I love seeing these projects make use of the wide-open namespace/references that git provides (outside of the basic `refs/heads` for git branches and `refs/tags` for tags). It looks like they store the data in the `bugs` namespace [1] (so refs/bugs/foo).
Other projects also make use of alternate namespaces. The oft-forgotten built-in "git notes" puts stuff in the `refs/notes/` namespace (specifically in `refs/notes/commits`). Gerrit uses the virtual `refs/for/` namespace for receiving commits for review, stores project config in `refs/meta/config`, and stores User data in `refs/users/` in a special repo [2]. I'm sure others do interesting things.
Alternate uses of git's DAG model are fascinating.
Yes, this really meant to be some sort of framework for storing entities in git, handle the conflicts, and let you buld easily your own tool (or add more features to git-bug).
Technically you could solve collisions through single indirection convention ie 'refs/manifest' which describes refs used with their "schema", "version" etc. If you want to decorate your repo with something, you'd register it there. This way you wouldn't have to keep global convention registry for different projects not to step on each other.
Could you expand on how you see that working? combining these extensions is what I’m guessing but I’d be super interested in hearing how it might work.
If all information for an asynchronous development process could be stored into git and distributed by it, this would be a very good idea for many projects.
The same way, one could add support for code review (aka PRs), todo list, custom entities that your workflow need (say, tracking documentation or custom requirement) ... It can also be entirely outside of the development process.
This would allow for really native linking between a tracked issue and corresponding commits/branches/tags, for modeling dependencies between issues as part of the git DAG,...
While I like the idea of tool consolidation, bug trackers aren't just a tool for the engineers. At most companies I've worked at, the support team, designers, QA team, managers, etc. all use the bug tracker on a daily basis.
It sounds like you can "bridge" to somehow show the tracker outside Engineering, but then you're having to do work around the consolidation, and I'd imagine the result won't be as nice as a full-featured tracker designed for everyone to use.
But, I am curious to hear from someone who has actually used this thing.
Fossil[0] has bug tracking as a standard feature, and through the HTTP role-based authentication, you are able to set up users with different privileges; for instance, being able to read and write the bug tracker without the ability to push new code.
Thanks for that. I kind of wondered how SQLite was relevant to SCM, but I hadn't really bothered to investigate it. Still, having it in a format that can be browsed and reasoned about is handy. Not that Git is any worse in this regard.
Plus, the only thing I need to do to transfer it to a new disk or network location is to move the `.fossil` file, since everything is nicely self-contained.
git-bug has a web ui that you can run on your git server, for example, that can be accessed through a browser.
it's fairly limited in functionality right now (create, comment on, and manage issues), but one of my goals is to refactor it to improve coverage of the existing features, and to add support for things like:
- authenticated access
- unauthenticated/anonymous access (e.g. a public, external contributor/user)
Improved user interfaces can always be added on top of the CLI/library functionality, and that’s the more flexible approach. Everyone can use and/or build their favorite UI, like people do with Git itself.
The monolithic web-first (often web-only) systems are a bit of a modern bane, you’re stuck with whatever user interface the one company/maintainer deems appropriate.
All those users are why bug trackers are annoying. I don't care about those fields "those other people" are demanding, why do I need to fill them out. Mean while they don't care about the fields that are critical for me and don't want to fill them out.
Every job has a part people don't like that's necessary. The company you work for pays you money to fill the fields out, you fill them out, you get paid.
That's why I do it. That doesn't explain why they even need to be there.
For example every project code drop down has this experience: my manager tells me what project code to put everything against, then I always pick the same option. Sometimes I've not been granted access to that option and waste a bunch of time getting that turned on.
At no point was any part of this necessary, because I neither defined the ticket, or could select the project code for myself, but we're all engaged in an elaborate game pretending I had agency over it.
If you're working somewhere competent then every budget code is an indicator that a manager with a budget has said that they are willing to pay for this piece of work to be done. It may even automatically keep track of how much of their budget was spent on you doing the work.
Designers, QA, support and managers are typically already using a VCS, and usually it's Git. And if they don't, this is as good as ever a chance to start.
From QA perspective, which is the one the closest to me: I wanted such a tool for many years, and even though I haven't tried this specific one, I endorse the idea with both of my hands.
In the context of QA, it's always a problem to coordinate between test development, feature development, feature description, and tracking the work progress. All bug trackers I used to date are awful at it. This has a potential to solve at least part of the problem by being physically connected to the work done by either development or QA, or, ideally, both. The holy grail here is the ability to determine what tests to run in response to a commit without making it too much of a burden on the committer.
Tests can easily be the most expensive ongoing activity a development company may undertake. In this context, being able to avoid running useless but expensive tests is very desirable.
Maybe the tool isn't intended for use in commercial environments. There is plenty of work done outside a those environments where this tool might be a better fit. E.g. most free software projects don't have all those other teams interacting with the bug trackers. To put it another way, this seems to fill a similar niche as Github's bug tracker and not Jira.
I think this is made for an audience who find it obvious. "Why would you do it any other way" would be more interesting: There is a weird divide between git supported features on one hand and pull requests/ merge requests/ patch lists/ issues/ bug trackers etc. If everything was supported in git all these features would be interoperable between forges like github, bitbucket and gitea, forgejo.
Not only interoperability but backups, tooling, distributed workflows and everything in between would work consistently and the same way.
That said, I cannot count the times this concept was brought up and tried to make work but despite how much i love the idea in theory, i have yet to see a way it could work in practice.
Some of the issues:
- no universal agreement on exact schema, feature set and workflows, do the competing implementations break each other? if its not interoperable why even bother vs just using an external solution
- how to handle issues not associated to one specific repo or to multiple repos, splitting repos etc.
- how to not confuse devs seeing issue branches or wherever the actual data is stored in the repo
> "Why would you do it any other way" would be more interesting:
That's the interesting question. Normally a bug tracker would basically be a SQL application. When you move it into a Git repo you lose that and now you have to think about how to represent all that relational data in your repository. It gets annoying. This is why for Fossil it's such a trivial thing to do: Fossil repositories _are_ relational and hosted on an RDBMS (SQLite3 or PG). If you don't have a SQL then referential integrity is easy to break (e.g., issues that refer to others that don't know they're being referred to), and querying your issue database becomes a problem as it gets huge because Git doesn't really have an appropriate index for this.
