task: Windows platform issues (8 bugs) #394

Activity

1. DeusData

   added

   bugSomething isn't workingSomething isn't working

   windowsWindows-specific issuesWindows-specific issues

   taskUmbrella task grouping multiple related issuesUmbrella task grouping multiple related issues

   on May 30, 2026
2. DeusData

   added 2 commits that reference this issue on May 31, 2026

   ci(temp): Windows repro harness for #394 group

   ...

   d198403

   ci: remove temporary Windows repro harness

   ...

   9dcc0e3
3. DeusData commented on Jun 1, 2026

   

   DeusData

   on Jun 1, 2026

   OwnerAuthor

   More actions

   All 8+ Windows bugs are now closed — fixed across v0.7.0: non-ASCII path support via wide-char API (#386), PATHEXT probing in find_in_path (#221), Cursor MCP registration (#222 via integration), fish PATH syntax (#319), lowercase/mapped-drive root_path integrity (#367/#227, c1dabc9), and the cmd.exe git-log history pass (df38e33). Closing the umbrella.
4. DeusData

   closed this as completedon Jun 1, 2026
5. bsun0000 commented on Jun 2, 2026

   

   bsun0000

   on Jun 2, 2026

   Last edited by bsun0000

   More actions

   It is marked as closed and completed, but it still does not work for me, Windows 10.
   Tested versions: v0.6.1, v0.7.0, git-latest (forked & compiled via git workflow).

   Cache sits in C:\Users\BlackSun.cache\codebase-memory-mcp\ - no weird symbols in the path, no spaces.
   My codebase is indexed here, stored in a .db file.

   codebase-memory-mcp.exe cli list_projects - does not list anything. Even if i explicitly set the CBM_CACHE_DIR env. parameter.
   {"projects":[],"hint":"No projects indexed. Call index_repository(repo_path=...) first."}

   If i copy _config.db to config.db, it now shows:
   {"projects":[{"name":"config","root_path":"","nodes":0,"edges":0,"size_bytes":12288}]}

   So it sees the path and can read it. Same behavior with --ui=true - webui ignores the actual indexed codebase.

   The web UI is still broken - "Select Repository Folder" shows folders with empty names.

   The index is 4.2Gb big, memory usage spiked to almost 128GB of RAM during indexing. But there were no errors in the console.. still, could this be the issue? The codebase was simply.. too big? Around 100k files to index, 5.85Gb in total, assuming this program didn't ignore both .gitignore and .cbmignore i setup to cut some large and problematic files.

   UPD: Both DB Browser and SQLite Expert fails to open the index: "database disk image is malformed". *sigh*..

   👀React with 👀1Reacted by Martin Vogel
6. DeusData

   reopened this on Jun 2, 2026
7. DeusData commented on Jun 2, 2026

   

   DeusData

   on Jun 2, 2026

   Last edited by DeusData

   OwnerAuthor

   More actions

   Hey @bsun0000 thx for reporting, will take another look. Sorry for the inconvenience. Can you give more details on what kinda project you try to index? In general a summary of your setup and indexing targets would be good. Maybe you can let your agent make a summary on this one. Need some way to reproduce this
8. janmajorczyk commented on Jun 2, 2026

   

   janmajorczyk

   on Jun 2, 2026

   More actions

   Adding a second data point for the reopen — this one is a minimal, deterministic, CI-reproducible repro of the lowercase-drive root_path integrity check that c1dabc9 was supposed to fix, on a plain local NTFS path with no SMB/mapped drive. It still fails on v0.7.0.

   This looks distinct from @bsun0000's case (100k files, 4.2 GB index, ~128 GB RAM, "database disk image is malformed") — that one smells like a size/OOM-driven corruption. Mine is the opposite: a normal ~1,200-file repo, no memory pressure, no crash. The only variable is the case of the drive letter, which isolates the integrity-check path cleanly.

   It also directly answers the two things asked in #227/#367 — "v0.7.0 retest" and "exact stored root_path".

   Environment

   • codebase-memory-mcp 0.7.0 (also repro'd that uppercase works on the same build)
   • Windows 11, local NTFS C: drive, cache at the default C:\Users\<user>\.cache\codebase-memory-mcp\
   • No SMB / UNC / mapped / subst drive. No spaces or non-ASCII in the path.

