Repository navigation
Add the start menu mod - #7573
babamohammed2022 wants to merge 4 commits into
Conversation
|
Thanks for the pull request! This repository uses a two-stage review: an AI review that you run yourself, followed by a human review. To get started, comment See the pull request review process for the full details. |
|
/ai-review |
Submission reviewNote: This review was done by Claude. Due to the amount of submissions, doing a fully manual review for each pull request is no longer feasible. Thank you for understanding. Remember: The AI reviewer can be wrong - it may misread code, flag correct code as broken, or suggest changes that make things worse. Treat its findings as suggestions to verify, not instructions to follow blindly. You're responsible for the code you submit, so if a finding doesn't hold up, say so instead of changing working code to satisfy it. Please address the following issues. The items in the collapsed sections are optional, so it's your call whether to address them. A few things to address before this is ready: 1. Substantial overlap with your own
About 3,900 of this file's 6,500 lines longer than 25 characters (~60%) appear verbatim in the 1507 mod. The differences:
The two mods can't run together (README: "Disable the regular Start menu mod and other Start replacements"). Neither the PR nor the README says why this is a separate mod. Two ways forward:
2. Mod storage can't hold some of what the mod writes. Portable Windhawk keeps mod storage in an INI file; a regular install keeps it in the registry.
3. Heap and stack overflow when converting search-result icons (4210–4216). For icons without alpha, the AND mask is read with a 1-bpp
This runs on the icon worker for file search results, so it can corrupt memory or crash the helper. 4. An invalid
Fall back to the stored layout when the import fails: bool ok = layout_.Deserialize(settings_.tileLayout, catalog_);
if (!ok && settings_.tileLayoutImportPending && !layoutApplied_)
ok = layout_.Deserialize(ReadStorageString(L"startMenuLayout"), catalog_);5. README: development notes, the helper process isn't explained, no screenshot.
Optional improvements
Minor polish — none of this affects users, so it's your call.
Functionality notes
Non-critical observations and ideas about the feature behavior itself.
Code overview
For the maintainer: a map of the mod and what it touches outside its own process. Descriptive only — anything that needs a change is listed above. Line numbers refer to the mod source at commit The mod runs in two roles. In the shell Code regions
Side effects of interest
Next steps:
See the review process for details. Important: Each reply that was generated by AI must be disclosed by including the following text in the reply: "Generated by AI. Model: (ai-model-name)". |
|
Thank you for the review. Regarding the overlap with Windows 10 1507 Start Menu Recreation, I would prefer to keep this as a separate mod. The two menus have a similar underlying structure and interaction model, so reusing substantial parts of the implementation is intentional. However, this mod recreates the Windows 8.1 Update 3 Metro appearance, which is visually distinct from the Windows 10 1507 menu. Combining both into one mod could make its purpose and settings confusing for users looking for one specific design. This mod also has its own search, cache, and helper-process implementation. Some of those improvements could be considered for the 1507 mod separately, but they are not essential to choosing its visual style. I understand that the amount of shared code is a maintenance concern, and I’m open to the maintainer’s decision on whether this justification is sufficient. I will address the other review points separately |
Updated documentation and settings descriptions for clarity and consistency. Adjusted layout handling and storage mechanisms for tile layouts.
|
/ai-review |
Submission reviewNote: This review was done by Claude. Due to the amount of submissions, doing a fully manual review for each pull request is no longer feasible. Thank you for understanding. Remember: The AI reviewer can be wrong - it may misread code, flag correct code as broken, or suggest changes that make things worse. Treat its findings as suggestions to verify, not instructions to follow blindly. You're responsible for the code you submit, so if a finding doesn't hold up, say so instead of changing working code to satisfy it. Please address the following issues. The items in the collapsed sections are optional, so it's your call whether to address them. Thanks for the fixes. These all check out:
Two things remain: 1. Substantial overlap with your own
About 60% of the file's lines longer than 25 characters appear verbatim in the 1507 mod. The differences:
The two can't run together; the README now lists the 1507 mod as a conflict. Your reason for a separate mod, from your PR comment: the Update 3 Metro look is visually distinct, and combining both designs in one mod could make its purpose and settings confusing for users who want one specific design. You also note that the search, cache and helper-process work could be brought to the 1507 mod separately, and that you leave the decision to the maintainer. The consolidation path would be the Update 3 look as a style option of the 1507 mod. Whether a separate mod is justified is the maintainer's decision. 2. Please confirm the hybrid paths were tested. The removed README note said the startup, elevation, Explorer restart and shutdown paths still "require testing on real Windows". The note is gone, but neither the PR nor your comment says they were tested. They are the riskiest parts of the mod: the shell Explorer relaunches the helper, and the bridge runs inside Explorer. Please say in the PR which of these you checked:
Optional improvements
Minor polish — none of this affects users, so it's your call.