What one might do to alleviate the relational issues is to just not try to maintain referential integrity but instead suck up the issues from Git into a local SQLite3 DB. Then as long as there are no non-fast-forward pushes to the issues DB it's always easy to catch up and have a functional relational database.
1. Fossil repositories are explicitly not relational, they are however stored in SQLite databases. The data model for everything SCM-relevant (that also includes all content like tickets, wiki, forum) is stored as artifacts in a blob table (+ delta), which references other artifacts by hash value, and that provides the referential integrity. That, and the code that handles it. There are relations (via auxiliary tables) to speed up queries, but these tables are transient, get updated by inserting new artifacts, and can be regenerated from the artifacts.
(Users and their metadata, and configuration is not part of this scheme, so these tables might be viewed as relational tables. They are local-only; and not synced.)
That being said, it's the only existent implementation and none other are planned. If you'd like to port it to PostgreSQL, go for it, but it probably would lose a lot of the appeal of being a distributed version control system.
> Normally a bug tracker would basically be a SQL application. When you move it into a Git repo you lose that and now you have to think about how to represent all that relational data in your repository. It gets annoying.
> just not try to maintain referential integrity but instead suck up the issues from Git into a local SQLite3 DB
Applications with SQL backends tackle the problem “how can we express our CRUD datastructures so they fit the relational model?
When you replace SQL with a file-based serialised format, you mainly lose ACID. This is arguably easier than SQL.
You are a single user locally: as long as your IDE plugin doesn’t fight with your terminal client, your need for ACID are low: you handle conflicts like git conflicts, but you may apply domain-specific conflict resolution; while conflicts in source code has no general resolution strategy, merging issue comments can be easy; changing issue status when people agree can be automatically resolved; handling conflicts manually can be tool-assisted.
So losing SQL is not that big of a deal, assuming you’re in a highly decentralised, highly async environment.
The answer to “Why though?” should be forge interop, keeping knowledge in-repo, and because it may fit the organisation (or lack of), e.g. just like you can commit code changes offline, you’re not prevented from updating issues when you’re offline.
I addressed that in a different sub-thread. This one was about "why would you do it any other way?". My commentary about relational data in non-relational media was specifically about why do it any other way. There are great advantages to doing what TFA does -and I acknowledged that in a different sub-thread- but it's still interesting to consider what gets lost, and how to get some of it back.
I don't think that is the main issue. Its not THAT hard to build a secondary sqlite index for data stored in git and keep in in sync especially because the data is already versioned and can be incrementally updated.
It's obvious. It's also not original. There are multiple things like this already. Fossil was the first tool to put bug tracking in the repository. Idk which VCS/forge was the first to put wikis in the repository, but it might have been GitHub or Fossil.
The point here is to be able to work with issues, PRs, and wikis offline just as one is now used to doing with code. And to use the same underlying content-addressed version control tooling for all those things.
honestly cleaning up the Readme and documentation would go a very long way, right now all the information feels fragmented behind all of the little pages. I clicked into the documentation and clicked the first link presented to me on each page and 5 clicks or so in I was on the command line docs but I hadn't seen anything that gave me a high level overview of what git-bug is, what it does, why I want to use it, etc...
I understand that documentation can be hard and you need docs for newbies and long time users, but as a newbie I cannot for the life of me figure out what this is.
This would be amazing as a Magit module for Emacs. I don’t relish the idea of using it in a terminal alongside Emacs while using Git from inside Emacs. Is there a lower-level interface that Magit could provide a porcelain for, maybe?
i have no understanding, beyond that of a lexical nature, what a "magit module" is. i'm a (neo)vim user, and heavy terminal junkie, and if i wanted to build a vim plugin for git-bug, that plugin would likely be shelling out to the command line (as git-bug doesn't expose an independent API today -- it's only started when you start the web ui).
is a "magit module" roughly synonymous with a vim plugin? if so, then would shelling out to the CLI work?
Magit is a Git porcelain for Emacs that implements an interactive TUI. It uses the lower-level APIs Git exposes, I think via the shell but I’m not positive. I’m really just saying if there’s not a primitive API for third-party interactive porcelains to use, that should be part of the design picture! :)
I personally hate how all these platforms like GitHub, GitLab, BitBucket, etc slapped a centralized relational database to manage issues, comments, merge requests, etc next to a distributed de-centralized system like Git.
I especially hate how they've integrated CI/CD into the Git platforms.
I loathe the fact that Microsoft has tied their AI to their Git platform.
I want my CI/CD to be agnostic.
I want my AI to be agnostic.
I want my issues, MRs, comments, etc to be decentralized and come along for the ride when I clone a repo.
The local first approach to dev tools and ecosystems does seem to be on the way out.
The pressures for this aren’t even explicitly corporate interest anymore, a lot of it is driven by non-software-experts who are kind of forced to participate in software dev (e.g. your friendly data science colleague who used to be, say, in material science or astrophysics), which is completely understandable. But more concerneing is the trend of actual software engineers who dislike consoles, terminal programs, and basically don’t believe much in understanding their tools.
You see this all the time with basic stuff from git UIs to kubernetes in IDEs. Productivity isn’t really the issue, although it’s always mentioned as an excuse, there’s just a big appeal to reducing any/every cognitive load no matter what the practical cost is for losing understanding/fluency. To give people the benefit of the doubt though, maybe this pressure is ultimately still corporate though and it started with the call for “full stack” devs, continued with devops/platform engineering etc, where specialists are often actively discouraged. Laziness and a higher tolerance for ignoring details may be a necessary virtue if the market forces everyone to be a generalist.
It’s very weird seeing the coping / seething about a useful tool like this even in HN comments. People have really drunk the proverbial kool-aid / joined the dark side.
> It’s very weird seeing the coping / seething about a useful tool like this even in HN comments. People have really drunk the proverbial kool-aid / joined the dark side.
Maybe because it’s less useful than you think. Companies and large orgs won’t be using it because they like centralized and on the small side your competition is a bunch of TODO files, not a bug tracker.