   Exact stored root_path

   • Fails: c:/WEBDEV/Signalyzed_Projects/Cardio-Cloud → project key c-WEBDEV-Signalyzed_Projects-Cardio-Cloud
   • Works: C:/WEBDEV/Signalyzed_Projects/Cardio-Cloud → project key C-WEBDEV-Signalyzed_Projects-Cardio-Cloud

   The only difference is c: vs C:.

   Minimal repro (CLI)

   # lowercase drive letter -> indexes, then self-deletes
   codebase-memory-mcp cli index_repository '{"repo_path":"c:/Projects/MyRepo","mode":"moderate"}'
   codebase-memory-mcp cli list_projects     # -> {"projects":[]}
   
   # identical repo, uppercase drive letter -> persists, queries work
   codebase-memory-mcp cli index_repository '{"repo_path":"C:/Projects/MyRepo","mode":"moderate"}'
   codebase-memory-mcp cli list_projects     # -> project present

   What the lowercase run logs — the build completes, then the projects-table integrity check rejects the row, deletes the DB, and still returns "indexed":

   level=info  msg=pipeline.done nodes=0 edges=15308 elapsed_ms=4947
   ERROR store.corrupt table=projects bad_root_path=c:/WEBDEV/Signalyzed_Projects/Cardio-Cloud
   level=error msg=store.auto_clean project=c-WEBDEV-Signalyzed_Projects-Cardio-Cloud \
               path=C:/Users/<user>/.cache/codebase-memory-mcp/c-WEBDEV-Signalyzed_Projects-Cardio-Cloud.db \
               action=deleting corrupt db — re-index required
   {"project":"c-WEBDEV-Signalyzed_Projects-Cardio-Cloud","status":"indexed"}
   

   Every later search_graph / index_status / get_architecture then returns "project not found or not indexed" — even though the response's own available_projects lists the project — because the DB was deleted right after being written.

   Why this should be CI-testable (the blocker mentioned in #394): no SMB/mapped drive is involved — just a lowercase drive letter. A smoke test that indexes a fixture with repo_path starting c:/… and asserts list_projects is non-empty afterward would catch it.

   Two things that may help:

   1. The "indexed" response after store.auto_clean deletes the DB is the real trap — index_repository reports success while list_projects is empty, so it looks like a query bug rather than an indexing one. Surfacing the corruption as an error (or dropping nodes/edges + adding a warning) would make this self-evident.
   2. This is very common with agent-driven indexing. Agents like Claude Code report the Windows CWD as lowercase c:\…, so anything that derives repo_path from the CWD hits this by default. (Workaround we're using: always pass an uppercase drive letter and set auto_index=false, since on a case-insensitive FS the c-….db and C-….db cache files collide and a lowercase auto-index clobbers a good uppercase index.)

   Suggested fix: upper-case (or case-fold) the drive letter wherever separator normalization already happens (cbm_normalize_path_sep() from v0.5.7?), before the projects-table integrity check computes/compares bad_root_path — or compare case-insensitively on Windows. Since c1dabc9 shipped in v0.7.0 but this still reproduces, the normalization likely isn't applied on the integrity-check read path for plain local drives.

   Happy to run a diagnostic build or grab fuller --verbose logs if useful.

   ❤️React with ❤️1Reacted by Martin Vogel
9. DeusData commented on Jun 2, 2026

   

   DeusData

   on Jun 2, 2026

   OwnerAuthor

   More actions

   @bsun0000 thanks for sharing. You can take down the link. Ideally provide me with an OSS project with similar traits, not feeling so comfortable with downloading files from anonymous people (nothing against you, hope you understand it). Maybe you can ask your agent to compare your repo with public available OSS ones and get the key characteristics matched so that I can use it for reproducibility. The size I can replicate easily by just copying the repo over and over and indexing them together. That is fine for me. @janmajorczyk also thanks, will have a look as well

   ❤️React with ❤️1Reacted by janmajorczyk
10. DeusData

    added a commit that references this issue on Jun 2, 2026

    feat(graph): queryable computation-bottleneck metrics + indexing perf…

    ...

