home.social

#libgit2 — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #libgit2, aggregated by home.social.

  1. Git в браузере. Расширяем возможности с помощью LFS

    Привет, Хабр! Я Паша, разработчик

    habr.com/ru/companies/gram_ax/

    #git #libgit2 #lfs #webassembly #rust #ffi #emscripten #docsascode #opensource #localfirst

  2. Git в браузере. Расширяем возможности с помощью LFS

    Привет, Хабр! Я Паша, разработчик

    habr.com/ru/companies/gram_ax/

    #git #libgit2 #lfs #webassembly #rust #ffi #emscripten #docsascode #opensource #localfirst

  3. Git в браузере. Расширяем возможности с помощью LFS

    Привет, Хабр! Я Паша, разработчик

    habr.com/ru/companies/gram_ax/

    #git #libgit2 #lfs #webassembly #rust #ffi #emscripten #docsascode #opensource #localfirst

  4. Git в браузере. Расширяем возможности с помощью LFS

    Привет, Хабр! Я Паша, разработчик

    habr.com/ru/companies/gram_ax/

    #git #libgit2 #lfs #webassembly #rust #ffi #emscripten #docsascode #opensource #localfirst

  5. Как мы засунули Git в браузер (и чего нам это стоило)

    Привет, Хабр! Меня зовут Паша, я разработчик

    habr.com/ru/companies/gram_ax/

    #emscripten #rust #webassembly #libgit2 #wasm #git #wysiwyg #open_source #localfirst

  6. Как мы засунули Git в браузер (и чего нам это стоило)

    Привет, Хабр! Меня зовут Паша, я разработчик

    habr.com/ru/companies/gram_ax/

    #emscripten #rust #webassembly #libgit2 #wasm #git #wysiwyg #open_source #localfirst

  7. Как мы засунули Git в браузер (и чего нам это стоило)

    Привет, Хабр! Меня зовут Паша, я разработчик

    habr.com/ru/companies/gram_ax/

    #emscripten #rust #webassembly #libgit2 #wasm #git #wysiwyg #open_source #localfirst

  8. Как мы засунули Git в браузер (и чего нам это стоило)

    Привет, Хабр! Меня зовут Паша, я разработчик

    habr.com/ru/companies/gram_ax/

    #emscripten #rust #webassembly #libgit2 #wasm #git #wysiwyg #open_source #localfirst

  9. In the course of tracking down a runtime error in #libgit2 (v1.9.0), I ran the #Xcode static analyzer against it, and it's kind of scary…

  10. In the course of tracking down a runtime error in #libgit2 (v1.9.0), I ran the #Xcode static analyzer against it, and it's kind of scary…

  11. In the course of tracking down a runtime error in #libgit2 (v1.9.0), I ran the #Xcode static analyzer against it, and it's kind of scary…

  12. In the course of tracking down a runtime error in #libgit2 (v1.9.0), I ran the #Xcode static analyzer against it, and it's kind of scary…

  13. In the course of tracking down a runtime error in #libgit2 (v1.9.0), I ran the #Xcode static analyzer against it, and it's kind of scary…

  14. Sytuacja: są trzy konkurujące ze sobą łatki na dodanie wsparcia #libgit2 1.8 do libgit2-glib. I każda z nich jest trochę inna, niektóre z jasną sugestią, że autor nie miał pojęcia, co robi.

    Serio, czy pokolenie GitHuba naprawdę nie potrafi poszukać, zanim wyśle własną łatkę?

    A tak przy okazji, wygląda na to, że ten projekt umarł.

    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2

    #GitHub

  15. Situation: there are 3 competing merge requests to add #libgit2 1.8 support to libgit2-glib. And all of them are slightly different, some with clear indication that the author had no clue what they were doing.

    Seriously, can't #GitHub generation actually search before sending something?

    Also, the project seems entirely dead.