The fact that nearly all of our source code is not only hosted on proprietary platforms that can (and do) delete it any time they like, but is ALSO integrated with many of our build systems so that it’s not trivially relocatable blows my mind every time I think about it.
If you want to be serious about this, use a self-hosted CI/CD server such as TeamCity, and then mirror all dependencies on a local git server (Go supports GOPROXY variable, but you can also probably fiddle with local DNS and self-signed certs, so that any mention of github.com forwarded you to the local server). This way, it's more controllable: if github.com is down or some repo is deleted, the CI/CD server won't even notice it.
This is really neat! I know Microsoft would never integrate this into github, but what about integrating this other less vendor-lockin-ey scm platforms like forgejo or gitlab? It might be a much better fit then retrofitting activity-pub (current efforts in this space).
Imagine having repository issues seamlessly propagated and replicated across all your git mirrors.
git-bug is built to be portable - today, the way git-bug interacts with other platforms which do not support reading from its namespaces directly is through bridges.
right now, git-bug has built-in bridges for github, gitlab, and jira. i am working on the design for a more modular system in which bridges can be built by anybody and used as "plugins".
really, though, the better, long-term goal is to work with $PLATFORMS to have them update their issue tracker to use git-bug's issues (that is, read from and publish to the refs/bugs namespace using git-bug). there's a bit missing right now to make this easy, but it's something that's very much top of mind as i think about git-bug's future.
I've been yelling 'omg why doesnt someone build a ticketing system on the basis of git, having a separate 'root' (no-parent git commit that is at the bottom of a git tree; technically a git repo can have more than one), with most of the conversation happening in git commit form' - for YEARS.
It'd only really matter if git-bug were to become part of the core Git features. Perhaps one or the other could relicense if that became a desirable outcome.
I have doubt it'll happen. GitHub/GitLab culture is pretty strong, few seem interested in having distributed project management features.
awesome to see a user in the wild! if you weren't aware, you can publish your git-bug issues to the project's issue tracker, assuming that it's on one of the supported bridges today (github, gitlab, jira).
the bridges exist within git-bug to support adoption of the tool and interop with existing platforms.
`git bug bridge pull` and `git bug bridge push` use the bridge's API, and don't attempt to pull from or push to the git remote.
hrm... what's your current workflow like? if there's anything you think git-bug could be doing to make that workflow easier, would you mind hopping in our matrix channel [1], or opening an issue [2]?
I create issues for myself. I sometimes create branches with informal issue keys like `<three letters>-<incrementing number>`. I use `git bug webui` to type edit issues. That’s it.
All I would like as a nice-to-have is support for making those issue keys that I use automatic. For me it’s all local to me so it doesn’t need to be globally unique. But those N-letter prefix does make collision less likely on a single project. I’ve mentioned it here: https://github.com/git-bug/git-bug/issues/75#issuecomment-19...
Very cool project — love the offline-first, Git-native approach. One question:
How do you handle conflict resolution when multiple users modify the same issue or thread concurrently across remotes? Is it purely Git's merge mechanics, or do you apply any domain-specific heuristics?
git-bug embeds a "lamport timestamp" [0] - that is, a logical clock, not a wall clock - in each operation (like the creation of a bug, or a comment, or an edit to a comment). this, combined with the data model [1] we use, allow activity to be recorded and replayed without ever encountering a merge conflict.
This is so cool! Much more elaborated and closer to the real bug tracker than my own tiny project (https://github.com/jhspetersson/git-task) (oh, sorry for the shameless plug)
looks neat! if you're interested in working on this sort of technology, git-bug needs more maintainers! (i also personally wouldn't mind a rust port, and have poked at this in the past).
Love this kind of thing. Particularly in the age of AI being cheap and even on device there are even more reasons to have issues be near the code for easy access.
Similarly things like automatically promoting bugs into the test suite as SHOULD FAIL and so on
This looks cool and has the opportunity to replace my old and trusted todo.txt, but I couldn’t find how to create or resolve bugs. The CLI has features related to syncing but nothing about this, or did I miss something?
It would be nice to see some screenshots of the tool.
But it could be me not looking in the right place.
I tried to find it among the documentation too.
I love this idea and any tool that reduces context switching out of the terminal (e.g. devdocs, fabric, tldr)
1. It may push the design boundaries, but a basic full text index with sqlite would be doable with the web interface. go-doc CLI does the same with the web UI. all indexing is client-side . The bug corpus is very small so indexing should be fast and consistent.
2. Tagging bugs with the observed and resolved git commit ID. that would improve bisect, merge, release management and changelog publication.
I appreciate that tools like Git Bug are not tied to a single repo host but they are tied inherently to a single VCS (git). BE was not reliant on a specific VCS, and it's a *really* simple format on-disk too.
With some motivation you could port git-bug to another VCS without too much problem. You would need to implement those interfaces [1]. The one you care about especially is RepoData, which mainly imply you can store a DAG, have references and push/pull. I believe other VCS (say mercurial) have similar concepts.
Or you could just as well plug a generic database there.
This really seems like an odd thing to make distributed. Do I now have to resolve conflicts in bug conversations? Am I going to find replies magically appearing before mine? The README doesn't even acknowledge that these difficulties might exist.
This sounds like it adds a ton of potential problems and solves some very minor ones:
* You can work offline. Great, but 90% of bug tracking is sending messages to other people so that's not particularly useful.
* You aren't tied to GitHub Issues or whatever. I guess that's good. Seems pretty marginal though.
> Do I now have to resolve conflicts in bug conversations?
> Am I going to find replies magically appearing before mine?
actually, no! git-bug objects embed a lamport timestamp [0] to handle time-based ordering, and actions like comment posting and editing are tracked as "operations", applied in order, and you will never have to deal with a merge conflict.
the data model documentation [1] provides deeper insight into how we handle time, describe why you'll never see a merge conflict, and more. through this post, i've gathered that many people would prefer this sort of documentation be made more visible in the README (instead of "buried" under //doc). the README is probably a bit too high level for a more technical audience, but i appreciate your feedback here, and will take it into consideration as the README is refactored.
In theory it could happen but it's unlikely in practice, for multiple reasons:
- git-bug use a form of logical clock (not wall clock) that order an action in relation to other actions in the repo. Clock drifting doesn't matter.