    12683b8
11. FDiskas commented on Jun 2, 2026

    

    FDiskas

    on Jun 2, 2026

    Last edited by FDiskas

    More actions

    CPU and 32GB of ram was used at max level to index my small project and everything crashed.
12. lichaoren commented on Jun 4, 2026

    

    lichaoren

    on Jun 4, 2026

    More actions

    I still experience (not so) random crash on indexing my large C++ codebase. As pointed out in another closed issue, running in a single thread delays the crash, and parses 99% of the files.

    PS D:\mcp\codebase-memory-mcp> ./codebase-memory-mcp.exe cli index_repository '{"repo_path":"D:/repo/fresh-gameservice"}'; echo "EXIT=$LASTEXITCODE"
    level=info msg=mem.init budget_mb=32728 total_ram_mb=65456
    level=info msg=pipeline.discover files=8459 elapsed_ms=235
    level=info msg=pipeline.route path=glimmer/amber/route
    level=info msg=pass.start pass=structure files=8459
    level=info msg=pass.done pass=structure nodes=6433 edges=6319
    level=info msg=pass.timing pass=structure elapsed_ms=27
    level=info msg=pipeline.mode mode=parallel workers=32 files=8459
    level=info msg=parallel.extract.start files=8459 workers=32
    level=info msg=parallel.mem.budget total_mb=32728 per_worker_mb=1022
    level=info msg=parallel.extract.file.start pos=0 size_kb=316 path=quartz/nova/ember/ridge/flux/static/css/toolkit/opal-responsive.css
    level=info msg=parallel.extract.file.start pos=1 size_kb=307 path=quartz/nova/ember/ridge/flux/static/css/toolkit/opal-web.css
    level=info msg=parallel.extract.file.start pos=2 size_kb=262 path=vendor/atlas/mysql-kit-linux/include/zephyr_error.h
    level=info msg=parallel.extract.file.start pos=3 size_kb=259 path=quartz/nova/ember/ridge/flux/static/js/toolkit/toolkit.js
    level=info msg=parallel.extract.file.start pos=5 size_kb=251 path=quartz/nova/ember/ridge/flux/static/css/toolkit/opal-web.min.css
    level=info msg=parallel.extract.file.start pos=6 size_kb=200 path=quartz/service/aurora/aurora_state.cpp
    level=info msg=parallel.extract.file.start pos=4 size_kb=257 path=quartz/nova/ember/ridge/flux/static/css/toolkit/opal-responsive.min.css
    level=info msg=parallel.extract.file.start pos=7 size_kb=181 path=quartz/service/meteor/meteor_controller.cpp
    level=info msg=parallel.extract.file.start pos=8 size_kb=152 path=quartz/service/comet/comet_controller.cpp
    level=info msg=parallel.extract.file.start pos=10 size_kb=137 path=quartz/core/network/neutron_cluster.cpp
    level=info msg=parallel.extract.file.start pos=11 size_kb=128 path=quartz/core/network/neutron_connection.cpp
    level=info msg=parallel.extract.file.start pos=12 size_kb=117 path=quartz/nova/ember/ridge/flux/static/js/toolkit/toolkit.min.js
    level=info msg=parallel.extract.file.start pos=9 size_kb=139 path=quartz/core/http/orbit_http2_server.cpp
    level=info msg=parallel.extract.file.start pos=16 size_kb=96 path=quartz/nova/ember/ridge/flux/static/js/third-party/sundown/sundown.min.js
    level=info msg=parallel.extract.file.start pos=14 size_kb=115 path=settings/nebula/common.yaml
    level=info msg=parallel.extract.file.start pos=15 size_kb=106 path=quartz/core/stream/nebula/consumer_impl.cpp
    level=info msg=parallel.extract.file.start pos=18 size_kb=92 path=quartz/service/accounting/storage/account_storage.cpp
    level=info msg=parallel.extract.file.start pos=13 size_kb=117 path=vendor/atlas/mysql-kit-win/include/zephyr_ername.h
    level=info msg=parallel.extract.file.start pos=20 size_kb=85 path=quartz/service/aurora/aurora_controller.cpp
    level=info msg=parallel.extract.file.start pos=17 size_kb=95 path=quartz/nova/ember/ridge/flux/static/css/develop/develop.min.css
    level=info msg=parallel.extract.file.start pos=22 size_kb=83 path=quartz/service/guild/guild_controller.cpp
    level=info msg=parallel.extract.file.start pos=19 size_kb=86 path=vendor/atlas/mysql-kit-win/include/mysql/psi/psi.h
    level=info msg=parallel.extract.file.start pos=23 size_kb=81 path=quartz/core/session/consul_session.cpp
    level=info msg=parallel.extract.file.start pos=21 size_kb=84 path=quartz/service/accounting/account_command.cpp
    level=info msg=parallel.extract.file.done pos=2 elapsed_ms=61 defs=1 path=vendor/atlas/mysql-kit-linux/include/zephyr_error.h
    level=info msg=parallel.extract.file.done pos=14 elapsed_ms=16 defs=9 path=settings/nebula/common.yaml
    level=info msg=parallel.extract.progress done=30 total=8459
    level=info msg=parallel.extract.file.done pos=0 elapsed_ms=121 defs=1 path=quartz/nova/ember/ridge/flux/static/css/toolkit/opal-responsive.css