    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2

  16. I was planning on using #libgit2 for my next project, and I wanted a bit of crash course on #git #internals, and came across this absolutely fantastic series of tutorials on #YouTube by Omer Rosenbaum. If you want to dig into git in more detail, these are a great place to start. youtu.be/fWMKue-WBok?si=yygfpQ

  17. #SwiftGit2 is the #Swift language binding that the #libgit2 project references. For reasons, I wanted SwiftGit2 to be available as a Swift package. I discovered that @mathijsb had made libgit2 itself available as a Swift package and then used that to simplify and clean up SwiftGit2 as a Swift package. This makes consuming SwiftGit2 with up-to-date libgit2 dependencies much easier! Links:

    #SPM support for SwiftGit2: github.com/mbernson/SwiftGit2

    #SPM support for libgit2: github.com/mbernson/libgit2

  18. It's been a while since I've complained about #RustLang itself, so…

    Cargo insists on interacting with #git repositories. At the same time, cargo insists on vendoring an old version of #LibGit2 (1.6.2 FWICS). So, if your system is using a new git version (2.44.0), you won't be able to `cargo build`:

    ```
    error: failed to determine package fingerprint for build script for uv v0.1.38 (/tmp/uv/crates/uv)

    Caused by:
    failed to determine the most recently modified file in /tmp/uv/crates/uv

    Caused by:
    failed to determine list of files in /tmp/uv/crates/uv

    Caused by:
    failed to open git index at /tmp/uv/.git/

    Caused by:
    invalid data in index - calculated checksum does not match expected; class=Index (10)
    ```

    You have to clone everything with `-c index.skipHash=false` to work around this.

    But yeah, I'm sure there's a great benefit to using an outdated vendored C library that NIHs git.

    github.com/rust-lang/cargo/iss

    #Gentoo #NIH

  19. Już dłuższą chwilę nie narzekałem na Rusta samego w sobie, więc…

    Cargo upiera się na interakcji z repozytoriami #git. Jednocześnie, Cargo upiera się przy używaniu włączonej (zamiast systemowej), starej wersji #LibGit2 (z tego, co widzę: 1.6.2). Tak więc, jeżeli wasz system używa nowego gita (2.44.0), to `cargo build` nie przejdzie:

    ```
    error: failed to determine package fingerprint for build script for uv v0.1.38 (/tmp/uv/crates/uv)

    Caused by:
    failed to determine the most recently modified file in /tmp/uv/crates/uv

    Caused by:
    failed to determine list of files in /tmp/uv/crates/uv

    Caused by:
    failed to open git index at /tmp/uv/.git/

    Caused by:
    invalid data in index - calculated checksum does not match expected; class=Index (10)
    ```

    Można to obejść, klonując wszystko z `-c index.skipHash=false`.

    Ale spoko, na pewno jest wielka zaleta z użycia przestarzałej, włączonej na sztywno biblioteki w C, która wynajduje gita na nowo.

    github.com/rust-lang/cargo/iss

    #Gentoo #RustLang

  20. Will #libgit2 ever gain true non-blocking #async IO using io_uring or alternatives?
    Cannot imagine the resulting speed up over network storage.

  21. #git is an object store with immutable data. Hoping to see a custom git ref backend like #libgit2 to store refs in a DB and start using object store backend (instead of #POSIX compliant filesystems)

  22. Sytuacja: są trzy konkurujące ze sobą łatki na dodanie wsparcia #libgit2 1.8 do libgit2-glib. I każda z nich jest trochę inna, niektóre z jasną sugestią, że autor nie miał pojęcia, co robi.

    Serio, czy pokolenie GitHuba naprawdę nie potrafi poszukać, zanim wyśle własną łatkę?

    A tak przy okazji, wygląda na to, że ten projekt umarł.

    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2

    #GitHub

  23. Sytuacja: są trzy konkurujące ze sobą łatki na dodanie wsparcia #libgit2 1.8 do libgit2-glib. I każda z nich jest trochę inna, niektóre z jasną sugestią, że autor nie miał pojęcia, co robi.

    Serio, czy pokolenie GitHuba naprawdę nie potrafi poszukać, zanim wyśle własną łatkę?

    A tak przy okazji, wygląda na to, że ten projekt umarł.

    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2

    #GitHub

  24. Sytuacja: są trzy konkurujące ze sobą łatki na dodanie wsparcia #libgit2 1.8 do libgit2-glib. I każda z nich jest trochę inna, niektóre z jasną sugestią, że autor nie miał pojęcia, co robi.

    Serio, czy pokolenie GitHuba naprawdę nie potrafi poszukać, zanim wyśle własną łatkę?

    A tak przy okazji, wygląda na to, że ten projekt umarł.

    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2

    #GitHub

  25. Sytuacja: są trzy konkurujące ze sobą łatki na dodanie wsparcia #libgit2 1.8 do libgit2-glib. I każda z nich jest trochę inna, niektóre z jasną sugestią, że autor nie miał pojęcia, co robi.