- pushing to git usually require some access to the repo, and therefore abuse can be dealt with socially (aka you get kicked out)
What can happen for example is someone write a comment, shut down the computer and only push the next day, but in that case the comment showing up before yours is the correct merging.
> pushing to git usually require some access to the repo
Wait, so to comment on an issue I now have to already have push access to that repo? How does that work? E.g. what if I want to comment on a VSCode issue? I'm not a VSCode developer...
Right now, yes, but the idea is to augment the webUI with external auth (e.g. Github OAuth and others) to make it a public portal where anyone can create issues and so on. In that case, the webUI would have access to the git repo, enforce any rules and prevent abuses.
With a single binary deployment, you'd just need a bit of config and a DNS, and you could host a forge-ish for your project.
to support the workflow where you, an individual, outside contributor, want to use git-bug to create or comment on an issue on a third-party platform that you do not control, you would:
- install git-bug
- create a directory (and `git init`), optionally fetch/clone the remote repo (but this is not needed)
- create a git-bug identity (`git bug user new`)
- configure a bridge to (for example, using vscode) github (`git bug bridge new`)
- pull issues from the bridge to your local repository's refs/bugs namespace (`git bug bridge pull`)
- create a new issue, or browse existing ones and comment on them at will
- export your activity to the bridge (`git bug bridge push`)
this works without push access to the repository, because when importing to or exporting from a bridge, the API credentials you provide when configuring the bridge are used -- `git bug bridge {push,pull}` does not push your local `refs/bugs` to the remote.
no - this is what using a lamport timestamp helps to avoid.
edit: re-read your comment and i see what you're getting at.
yes, there is the chance that you don't interact with the remote for X days, and neither does someone else, and when you both finally do, their comment will "magically show up before yours" because in reality they _did_ leave the comment before you.
this is not dissimilar to looking at normal git commits ordered by "author date" vs. "commit date", and seeing "weird date ordering" in a linear tree.
git-bug shows items in "the real order", so in a workflow where you are not fetching frequently, yes, other peoples' activity may be applied before yours when you finally do.
this is just like on a centralized platform like github, where if you are writing a lengthy response or review of a PR, you can end up posting it and requesting changes or approving it after the PR has been merged.
I would think in general yes, but hopefully not like the example you gave because I would expect if you are replying to a comment, your clock would fast-forward past the timestamp of the original comment (NB I haven't looked at OP's implementation).
I think there is multiple cases where repos are mirrored between Codeberg, GitHub and internal and public instances of Gitlab. People want to open issues where they are and they can get accounts. Also I had to migrate issues from one repo to the other. Having wikis moved to git repos is a big advantage over trac and redmine (we just recently moved old projects and it is a pain each time). So I highly welcome anyone who moves issue tracking to git as well.
I love seeing these projects make use of the wide-open namespace/references that git provides (outside of the basic `refs/heads` for git branches and `refs/tags` for tags). It looks like they store the data in the `bugs` namespace [1] (so refs/bugs/foo).
Other projects also make use of alternate namespaces. The oft-forgotten built-in "git notes" puts stuff in the `refs/notes/` namespace (specifically in `refs/notes/commits`). Gerrit uses the virtual `refs/for/` namespace for receiving commits for review, stores project config in `refs/meta/config`, and stores User data in `refs/users/` in a special repo [2]. I'm sure others do interesting things.
Alternate uses of git's DAG model are fascinating.
[1] https://github.com/git-bug/git-bug/blob/bd936650ccf44ca33cf9...
[2] https://gerrit-review.googlesource.com/Documentation/config-...
Yes, this really meant to be some sort of framework for storing entities in git, handle the conflicts, and let you buld easily your own tool (or add more features to git-bug).
See also https://github.com/git-bug/git-bug/blob/master/doc/design/da... and https://github.com/git-bug/git-bug/blob/master/entity/dag/ex...
I'd love to see this used in the wild for other use cases.
Is there a registry of namespaces to avoid collisions?
Technically you could solve collisions through single indirection convention ie 'refs/manifest' which describes refs used with their "schema", "version" etc. If you want to decorate your repo with something, you'd register it there. This way you wouldn't have to keep global convention registry for different projects not to step on each other.
Wow!So this can be used for full blown project management too.
Could you expand on how you see that working? combining these extensions is what I’m guessing but I’d be super interested in hearing how it might work.
If all information for an asynchronous development process could be stored into git and distributed by it, this would be a very good idea for many projects.
So for example, git-bug already has a PR to add support for a project board: https://github.com/git-bug/git-bug/pull/843
The same way, one could add support for code review (aka PRs), todo list, custom entities that your workflow need (say, tracking documentation or custom requirement) ... It can also be entirely outside of the development process.
This would allow for really native linking between a tracked issue and corresponding commits/branches/tags, for modeling dependencies between issues as part of the git DAG,...
Git Bug screenshots:
- TUI recording (GIF): https://github.com/git-bug/git-bug/blob/master/doc/assets/tu...
- Web comments (PNG): https://github.com/git-bug/git-bug/blob/master/doc/assets/we...
- Web feed (PNG): https://github.com/git-bug/git-bug/blob/master/doc/assets/we...
While I like the idea of tool consolidation, bug trackers aren't just a tool for the engineers. At most companies I've worked at, the support team, designers, QA team, managers, etc. all use the bug tracker on a daily basis.
It sounds like you can "bridge" to somehow show the tracker outside Engineering, but then you're having to do work around the consolidation, and I'd imagine the result won't be as nice as a full-featured tracker designed for everyone to use.
But, I am curious to hear from someone who has actually used this thing.
Fossil[0] has bug tracking as a standard feature, and through the HTTP role-based authentication, you are able to set up users with different privileges; for instance, being able to read and write the bug tracker without the ability to push new code.
[0]: https://fossil-scm.org/home/doc/trunk/www/index.wiki
+1 for this. I love having a self-contained, syncable GitHub-lite. It uses SQLite for the format, too, which makes it easy to discover the internals.
It just needs some more 'modern' themes
> It uses SQLite for the format
My chance at sharing Fossil is not Relational. The two times it's been submitted to HN it didn't gain votes.
https://fossil-scm.org/home/doc/trunk/www/fossil-is-not-rela...
