
Application and support tools	345	95,527	114,320
Tests and test tools	175	32,403	34,722
Legacy rendering test references	10	1,641	1,774
Identified third-party or derived source	60	20,998	34,073
Generated source	6	3,681	4,197
Auxiliary scripts and SQL	5	563	694
All selected source files	601	154,813	189,780

Code lines are nonblank, non-comment-only physical source lines as classified by cloc. Physical lines include code, comments and blank lines. Neither measurement is a count of statements or a measure of algorithmic complexity.

A useful rounded description is approximately 130,000 code lines including tests, or approximately 95,500 application and support code lines excluding tests. The exact primary project subtotal is 129,571 code lines and 150,816 physical lines across 530 files.

Tests, test tools and rendering references account for 26.3% of that code subtotal. This is a proportion of source size, not a test-coverage percentage.

Language breakdown

The following breakdown covers the primary project subtotal, including tests but excluding the third-party/derived, generated and auxiliary categories.

Language or source type	Code lines
C++ implementations	97,734
C/C++ headers	25,414
HLSL	3,485
Python	2,345
PowerShell	504
Windows resource files	69
Batch scripts	20
Total	129,571

C++ implementations and C/C++ headers account for approximately 95% of this subtotal.

Media and tracked files

The committed tree contains 4,798 files totaling 2,428,643,862 bytes (2.43 decimal GB, or 2.26 GiB). This measures the files in the audited tree, not compressed Git history or the size of a complete working installation.

Directory	Tracked files
Textures	2,265
Sounds	1,470
Meshes	190

Textures and sounds account for 97.9% of the tracked bytes. These directory inventories include unused or original assets and support files; they are not counts of distinct assets used in a production session. Repository size is therefore dominated by media and is a poor proxy for code size.

Revision history

Scope	Distinct commits	Notes
Active develop ancestry	4,302	Includes 189 merge commits
Union of locally known origin branch ancestries	4,386	GitHub branch tips were checked and matched these remote-tracking refs
All locally referenced history	4,396	Includes 10 commits not reachable from those origin refs, and 222 merge commits overall
GitHub default branch, master	2,540	Last commit dated 15 June 2025 in America/Chicago

These are distinct scopes and must not be added together. Distinct commit objects are also not necessarily distinct features: merges, parallel development and cherry-picks affect the history.

GitHub records 52 pull requests: 51 merged, 1 closed without merging, and 0 open at the time of the audit. Git merge commits and merged GitHub pull requests are different measurements; the pull-request count was retrieved independently from GitHub.

The earliest preserved commit is 50ff4fdaedd5f3123e7da32daf4d160a067e26e4, dated 27 January 2017, with the message Initial versioning. The inspected checkout is not shallow and has one reachable root. The latest preserved author date is 13 October 2026.

Across all locally referenced history, commits occur on 1,239 distinct author-date calendar days and in 104 distinct months. These are measurements of recorded commit activity, not days or hours worked.

Current development branch

At the audit date, develop contains the current implementation. GitHub's default master branch stops in June 2025 and omits much of the recent development. Anyone evaluating the current client should inspect develop.

The total commit count on develop exceeds the total on master by 1,762. This is a subtraction of ancestry counts, not a calculation of commits unique to either branch.

Contributors and attribution

The preserved history contains 15 distinct raw author name/email pairs. Consolidating apparent aliases identifies two human contributors, Jason Reskin and Martin Elberts, plus a Codex identity.

Inferred author grouping	All locally referenced history	develop ancestry
Jason Reskin and apparent aliases	4,339	4,268
Martin Elberts and apparent aliases	55	32
Codex identity	2	2
Total	4,396	4,302

Five coauthor trailers also mention Codex. Commit attribution does not quantify AI involvement or the amount of code written by any person.

Alias consolidation is inferred from names and shared email identities. It is not an independently verified lifetime contributor census: contributors before the preserved history, and non-code collaborators, may be absent. The metadata nevertheless shows a project overwhelmingly led by one principal maintainer.

Annual activity and source growth

Year	Commits in develop ancestry	Application/support code lines	Tests/tools/reference code lines
2017	99	38,118	0
2018	568	43,156	0
2019	968	49,104	0
2020	1,037	56,331	0
2021	285	58,117	0
2022	14	58,383	0
2023	284	62,167	0
2024	171	64,201	0
2025	340	65,386	0
2026, through 13 October	536	95,527	34,044

Source counts are year-end first-parent snapshots, except that 2026 uses the audited head. Commit counts use author-date years across the full develop ancestry, including merges. A zero in the test column means that no source was present in the classified Tests directory at that snapshot; it is not proof that no testing existed elsewhere in the project.

Application/support code is now 2.43 times the size of the first preserved revision, which contained 39,328 code lines. It grew 46.1% between the end of 2025 and the audited 2026 snapshot. These measurements describe source size, not developer productivity or software quality.