Functionality notes
Non-critical observations and ideas about the feature behavior itself.
Code overview
For the maintainer: a map of the mod and what it touches outside its own process. Descriptive only — anything that needs a change is listed above. Line numbers refer to the mod source at commit The mod runs in two roles. In the shell Code regions
Side effects of interest
Next steps:
See the review process for details. Important: Each reply that was generated by AI must be disclosed by including the following text in the reply: "Generated by AI. Model: (ai-model-name)". |
|
After an explorer.exe restart the mod's start menu appears after a few seconds (the same applies when re-enabling the mod); at startup it seems to work. I can't confirm if it works with an update, as the mod has not been approved on the catalog, so there isn't an update to receive. I have tested it with taskmgr.exe and cmd.exe as administrator, and the mod's start menu opens. For shutdown, I don't think there is anything major to test, as the computer is shut down so there isn't a mod to test. However, I have restarted the PC (and shut it down and started it normally) to check for problems, and the mod's start menu started after 5-10 seconds. On the elevated Windows case, the mod seems to function correctly with both WIN and the click on the start menu after the update. If the mod is tried on another Windows environment and problems are found, please inform me. [StartMenu] isolated Win tap detected [StartMenu] isolated Win tap detected [StartMenu] isolated Win tap detected [StartMenu] isolated Win tap detected [StartMenu] isolated Win tap detected [Toggle] request from Explorer bridge posted=1 [StartMenu] isolated Win tap detected Regarding the overlap, there are various additional reasons for which keeping it as a separate mod is useful:
I'd also like to offer a different perspective on the shared code, though I understand this is ultimately the maintainer's call. Reusing the existing implementation seemed safer than rewriting it. The parts shared with the 1507 mod are the delicate ones — the WH_KEYBOARD_LL/WH_MOUSE_LL hooks, the tile drag/resize/pin model, and the WHSM2# layout parsing. Since that code is already merged and tested, I chose to build on it rather than start from scratch, partly because a mistake in those paths can crash Explorer. That said, I realize this is a trade-off, not an obvious win. The similarity is mostly structural, not accidental. Both mods recreate the same interaction model because they target the same OS APIs and user expectations. The parts that differ are the ones I'd expect to differ: the visual style, search, and helper-process architecture. I understand the maintenance concern is real. If the two copies diverge over time, a fix applied to one might not reach the other. Extracting a shared module would address that, but it would also mean designing an abstraction, testing it against both designs, and maintaining it — which felt like a lot for a single volunteer contributor. I'm open to doing that work if the maintainer thinks it's worth it, but I didn't want to block this PR on it. So I see the duplication as a reasonable trade-off rather than a flaw, but I'm happy to defer to the maintainer on whether it's acceptable. Generated by AI. Model: Claude |
|
/ai-review |
Submission reviewNote: This review was done by Claude. Due to the amount of submissions, doing a fully manual review for each pull request is no longer feasible. Thank you for understanding. Remember: The AI reviewer can be wrong - it may misread code, flag correct code as broken, or suggest changes that make things worse. Treat its findings as suggestions to verify, not instructions to follow blindly. You're responsible for the code you submit, so if a finding doesn't hold up, say so instead of changing working code to satisfy it. Please address the following issues. The items in the collapsed sections are optional, so it's your call whether to address them. Thanks for the test report. It covers what the last round asked for: startup, Explorer restart, re-enabling, elevated windows in the foreground (Task Manager and an admin One item remains: 1. Substantial overlap with your own
About 60% of the file's lines longer than 25 characters appear verbatim in the 1507 mod. The differences:
The two can't run together; the README lists the 1507 mod as a conflict. Your reasons for a separate mod, from your two PR comments:
On the shared code, you write that:
The consolidation path would be the Update 3 look as a style option of the 1507 mod. Whether a separate mod is justified is the maintainer's decision. Optional improvements
Minor polish — none of this affects users, so it's your call.
Functionality notes
Non-critical observations and ideas about the feature behavior itself.