Thanks for that. I kind of wondered how SQLite was relevant to SCM, but I hadn't really bothered to investigate it. Still, having it in a format that can be browsed and reasoned about is handy. Not that Git is any worse in this regard.
Plus, the only thing I need to do to transfer it to a new disk or network location is to move the `.fossil` file, since everything is nicely self-contained.
hey! maintainer here.
git-bug has a web ui that you can run on your git server, for example, that can be accessed through a browser.
it's fairly limited in functionality right now (create, comment on, and manage issues), but one of my goals is to refactor it to improve coverage of the existing features, and to add support for things like:
- authenticated access
- unauthenticated/anonymous access (e.g. a public, external contributor/user)
- issue privacy levels
- sprints, projects, report generation
Improved user interfaces can always be added on top of the CLI/library functionality, and that’s the more flexible approach. Everyone can use and/or build their favorite UI, like people do with Git itself.
The monolithic web-first (often web-only) systems are a bit of a modern bane, you’re stuck with whatever user interface the one company/maintainer deems appropriate.
It’s expensive to maintain more front ends than you absolutely require. Web+mobile is one too many already.
This is a bug tracker. What you are describing is much closer to a project management tool, just to make the difference clear.
Those should be tightly integrated, if not the same tool.
It depends on who manages the project. In an open-source project, those will likely also be engineering types, not non-technical managers.
It's hard to make a product that's all things to all people, and it's wise to make a product that has a well-understood, if more narrow, audience.
To some people, engineers without project management doesn't make any sense.
All those users are why bug trackers are annoying. I don't care about those fields "those other people" are demanding, why do I need to fill them out. Mean while they don't care about the fields that are critical for me and don't want to fill them out.
There's no I in team...
Or to put it another way, those other 'useless' fields that take minutes may save the company hours of time in places that you don't see.
> There's no I in team...
But there's U and I in unity!
/s aside, you have a valid point there: just because "I" find the field useless, it doesn't mean that it isn't out to actual good use downstream.
Every job has a part people don't like that's necessary. The company you work for pays you money to fill the fields out, you fill them out, you get paid.
That's why I do it. That doesn't explain why they even need to be there.
For example every project code drop down has this experience: my manager tells me what project code to put everything against, then I always pick the same option. Sometimes I've not been granted access to that option and waste a bunch of time getting that turned on.
At no point was any part of this necessary, because I neither defined the ticket, or could select the project code for myself, but we're all engaged in an elaborate game pretending I had agency over it.
If you're working somewhere competent then every budget code is an indicator that a manager with a budget has said that they are willing to pay for this piece of work to be done. It may even automatically keep track of how much of their budget was spent on you doing the work.
I think that is more useful for communities whose members don't have reliable, always-on networks, rather than workflows within companies.
Designers, QA, support and managers are typically already using a VCS, and usually it's Git. And if they don't, this is as good as ever a chance to start.
From QA perspective, which is the one the closest to me: I wanted such a tool for many years, and even though I haven't tried this specific one, I endorse the idea with both of my hands.
In the context of QA, it's always a problem to coordinate between test development, feature development, feature description, and tracking the work progress. All bug trackers I used to date are awful at it. This has a potential to solve at least part of the problem by being physically connected to the work done by either development or QA, or, ideally, both. The holy grail here is the ability to determine what tests to run in response to a commit without making it too much of a burden on the committer.
Tests can easily be the most expensive ongoing activity a development company may undertake. In this context, being able to avoid running useless but expensive tests is very desirable.
Maybe the tool isn't intended for use in commercial environments. There is plenty of work done outside a those environments where this tool might be a better fit. E.g. most free software projects don't have all those other teams interacting with the bug trackers. To put it another way, this seems to fill a similar niche as Github's bug tracker and not Jira.
P.S. The GitHub readme for this project desperately needs a "Why?" (... would anyone use it, ie. what benefits does it offer vs. say Jira?)
I think this is made for an audience who find it obvious. "Why would you do it any other way" would be more interesting: There is a weird divide between git supported features on one hand and pull requests/ merge requests/ patch lists/ issues/ bug trackers etc. If everything was supported in git all these features would be interoperable between forges like github, bitbucket and gitea, forgejo.
Not only interoperability but backups, tooling, distributed workflows and everything in between would work consistently and the same way.
That said, I cannot count the times this concept was brought up and tried to make work but despite how much i love the idea in theory, i have yet to see a way it could work in practice.
Some of the issues: - no universal agreement on exact schema, feature set and workflows, do the competing implementations break each other? if its not interoperable why even bother vs just using an external solution
- how to handle issues not associated to one specific repo or to multiple repos, splitting repos etc.
- how to not confuse devs seeing issue branches or wherever the actual data is stored in the repo
- how to best make this usable to non devs
The list goes on
> "Why would you do it any other way" would be more interesting:
That's the interesting question. Normally a bug tracker would basically be a SQL application. When you move it into a Git repo you lose that and now you have to think about how to represent all that relational data in your repository. It gets annoying. This is why for Fossil it's such a trivial thing to do: Fossil repositories _are_ relational and hosted on an RDBMS (SQLite3 or PG). If you don't have a SQL then referential integrity is easy to break (e.g., issues that refer to others that don't know they're being referred to), and querying your issue database becomes a problem as it gets huge because Git doesn't really have an appropriate index for this.
What one might do to alleviate the relational issues is to just not try to maintain referential integrity but instead suck up the issues from Git into a local SQLite3 DB. Then as long as there are no non-fast-forward pushes to the issues DB it's always easy to catch up and have a functional relational database.
Two corrections:
1. Fossil repositories are explicitly not relational, they are however stored in SQLite databases. The data model for everything SCM-relevant (that also includes all content like tickets, wiki, forum) is stored as artifacts in a blob table (+ delta), which references other artifacts by hash value, and that provides the referential integrity. That, and the code that handles it. There are relations (via auxiliary tables) to speed up queries, but these tables are transient, get updated by inserting new artifacts, and can be regenerated from the artifacts.
(Users and their metadata, and configuration is not part of this scheme, so these tables might be viewed as relational tables. They are local-only; and not synced.)