Earlier source-control history

Period or date	History	Evidence
Around 2001–2003	Early development without source control	Principal developer's recollection
Around 2003–2007	SourceGear Vault	Principal developer's recollection; the current solution retains a VaultVsipSolution-v2 section
From around 2007	Mercurial	Principal developer's recollection
27 January 2017	First revision preserved in the current Git repository	Commit 50ff4fda; the tree already contains .hgignore and a substantial client
5 June 2020	GitHub repository created	GitHub repository metadata
8 June 2020	.hgignore still being modified	Commit c2baeac82
27 June 2020	.hgignore deleted	Commit db92ca610

The evidence is consistent with imported Mercurial history and a transition around the 2020 GitHub repository creation. The exact date of the project's first use of Git is not established by this copy; a tentative recollection of a 2012 switch could not be verified.

Git author and committer timestamps can survive conversion from another version-control system. They therefore do not establish when a revision first entered Git. The project has more than 20 years of development history, but this repository does not retain a complete 20-year revision history.

Notable characteristics

A mature starting snapshot

The first preserved revision already contains 39,328 application/support code lines in 97 files, plus a generated header, and 3,490 tracked files overall. A substantial part of the product's development predates the visible history.

Recent regression infrastructure

The audited tree has 85 Visual C++ projects under Tests, including harnesses and benchmarks, and 84 C++ files named *Tests.cpp. The earliest addition under Tests in the preserved history is dated 11 August 2026. Earlier year-end snapshots contain no Tests directory.

Established simulation behavior therefore coexists with recently added regression tools and rendering comparisons. Project counts do not establish test coverage or indicate that every project is a standalone test suite. This analysis did not run those tests.

Concentration of application code

The ten largest application/support source files contain 38,158 code lines, approximately 40% of that category.

File	Code lines
Viewscreen/Viewscreen.cpp	5,737
gui.cpp	4,635
Sound.cpp	4,576
Viewscreen/cockpit.cpp	4,136
Networking.cpp	3,752
Viewscreen/propulsion.cpp	3,701
Viewscreen/framemove.cpp	3,524
Viewscreen/renderer.cpp	3,129
Viewscreen/buildings.cpp	2,632
Viewscreen/render.cpp	2,336

Large-file concentration identifies useful starting points for architectural investigation. It is not a defect score.

Frequently changed paths

Current source path	Non-merge commits touching the path in develop
Viewscreen/Viewscreen.cpp	789
Networking.cpp	736
Viewscreen/propulsion.cpp	624
Viewscreen/framemove.cpp	550
GameClass.cpp	538

These counts track exact paths without following renames. They highlight persistent simulation and integration work, not necessarily defects.

Evaluation context

The repository offers specialized, evolving C++ behavior and interactions among rendering, simulation, audio, networking and UI systems. Its relatively small pull-request count compared with its commit count means that PR counts alone understate its development history.

A long-lived product, long preserved revision history, many independent contributors, and a large PR corpus are different evaluation criteria. This repository provides different amounts of evidence for each.

The client is also part of a larger system. A reproducible evaluation environment would need the sibling dependencies and intended test subset identified. The audit did not attempt a clean-room build.

Counting method and limitations

Snapshot and tool

Counts were taken from Git objects at the immutable audited commit, rather than from the modified working tree. Untracked files, build outputs and uncommitted changes were excluded. No application source was changed by the audit, and neither Debug nor Release builds were run for this read-only analysis.

cloc 2.06 was run with --by-file --json --unicode --skip-uniqueness. HLSL was explicitly assigned to .fx, .fxh and .hlsl files. Duplicate files were counted separately when they were separate tracked files.

The selected source extensions were .cpp, .c, .h, .hpp, .hh, .cc, .cxx, .inl, .hlsl, .fx, .fxh, .py, .ps1, .cmd, .bat, .js, .ts, .cs, .lua, .rc and .sql. XML/project/configuration files, documentation, translations and media were not included in code totals.

Classification

• Third-party or derived source: ThirdParty/*, plus UI/Controls.cpp and UI/Controls.h, whose Microsoft notices identify their DXUT origin.
• Generated source: filenames containing .generated., source bearing the Tests/PortShaders11.py generated marker, and the generated resource.h header.
• Test references: source under Tests/Reference9, reported separately within the overall test subtotal.
• Tests and test tools: remaining selected source under Tests.
• Auxiliary scripts and SQL: selected source under Sounds and docs.
• Application and support tools: remaining selected source.

This is a conservative inventory classification, not a legal provenance audit. Adapted snippets may remain within project source.

Historical comparisons

Historical snapshots use the same classification rules. Each year-end sample is the latest first-parent commit before the local year-end cutoff. History statistics use author-date calendar years and include merges unless explicitly excluded.

Local refs can contain parallel or cherry-picked work. Commit attribution cannot measure individual effort, productivity, or the extent of AI assistance. Line counts do not measure test coverage, correctness, complexity or runtime performance.

Sources and reproducibility

The analysis combines Git tree and history inspection, cloc output, GitHub repository and pull-request metadata, and the principal developer's recollection of early development. GitHub links below require repository access.

• Audited source snapshot
• First preserved revision
• Native Direct3D 11 migration
• Build configuration
• Rendering migration documentation

The audit retained its counting script, pinned cloc source, per-file results, historical snapshots, annual summaries and path-change counts as local reproducibility artifacts. These artifacts are not attachments to this wiki page.