    Serio, czy pokolenie GitHuba naprawdę nie potrafi poszukać, zanim wyśle własną łatkę?

    A tak przy okazji, wygląda na to, że ten projekt umarł.

    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2

    #GitHub

  26. Situation: there are 3 competing merge requests to add #libgit2 1.8 support to libgit2-glib. And all of them are slightly different, some with clear indication that the author had no clue what they were doing.

    Seriously, can't #GitHub generation actually search before sending something?

    Also, the project seems entirely dead.

    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2

  27. Situation: there are 3 competing merge requests to add #libgit2 1.8 support to libgit2-glib. And all of them are slightly different, some with clear indication that the author had no clue what they were doing.

    Seriously, can't #GitHub generation actually search before sending something?

    Also, the project seems entirely dead.

    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2

  28. Situation: there are 3 competing merge requests to add #libgit2 1.8 support to libgit2-glib. And all of them are slightly different, some with clear indication that the author had no clue what they were doing.

    Seriously, can't #GitHub generation actually search before sending something?

    Also, the project seems entirely dead.

    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2

  29. Situation: there are 3 competing merge requests to add #libgit2 1.8 support to libgit2-glib. And all of them are slightly different, some with clear indication that the author had no clue what they were doing.

    Seriously, can't #GitHub generation actually search before sending something?

    Also, the project seems entirely dead.

    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2
    gitlab.gnome.org/GNOME/libgit2

  30. I was planning on using #libgit2 for my next project, and I wanted a bit of crash course on #git #internals, and came across this absolutely fantastic series of tutorials on #YouTube by Omer Rosenbaum. If you want to dig into git in more detail, these are a great place to start. youtu.be/fWMKue-WBok?si=yygfpQ

  31. I was planning on using #libgit2 for my next project, and I wanted a bit of crash course on #git #internals, and came across this absolutely fantastic series of tutorials on #YouTube by Omer Rosenbaum. If you want to dig into git in more detail, these are a great place to start. youtu.be/fWMKue-WBok?si=yygfpQ

  32. #SwiftGit2 is the #Swift language binding that the #libgit2 project references. For reasons, I wanted SwiftGit2 to be available as a Swift package. I discovered that @mathijsb had made libgit2 itself available as a Swift package and then used that to simplify and clean up SwiftGit2 as a Swift package. This makes consuming SwiftGit2 with up-to-date libgit2 dependencies much easier! Links:

    #SPM support for SwiftGit2: github.com/mbernson/SwiftGit2

    #SPM support for libgit2: github.com/mbernson/libgit2

  33. #SwiftGit2 is the #Swift language binding that the #libgit2 project references. For reasons, I wanted SwiftGit2 to be available as a Swift package. I discovered that @mathijsb had made libgit2 itself available as a Swift package and then used that to simplify and clean up SwiftGit2 as a Swift package. This makes consuming SwiftGit2 with up-to-date libgit2 dependencies much easier! Links:

    #SPM support for SwiftGit2: github.com/mbernson/SwiftGit2

    #SPM support for libgit2: github.com/mbernson/libgit2

  34. #SwiftGit2 is the #Swift language binding that the #libgit2 project references. For reasons, I wanted SwiftGit2 to be available as a Swift package. I discovered that @mathijsb had made libgit2 itself available as a Swift package and then used that to simplify and clean up SwiftGit2 as a Swift package. This makes consuming SwiftGit2 with up-to-date libgit2 dependencies much easier! Links:

    #SPM support for SwiftGit2: github.com/mbernson/SwiftGit2

    #SPM support for libgit2: github.com/mbernson/libgit2

  35. #SwiftGit2 is the #Swift language binding that the #libgit2 project references. For reasons, I wanted SwiftGit2 to be available as a Swift package. I discovered that @mathijsb had made libgit2 itself available as a Swift package and then used that to simplify and clean up SwiftGit2 as a Swift package. This makes consuming SwiftGit2 with up-to-date libgit2 dependencies much easier! Links:

    #SPM support for SwiftGit2: github.com/mbernson/SwiftGit2

    #SPM support for libgit2: github.com/mbernson/libgit2

  36. It's been a while since I've complained about #RustLang itself, so…

    Cargo insists on interacting with #git repositories. At the same time, cargo insists on vendoring an old version of #LibGit2 (1.6.2 FWICS). So, if your system is using a new git version (2.44.0), you won't be able to `cargo build`:

    ```
    error: failed to determine package fingerprint for build script for uv v0.1.38 (/tmp/uv/crates/uv)

    Caused by:
    failed to determine the most recently modified file in /tmp/uv/crates/uv

    Caused by:
    failed to determine list of files in /tmp/uv/crates/uv

    Caused by:
    failed to open git index at /tmp/uv/.git/

    Caused by:
    invalid data in index - calculated checksum does not match expected; class=Index (10)
    ```

    You have to clone everything with `-c index.skipHash=false` to work around this.