See https://fossil-scm.org/home/doc/trunk/www/fossil-is-not-rela... and https://fossil-scm.org/home/doc/tip/www/theory1.wiki for more details.
2. There are no other databases like PostgreSQL to choose from.
I get that the auxiliary tables speed up queries, but the data is relational in nature on some level.
I thought that Fossil did or at least aimed to support different SQL RDBMSes. Did that go away at some point?
Fossil's "file format" is not intimately tied to SQLite: https://fossil-scm.org/home/doc/trunk/www/fileformat.wiki
That being said, it's the only existent implementation and none other are planned. If you'd like to port it to PostgreSQL, go for it, but it probably would lose a lot of the appeal of being a distributed version control system.
> Normally a bug tracker would basically be a SQL application. When you move it into a Git repo you lose that and now you have to think about how to represent all that relational data in your repository. It gets annoying.
> just not try to maintain referential integrity but instead suck up the issues from Git into a local SQLite3 DB
Applications with SQL backends tackle the problem “how can we express our CRUD datastructures so they fit the relational model?
When you replace SQL with a file-based serialised format, you mainly lose ACID. This is arguably easier than SQL.
You are a single user locally: as long as your IDE plugin doesn’t fight with your terminal client, your need for ACID are low: you handle conflicts like git conflicts, but you may apply domain-specific conflict resolution; while conflicts in source code has no general resolution strategy, merging issue comments can be easy; changing issue status when people agree can be automatically resolved; handling conflicts manually can be tool-assisted.
So losing SQL is not that big of a deal, assuming you’re in a highly decentralised, highly async environment.
The answer to “Why though?” should be forge interop, keeping knowledge in-repo, and because it may fit the organisation (or lack of), e.g. just like you can commit code changes offline, you’re not prevented from updating issues when you’re offline.
> The answer to “Why though?”
I addressed that in a different sub-thread. This one was about "why would you do it any other way?". My commentary about relational data in non-relational media was specifically about why do it any other way. There are great advantages to doing what TFA does -and I acknowledged that in a different sub-thread- but it's still interesting to consider what gets lost, and how to get some of it back.
I don't think that is the main issue. Its not THAT hard to build a secondary sqlite index for data stored in git and keep in in sync especially because the data is already versioned and can be incrementally updated.
It's obvious. It's also not original. There are multiple things like this already. Fossil was the first tool to put bug tracking in the repository. Idk which VCS/forge was the first to put wikis in the repository, but it might have been GitHub or Fossil.
The point here is to be able to work with issues, PRs, and wikis offline just as one is now used to doing with code. And to use the same underlying content-addressed version control tooling for all those things.
> Fossil was the first tool to put bug tracking in the repository
I'm pretty sure Aegis did it 15 years before Fossil.
> what benefits does it offer vs. say Jira?
Not being Jira is already a huge benefit. It says offline, local first. Isn't that nice?
Some screenshots would be nice. I found this one [0] of the TUI from 2018, but not much else.
[0] https://github.com/git-bug/git-bug/releases/tag/0.4.0
maintainer here - this is great feedback!
i recently rewrote the README because i felt like its previous iteration was a bit _too_ dense. i may have gone a bit overboard on moving things :)
FWIW, the screenshots you're looking for currently live in: https://github.com/git-bug/git-bug/blob/bd936650ccf44ca33cf9...
honestly cleaning up the Readme and documentation would go a very long way, right now all the information feels fragmented behind all of the little pages. I clicked into the documentation and clicked the first link presented to me on each page and 5 clicks or so in I was on the command line docs but I hadn't seen anything that gave me a high level overview of what git-bug is, what it does, why I want to use it, etc...
I understand that documentation can be hard and you need docs for newbies and long time users, but as a newbie I cannot for the life of me figure out what this is.
This would be amazing as a Magit module for Emacs. I don’t relish the idea of using it in a terminal alongside Emacs while using Git from inside Emacs. Is there a lower-level interface that Magit could provide a porcelain for, maybe?
i have no understanding, beyond that of a lexical nature, what a "magit module" is. i'm a (neo)vim user, and heavy terminal junkie, and if i wanted to build a vim plugin for git-bug, that plugin would likely be shelling out to the command line (as git-bug doesn't expose an independent API today -- it's only started when you start the web ui).
is a "magit module" roughly synonymous with a vim plugin? if so, then would shelling out to the CLI work?
Magit is a Git porcelain for Emacs that implements an interactive TUI. It uses the lower-level APIs Git exposes, I think via the shell but I’m not positive. I’m really just saying if there’s not a primitive API for third-party interactive porcelains to use, that should be part of the design picture! :)
TUI?
EDIT: "rich terminal users interfaces"
> TUI?
Text User Interface
I personally hate how all these platforms like GitHub, GitLab, BitBucket, etc slapped a centralized relational database to manage issues, comments, merge requests, etc next to a distributed de-centralized system like Git.
I especially hate how they've integrated CI/CD into the Git platforms.
I loathe the fact that Microsoft has tied their AI to their Git platform.
I want my CI/CD to be agnostic. I want my AI to be agnostic. I want my issues, MRs, comments, etc to be decentralized and come along for the ride when I clone a repo.
The local first approach to dev tools and ecosystems does seem to be on the way out.
The pressures for this aren’t even explicitly corporate interest anymore, a lot of it is driven by non-software-experts who are kind of forced to participate in software dev (e.g. your friendly data science colleague who used to be, say, in material science or astrophysics), which is completely understandable. But more concerneing is the trend of actual software engineers who dislike consoles, terminal programs, and basically don’t believe much in understanding their tools.
You see this all the time with basic stuff from git UIs to kubernetes in IDEs. Productivity isn’t really the issue, although it’s always mentioned as an excuse, there’s just a big appeal to reducing any/every cognitive load no matter what the practical cost is for losing understanding/fluency. To give people the benefit of the doubt though, maybe this pressure is ultimately still corporate though and it started with the call for “full stack” devs, continued with devops/platform engineering etc, where specialists are often actively discouraged. Laziness and a higher tolerance for ignoring details may be a necessary virtue if the market forces everyone to be a generalist.
Have you tried Fossil? https://fossil-scm.org/home/doc/trunk/www/index.wiki
It’s very weird seeing the coping / seething about a useful tool like this even in HN comments. People have really drunk the proverbial kool-aid / joined the dark side.