Code overview
For the maintainer: a map of the mod and what it touches outside its own process. Descriptive only — anything that needs a change is listed above. Line numbers refer to the mod source at commit The mod runs in two roles. In the shell Code regions
Side effects of interest
Next steps:
See the review process for details. Important: Each reply that was generated by AI must be disclosed by including the following text in the reply: "Generated by AI. Model: (ai-model-name)". |
|
Note for the maintainer on the 1507 overlap The AI review has passed the remaining design question to you, so here is the case in one place. The internal base is shared on purpose. The two mods share the same technical foundation — the low-level keyboard/mouse hooks, the tile drag/resize/pin model, and the WHSM2# layout parsing. These are the delicate parts of a Start replacement: a mistake there can crash Explorer. That code is already merged and tested in the 1507 mod, so I reused it instead of rewriting it. Reinventing the wheel here would have added risk, not value, and would have delayed the parts that are actually new. For the end user, the two mods are completely different. The shared code is invisible to the person using the mod. What they see and interact with is a different product: Different visual generation. This mod recreates the Windows 8.1 Update 3 Metro look — the purple background, the flat tiles, the Windows 8-style popups. The 1507 mod recreates the Windows 10 1507 design. These are two distinct Metro generations, not a skin on top of the same interface. Windows 8.1 is a separate product generation, not a variant of 1507. Different target audience. Someone who wants Windows 10/11 to look like Windows 8.1 is looking for a specific, recognizable design. That user is unlikely to find it inside a mod titled "Windows 10 1507 Start Menu Recreation", because 1507 represents a major rewrite compared to 8.1, not a minor OS change. Keeping the mod separate makes it discoverable by the audience it is actually for. Different settings surface. A merged mod would need a top-level style switch plus options that apply to only one of the two designs. Users would have to reason about which settings affect which look, and the risk of misconfiguring a design they did not intend to use would grow. Two focused mods keep each settings page coherent for its own audience. On the duplication trade-off. I understand the maintenance concern: if the two copies diverge, a fix applied to one might not reach the other. I see three ways forward, and I am open to whichever you prefer: A — Consolidate: add the Update 3 look as a style option of the 1507 mod. The search, cache and helper-process work would then benefit both. I can do this if you think it is the right call. B — Keep them separate, with an explicit commitment to keep critical fixes in sync (for example, a tracking issue or a comment in both files pointing to the shared logic). I did not want to block this PR on a refactor you may not consider necessary, which is why I am deferring the choice to you rather than deciding unilaterally. What is already done. The technical findings from the first review are fixed and verified: storage moved to per-user files with temp+rename, tab-separated launch counts, fingerprint import marker, shared icon converter, fallback for invalid tileLayout pastes, and the README now documents the helper process, the conflicting mods and a screenshot. The hybrid paths have been tested and reported (startup, Explorer restart, re-enabling, elevated windows in the foreground with both the Win key and a Start-button click, restart/shutdown). The rest of the PR is ready for your review. Generated by AI. Model: Claude /ready-for-reviewer |
Why not? Is it not like "Windows 7/8/8.1 Taskbar Recreation" where you can select the skin? |
Thanks for the question. The discussed point is partially true. It is indeed doable to add a skin, but having it as a separate mod in this distinct case is useful especially for the final user because the mods have almost the same internal code but the interface is entirely different, and from a final user perspective it would be harder to find a mod that combines both skins together. In the case of the taskbar recreation, the name itself is explanatory, but something like "Windows 8.1 Update 3/Windows 10 1507 Start Menu Recreation" reads awkwardly from my perspective. The taskbar recreation name works because "7/8/8.1" is a compact list of variants of the same object: the taskbar looks broadly similar across those versions. In this case, "Windows 8.1 Update 3" and "Windows 10 1507" are two complete product names describing two different Start menu designs, not variants of one. A combined name would read like two products stapled together, not like a list of skins for the same product. That is why the "skin option" pattern fits the taskbar recreation but is awkward here. At least, this is my opinion. If it is specifically required to combine them, I'll do it, but it might make discoverability worse as standard Windows 8.1 did not include a Start Menu. Additionally, I planned to add more Windows 8.1-related features in future updates for it, making the mod more distinct than the 1507 one. |
|
I'd use a less specific name such as "Windows 8.1/10 Start Menu Recreation", and then name the specific revisions in the description or the readme. Otherwise, we can end up with dozens of mods for various Windows revisions, that's unlikely to be more convenient for users. |
Changelog
If this pull request updates an existing mod, describe the changes inside the changelog markers:
Mod authorship
If this pull request introduces a new mod, please complete the section below.
This mod was created by:
Please select the options that best apply. Your selection does not affect the acceptance criteria, but it helps reviewers understand the context of the code and provide relevant feedback.