    But yeah, I'm sure there's a great benefit to using an outdated vendored C library that NIHs git.

    github.com/rust-lang/cargo/iss

    #Gentoo #NIH

  37. It's been a while since I've complained about #RustLang itself, so…

    Cargo insists on interacting with #git repositories. At the same time, cargo insists on vendoring an old version of #LibGit2 (1.6.2 FWICS). So, if your system is using a new git version (2.44.0), you won't be able to `cargo build`:

    ```
    error: failed to determine package fingerprint for build script for uv v0.1.38 (/tmp/uv/crates/uv)

    Caused by:
    failed to determine the most recently modified file in /tmp/uv/crates/uv

    Caused by:
    failed to determine list of files in /tmp/uv/crates/uv

    Caused by:
    failed to open git index at /tmp/uv/.git/

    Caused by:
    invalid data in index - calculated checksum does not match expected; class=Index (10)
    ```

    You have to clone everything with `-c index.skipHash=false` to work around this.

    But yeah, I'm sure there's a great benefit to using an outdated vendored C library that NIHs git.

    github.com/rust-lang/cargo/iss

    #Gentoo #NIH

  38. It's been a while since I've complained about #RustLang itself, so…

    Cargo insists on interacting with #git repositories. At the same time, cargo insists on vendoring an old version of #LibGit2 (1.6.2 FWICS). So, if your system is using a new git version (2.44.0), you won't be able to `cargo build`:

    ```
    error: failed to determine package fingerprint for build script for uv v0.1.38 (/tmp/uv/crates/uv)

    Caused by:
    failed to determine the most recently modified file in /tmp/uv/crates/uv

    Caused by:
    failed to determine list of files in /tmp/uv/crates/uv

    Caused by:
    failed to open git index at /tmp/uv/.git/

    Caused by:
    invalid data in index - calculated checksum does not match expected; class=Index (10)
    ```

    You have to clone everything with `-c index.skipHash=false` to work around this.

    But yeah, I'm sure there's a great benefit to using an outdated vendored C library that NIHs git.

    github.com/rust-lang/cargo/iss

    #Gentoo #NIH

  39. It's been a while since I've complained about #RustLang itself, so…

    Cargo insists on interacting with #git repositories. At the same time, cargo insists on vendoring an old version of #LibGit2 (1.6.2 FWICS). So, if your system is using a new git version (2.44.0), you won't be able to `cargo build`:

    ```
    error: failed to determine package fingerprint for build script for uv v0.1.38 (/tmp/uv/crates/uv)

    Caused by:
    failed to determine the most recently modified file in /tmp/uv/crates/uv

    Caused by:
    failed to determine list of files in /tmp/uv/crates/uv

    Caused by:
    failed to open git index at /tmp/uv/.git/

    Caused by:
    invalid data in index - calculated checksum does not match expected; class=Index (10)
    ```

    You have to clone everything with `-c index.skipHash=false` to work around this.

    But yeah, I'm sure there's a great benefit to using an outdated vendored C library that NIHs git.

    github.com/rust-lang/cargo/iss

    #Gentoo #NIH

  40. Już dłuższą chwilę nie narzekałem na Rusta samego w sobie, więc…

    Cargo upiera się na interakcji z repozytoriami #git. Jednocześnie, Cargo upiera się przy używaniu włączonej (zamiast systemowej), starej wersji #LibGit2 (z tego, co widzę: 1.6.2). Tak więc, jeżeli wasz system używa nowego gita (2.44.0), to `cargo build` nie przejdzie:

    ```
    error: failed to determine package fingerprint for build script for uv v0.1.38 (/tmp/uv/crates/uv)

    Caused by:
    failed to determine the most recently modified file in /tmp/uv/crates/uv

    Caused by:
    failed to determine list of files in /tmp/uv/crates/uv

    Caused by:
    failed to open git index at /tmp/uv/.git/

    Caused by:
    invalid data in index - calculated checksum does not match expected; class=Index (10)
    ```

    Można to obejść, klonując wszystko z `-c index.skipHash=false`.