> It’s very weird seeing the coping / seething about a useful tool like this even in HN comments. People have really drunk the proverbial kool-aid / joined the dark side.
It’s unclear to me what you mean.
Maybe because it’s less useful than you think. Companies and large orgs won’t be using it because they like centralized and on the small side your competition is a bunch of TODO files, not a bug tracker.
Yes, I understand that Jira exists, thanks.
The question was who wants to use git-bug, not why jira exists.
This is incredibly cool. I love seeing local first software starting to make a comeback. Github is becoming painful to use, even on a fast connection.
The fact that nearly all of our source code is not only hosted on proprietary platforms that can (and do) delete it any time they like, but is ALSO integrated with many of our build systems so that it’s not trivially relocatable blows my mind every time I think about it.
If you want to be serious about this, use a self-hosted CI/CD server such as TeamCity, and then mirror all dependencies on a local git server (Go supports GOPROXY variable, but you can also probably fiddle with local DNS and self-signed certs, so that any mention of github.com forwarded you to the local server). This way, it's more controllable: if github.com is down or some repo is deleted, the CI/CD server won't even notice it.
I know there are workarounds, I’m talking about the general principle or the situation at large.
This is really neat! I know Microsoft would never integrate this into github, but what about integrating this other less vendor-lockin-ey scm platforms like forgejo or gitlab? It might be a much better fit then retrofitting activity-pub (current efforts in this space).
Imagine having repository issues seamlessly propagated and replicated across all your git mirrors.
git-bug is built to be portable - today, the way git-bug interacts with other platforms which do not support reading from its namespaces directly is through bridges.
right now, git-bug has built-in bridges for github, gitlab, and jira. i am working on the design for a more modular system in which bridges can be built by anybody and used as "plugins".
really, though, the better, long-term goal is to work with $PLATFORMS to have them update their issue tracker to use git-bug's issues (that is, read from and publish to the refs/bugs namespace using git-bug). there's a bit missing right now to make this easy, but it's something that's very much top of mind as i think about git-bug's future.
There's also https://github.com/dspinellis/git-issue which has much fewer dependencies.
I've been yelling 'omg why doesnt someone build a ticketing system on the basis of git, having a separate 'root' (no-parent git commit that is at the bottom of a git tree; technically a git repo can have more than one), with most of the conversation happening in git commit form' - for YEARS.
This is wildly exciting.
FWIW, this project made its first release in 2018. =) It's been posted to HN several times, though under its previous project URL, https://github.com/MichaelMure/git-bug (see https://news.ycombinator.com/from?site=github.com/michaelmur...)
you aren't alone! linus thinks we need this, too:
https://youtu.be/sCr_gb8rdEI?t=1533
I had the same idea for a talk a long time ago the result is https://git.esy.fun/yogsototh/gpm
I never used it seriously, but the concepts are easy enough to grasp.
Have a separate dedicated branch that contains all the metadata of your project. Like issues, todo-list, review comments, etc...
Hopefully, something decentralised like this could become more popular maybe.
Interesting that GPL3 was selected for this, while Git itself is GPL2 (these are incompatible licenses)
It'd only really matter if git-bug were to become part of the core Git features. Perhaps one or the other could relicense if that became a desirable outcome.
I have doubt it'll happen. GitHub/GitLab culture is pretty strong, few seem interested in having distributed project management features.
I use this to note potential issues in my copy of the Git project. For outright bugs though I report them to the project.
awesome to see a user in the wild! if you weren't aware, you can publish your git-bug issues to the project's issue tracker, assuming that it's on one of the supported bridges today (github, gitlab, jira).
the bridges exist within git-bug to support adoption of the tool and interop with existing platforms.
`git bug bridge pull` and `git bug bridge push` use the bridge's API, and don't attempt to pull from or push to the git remote.
Thanks. I can’t push it to the issue tracker since they don’t have one. Bug reports go to the mailing list.
hrm... what's your current workflow like? if there's anything you think git-bug could be doing to make that workflow easier, would you mind hopping in our matrix channel [1], or opening an issue [2]?
[1]: https://matrix.to/#/#git-bug:matrix.org
[2]: https://github.com/git-bug/git-bug/issues
I create issues for myself. I sometimes create branches with informal issue keys like `<three letters>-<incrementing number>`. I use `git bug webui` to type edit issues. That’s it.
All I would like as a nice-to-have is support for making those issue keys that I use automatic. For me it’s all local to me so it doesn’t need to be globally unique. But those N-letter prefix does make collision less likely on a single project. I’ve mentioned it here: https://github.com/git-bug/git-bug/issues/75#issuecomment-19...
Funny.
Some days ago, on the Firefox mover to github topic, people were wondering if issue trackers should also be distributed.
Seems an interesting idea.
Very cool project — love the offline-first, Git-native approach. One question:
How do you handle conflict resolution when multiple users modify the same issue or thread concurrently across remotes? Is it purely Git's merge mechanics, or do you apply any domain-specific heuristics?
maintainer here. great question!
git-bug embeds a "lamport timestamp" [0] - that is, a logical clock, not a wall clock - in each operation (like the creation of a bug, or a comment, or an edit to a comment). this, combined with the data model [1] we use, allow activity to be recorded and replayed without ever encountering a merge conflict.
[0]: https://en.wikipedia.org/wiki/Lamport_timestamp
[1]: https://github.com/git-bug/git-bug/blob/master/doc/design/da...
This is so cool! Much more elaborated and closer to the real bug tracker than my own tiny project (https://github.com/jhspetersson/git-task) (oh, sorry for the shameless plug)
looks neat! if you're interested in working on this sort of technology, git-bug needs more maintainers! (i also personally wouldn't mind a rust port, and have poked at this in the past).
Love this kind of thing. Particularly in the age of AI being cheap and even on device there are even more reasons to have issues be near the code for easy access.
Similarly things like automatically promoting bugs into the test suite as SHOULD FAIL and so on
This looks cool and has the opportunity to replace my old and trusted todo.txt, but I couldn’t find how to create or resolve bugs. The CLI has features related to syncing but nothing about this, or did I miss something?