    level=info msg=parallel.extract.file.done pos=1 elapsed_ms=123 defs=1 path=quartz/nova/ember/ridge/flux/static/css/toolkit/opal-web.css
    level=info msg=parallel.extract.file.done pos=17 elapsed_ms=33 defs=1 path=quartz/nova/ember/ridge/flux/static/css/develop/develop.min.css
    level=info msg=parallel.extract.file.done pos=16 elapsed_ms=66 defs=1 path=quartz/nova/ember/ridge/flux/static/js/third-party/sundown/sundown.min.js
    level=info msg=parallel.extract.file.done pos=5 elapsed_ms=117 defs=1 path=quartz/nova/ember/ridge/flux/static/css/toolkit/opal-web.min.css
    level=info msg=parallel.extract.file.done pos=19 elapsed_ms=31 defs=681 path=vendor/atlas/mysql-kit-win/include/mysql/psi/psi.h
    level=info msg=parallel.extract.file.done pos=4 elapsed_ms=122 defs=1 path=quartz/nova/ember/ridge/flux/static/css/toolkit/opal-responsive.min.css
    level=info msg=parallel.extract.file.done pos=3 elapsed_ms=174 defs=12 path=quartz/nova/ember/ridge/flux/static/js/toolkit/toolkit.js
    level=info msg=parallel.extract.file.done pos=12 elapsed_ms=152 defs=12 path=quartz/nova/ember/ridge/flux/static/js/toolkit/toolkit.min.js
    level=info msg=parallel.extract.file.done pos=22 elapsed_ms=111 defs=98 path=quartz/service/guild/guild_controller.cpp
    level=info msg=parallel.extract.progress done=20 total=8459
    level=info msg=parallel.extract.file.done pos=21 elapsed_ms=103 defs=47 path=quartz/service/accounting/account_command.cpp
    level=info msg=parallel.extract.file.done pos=13 elapsed_ms=149 defs=1 path=vendor/atlas/mysql-kit-win/include/zephyr_ername.h
    level=info msg=parallel.extract.progress done=40 total=8459
    level=info msg=parallel.extract.file.done pos=11 elapsed_ms=204 defs=81 path=quartz/core/network/neutron_connection.cpp
    level=info msg=parallel.extract.file.done pos=10 elapsed_ms=213 defs=79 path=quartz/core/network/neutron_cluster.cpp
    level=info msg=parallel.extract.file.done pos=18 elapsed_ms=187 defs=77 path=quartz/service/accounting/storage/account_storage.cpp
    level=info msg=parallel.extract.file.done pos=23 elapsed_ms=154 defs=60 path=quartz/core/session/consul_session.cpp
    level=info msg=parallel.extract.file.done pos=9 elapsed_ms=225 defs=121 path=quartz/core/http/orbit_http2_server.cpp
    level=info msg=parallel.extract.file.done pos=15 elapsed_ms=214 defs=64 path=quartz/core/stream/nebula/consumer_impl.cpp
    level=info msg=parallel.extract.file.done pos=20 elapsed_ms=231 defs=128 path=quartz/service/aurora/aurora_controller.cpp
    level=info msg=parallel.extract.progress done=60 total=8459
    level=info msg=parallel.extract.progress done=10 total=8459
    level=info msg=parallel.extract.progress done=50 total=8459
    level=info msg=parallel.extract.progress done=70 total=8459
    level=info msg=parallel.extract.file.done pos=6 elapsed_ms=356 defs=121 path=quartz/service/aurora/aurora_state.cpp
    level=info msg=parallel.extract.progress done=100 total=8459
    level=info msg=parallel.extract.file.done pos=7 elapsed_ms=352 defs=96 path=quartz/service/meteor/meteor_controller.cpp
    level=info msg=parallel.extract.file.done pos=8 elapsed_ms=370 defs=88 path=quartz/service/comet/comet_controller.cpp
    level=info msg=parallel.extract.progress done=90 total=8459
    level=info msg=parallel.extract.progress done=80 total=8459
    level=info msg=parallel.extract.progress done=110 total=8459
    level=info msg=parallel.extract.progress done=120 total=8459
    level=info msg=parallel.extract.progress done=140 total=8459
    level=info msg=parallel.extract.progress done=130 total=8459
    level=info msg=parallel.extract.progress done=160 total=8459
    level=info msg=parallel.extract.progress done=180 total=8459
    level=info msg=parallel.extract.progress done=150 total=8459
    level=info msg=parallel.extract.progress done=170 total=8459
    level=info msg=parallel.extract.progress done=190 total=8459
    level=info msg=parallel.extract.progress done=230 total=8459
    level=info msg=parallel.extract.progress done=210 total=8459
    level=info msg=parallel.extract.progress done=200 total=8459
    level=info msg=parallel.extract.progress done=240 total=8459
    level=info msg=parallel.extract.progress done=220 total=8459
    level=info msg=parallel.extract.progress done=270 total=8459
    level=info msg=parallel.extract.progress done=260 total=8459
    level=info msg=parallel.extract.progress done=250 total=8459
    level=info msg=parallel.extract.progress done=280 total=8459
    level=info msg=parallel.extract.progress done=290 total=8459
    level=info msg=parallel.extract.progress done=300 total=8459
    EXIT=-1073741819
    