    Ale spoko, na pewno jest wielka zaleta z użycia przestarzałej, włączonej na sztywno biblioteki w C, która wynajduje gita na nowo.

    github.com/rust-lang/cargo/iss

    #Gentoo #RustLang

  41. Już dłuższą chwilę nie narzekałem na Rusta samego w sobie, więc…

    Cargo upiera się na interakcji z repozytoriami #git. Jednocześnie, Cargo upiera się przy używaniu włączonej (zamiast systemowej), starej wersji #LibGit2 (z tego, co widzę: 1.6.2). Tak więc, jeżeli wasz system używa nowego gita (2.44.0), to `cargo build` nie przejdzie:

    ```
    error: failed to determine package fingerprint for build script for uv v0.1.38 (/tmp/uv/crates/uv)

    Caused by:
    failed to determine the most recently modified file in /tmp/uv/crates/uv

    Caused by:
    failed to determine list of files in /tmp/uv/crates/uv

    Caused by:
    failed to open git index at /tmp/uv/.git/

    Caused by:
    invalid data in index - calculated checksum does not match expected; class=Index (10)
    ```

    Można to obejść, klonując wszystko z `-c index.skipHash=false`.

    Ale spoko, na pewno jest wielka zaleta z użycia przestarzałej, włączonej na sztywno biblioteki w C, która wynajduje gita na nowo.

    github.com/rust-lang/cargo/iss

    #Gentoo #RustLang

  42. Już dłuższą chwilę nie narzekałem na Rusta samego w sobie, więc…

    Cargo upiera się na interakcji z repozytoriami #git. Jednocześnie, Cargo upiera się przy używaniu włączonej (zamiast systemowej), starej wersji #LibGit2 (z tego, co widzę: 1.6.2). Tak więc, jeżeli wasz system używa nowego gita (2.44.0), to `cargo build` nie przejdzie:

    ```
    error: failed to determine package fingerprint for build script for uv v0.1.38 (/tmp/uv/crates/uv)

    Caused by:
    failed to determine the most recently modified file in /tmp/uv/crates/uv

    Caused by:
    failed to determine list of files in /tmp/uv/crates/uv

    Caused by:
    failed to open git index at /tmp/uv/.git/

    Caused by:
    invalid data in index - calculated checksum does not match expected; class=Index (10)
    ```

    Można to obejść, klonując wszystko z `-c index.skipHash=false`.

    Ale spoko, na pewno jest wielka zaleta z użycia przestarzałej, włączonej na sztywno biblioteki w C, która wynajduje gita na nowo.

    github.com/rust-lang/cargo/iss

    #Gentoo #RustLang

  43. Już dłuższą chwilę nie narzekałem na Rusta samego w sobie, więc…

    Cargo upiera się na interakcji z repozytoriami #git. Jednocześnie, Cargo upiera się przy używaniu włączonej (zamiast systemowej), starej wersji #LibGit2 (z tego, co widzę: 1.6.2). Tak więc, jeżeli wasz system używa nowego gita (2.44.0), to `cargo build` nie przejdzie:

    ```
    error: failed to determine package fingerprint for build script for uv v0.1.38 (/tmp/uv/crates/uv)

    Caused by:
    failed to determine the most recently modified file in /tmp/uv/crates/uv

    Caused by:
    failed to determine list of files in /tmp/uv/crates/uv

    Caused by:
    failed to open git index at /tmp/uv/.git/

    Caused by:
    invalid data in index - calculated checksum does not match expected; class=Index (10)
    ```

    Można to obejść, klonując wszystko z `-c index.skipHash=false`.

    Ale spoko, na pewno jest wielka zaleta z użycia przestarzałej, włączonej na sztywno biblioteki w C, która wynajduje gita na nowo.

    github.com/rust-lang/cargo/iss

    #Gentoo #RustLang

  44. Will ever gain true non-blocking IO using io_uring or alternatives?
    Cannot imagine the resulting speed up over network storage.

  45. Will #libgit2 ever gain true non-blocking #async IO using io_uring or alternatives?
    Cannot imagine the resulting speed up over network storage.

  46. is an object store with immutable data. Hoping to see a custom git ref backend like to store refs in a DB and start using object store backend (instead of compliant filesystems)

  47. #git is an object store with immutable data. Hoping to see a custom git ref backend like #libgit2 to store refs in a DB and start using object store backend (instead of #POSIX compliant filesystems)