If you're looking at the CLI there are specific man pages for it.
You can find creation here https://github.com/git-bug/git-bug/blob/master/doc%2Fmd%2Fgi...
And status update here https://github.com/git-bug/git-bug/blob/master/doc%2Fmd%2Fgi...
Yeah it was right in front of me and I missed it. Thanks a bunch!
It would be nice to see some screenshots of the tool. But it could be me not looking in the right place. I tried to find it among the documentation too.
Is there anybody using a sub directory in the fit repo with one markdown per ticket?
Docs seem to miss instructions on how you add a bug! Did I miss it?
https://github.com/git-bug/git-bug/blob/master/doc/md/git-bu...
Might be interesting as a personal cross project todo tracker?
I love this idea and any tool that reduces context switching out of the terminal (e.g. devdocs, fabric, tldr)
1. It may push the design boundaries, but a basic full text index with sqlite would be doable with the web interface. go-doc CLI does the same with the web UI. all indexing is client-side . The bug corpus is very small so indexing should be fast and consistent.
2. Tagging bugs with the observed and resolved git commit ID. that would improve bisect, merge, release management and changelog publication.
It's a little disappointing that BugsEverywhere/BE (https://github.com/aaiyer/bugseverywhere) never gained popularity.
I appreciate that tools like Git Bug are not tied to a single repo host but they are tied inherently to a single VCS (git). BE was not reliant on a specific VCS, and it's a *really* simple format on-disk too.
With some motivation you could port git-bug to another VCS without too much problem. You would need to implement those interfaces [1]. The one you care about especially is RepoData, which mainly imply you can store a DAG, have references and push/pull. I believe other VCS (say mercurial) have similar concepts.
Or you could just as well plug a generic database there.
[1]: https://github.com/git-bug/git-bug/blob/master/repository/re...
Excited to try this. I like the look of that TUI.
This really seems like an odd thing to make distributed. Do I now have to resolve conflicts in bug conversations? Am I going to find replies magically appearing before mine? The README doesn't even acknowledge that these difficulties might exist.
This sounds like it adds a ton of potential problems and solves some very minor ones:
* You can work offline. Great, but 90% of bug tracking is sending messages to other people so that's not particularly useful.
* You aren't tied to GitHub Issues or whatever. I guess that's good. Seems pretty marginal though.
hey, maintainer here!
> Do I now have to resolve conflicts in bug conversations? > Am I going to find replies magically appearing before mine?
actually, no! git-bug objects embed a lamport timestamp [0] to handle time-based ordering, and actions like comment posting and editing are tracked as "operations", applied in order, and you will never have to deal with a merge conflict.
the data model documentation [1] provides deeper insight into how we handle time, describe why you'll never see a merge conflict, and more. through this post, i've gathered that many people would prefer this sort of documentation be made more visible in the README (instead of "buried" under //doc). the README is probably a bit too high level for a more technical audience, but i appreciate your feedback here, and will take it into consideration as the README is refactored.
[0]: https://en.wikipedia.org/wiki/Lamport_timestamp [1]: https://github.com/git-bug/git-bug/blob/bd936650ccf44ca33cf9...
Interesting, thanks. This must be true though right?
> Am I going to find replies magically appearing before mine?
In theory it could happen but it's unlikely in practice, for multiple reasons:
- git-bug use a form of logical clock (not wall clock) that order an action in relation to other actions in the repo. Clock drifting doesn't matter.
- pushing to git usually require some access to the repo, and therefore abuse can be dealt with socially (aka you get kicked out)
What can happen for example is someone write a comment, shut down the computer and only push the next day, but in that case the comment showing up before yours is the correct merging.
> pushing to git usually require some access to the repo
Wait, so to comment on an issue I now have to already have push access to that repo? How does that work? E.g. what if I want to comment on a VSCode issue? I'm not a VSCode developer...
Right now, yes, but the idea is to augment the webUI with external auth (e.g. Github OAuth and others) to make it a public portal where anyone can create issues and so on. In that case, the webUI would have access to the git repo, enforce any rules and prevent abuses.
With a single binary deployment, you'd just need a bit of config and a DNS, and you could host a forge-ish for your project.
We are not there yet but it's really not far.
to support the workflow where you, an individual, outside contributor, want to use git-bug to create or comment on an issue on a third-party platform that you do not control, you would:
- install git-bug
- create a directory (and `git init`), optionally fetch/clone the remote repo (but this is not needed)
- create a git-bug identity (`git bug user new`)
- configure a bridge to (for example, using vscode) github (`git bug bridge new`)
- pull issues from the bridge to your local repository's refs/bugs namespace (`git bug bridge pull`)
- create a new issue, or browse existing ones and comment on them at will
- export your activity to the bridge (`git bug bridge push`)
this works without push access to the repository, because when importing to or exporting from a bridge, the API credentials you provide when configuring the bridge are used -- `git bug bridge {push,pull}` does not push your local `refs/bugs` to the remote.
no - this is what using a lamport timestamp helps to avoid.
edit: re-read your comment and i see what you're getting at.
yes, there is the chance that you don't interact with the remote for X days, and neither does someone else, and when you both finally do, their comment will "magically show up before yours" because in reality they _did_ leave the comment before you.
this is not dissimilar to looking at normal git commits ordered by "author date" vs. "commit date", and seeing "weird date ordering" in a linear tree.
git-bug shows items in "the real order", so in a workflow where you are not fetching frequently, yes, other peoples' activity may be applied before yours when you finally do.
this is just like on a centralized platform like github, where if you are writing a lengthy response or review of a PR, you can end up posting it and requesting changes or approving it after the PR has been merged.
I would think in general yes, but hopefully not like the example you gave because I would expect if you are replying to a comment, your clock would fast-forward past the timestamp of the original comment (NB I haven't looked at OP's implementation).
I think there is multiple cases where repos are mirrored between Codeberg, GitHub and internal and public instances of Gitlab. People want to open issues where they are and they can get accounts. Also I had to migrate issues from one repo to the other. Having wikis moved to git repos is a big advantage over trac and redmine (we just recently moved old projects and it is a pain each time). So I highly welcome anyone who moves issue tracking to git as well.
Just the fact that I can use it without a website is a game changer.