    Single thread was in an ubuntu WSL, the other variable is smaller RAM size there, FYI
13. 9 remaining items

    Load more
14. Flipper1994

    added a commit that references this issue on Jul 4, 2026

    test(windows): note umbrella tracker DeusData#394 in red-test analysis

    ...

    2b4fa2d
15. DeusData

    added this to the 0.9.1-rc milestone on Jul 8, 2026
16. DeusData

    added sub-issues on Jul 24, 2026

    • Installer targets legacy Gemini-CLI settings.json instead of Antigravity 2.0 config directory (resulting in duplicate 273MB binaries) #1161
    • Windows: CBM_CACHE_DIR / CBM_ALLOWED_ROOT are lossily decoded — every non-ASCII char becomes U+FFFD, so the server silently uses a different path (v0.9.0) #1165
    • Can't upgrade codebase-memory-mcp on windows platform. Can't uninstall. #1168
    • [v0.9.0][Windows] persistence=true crashes the indexing worker for repositories under CJK paths #1171
    • Windows Defender report: Win32/Peardis.C #1194
    • [v0.9.0][Windows] search_code content pipeline transcodes UTF-8 through the ANSI codepage — CJK pattern search always returns 0 matches, raw_matches show mojibake #1209
    • Windows: Git discovery fails for valid repository at Unicode path #1240
    • [v0.9.0] install.ps1 expects codebase-memory-mcp.payload.exe missing from release archive #1218
17. github-actions

    added

    editor/integrationEditor compatibility and CLI integrationEditor compatibility and CLI integration

    parsing/qualityGraph extraction bugs, false positives, missing edgesGraph extraction bugs, false positives, missing edges

    stability/performanceServer crashes, OOM, hangs, high CPU/memoryServer crashes, OOM, hangs, high CPU/memory

    ux/behaviorDisplay bugs, docs, adoption UXDisplay bugs, docs, adoption UX

    on Aug 3, 2026
18. DeusData commented on Oct 25, 2026

    

    DeusData

    on Oct 25, 2026

    OwnerAuthor

    More actions

    Thank you to everyone who supplied the Windows retests and workarounds. The later comments show why the original completed checklist cannot represent the whole umbrella: installer registration, UI indexing, and path/CLI behavior have separate follow-ups. I have no new blanket resolution to report, and the supplied examples do not need repeating. Sorry for the delay—this is waiting on maintainer follow-up, and I’ll follow up soon.