Bug reporting guidelines

From ArchWiki
(Redirected from Reporting bug guidelines)

Opening (and closing) bug reports on the Arch Linux bug tracker in Gitlab is one of many possible ways to help the community. However, poorly-formed bug reports can be counter-productive. When bugs are incorrectly reported, developers waste time investigating and closing invalid reports. This document will guide anyone wanting to help the community by efficiently reporting and hunting bugs.

See also: How to Report Bugs Effectively by Simon Tatham

Before reporting

Preparing a detailed and well-formed bug report is not difficult, but requires an effort on behalf of the reporter. The work done before reporting a bug is arguably the most useful part of the job. Unfortunately, few people take the time to do this work properly.

The following steps will guide you in preparing your bug report or feature request.

Search for duplicates

If you encounter a problem or desire a new feature, there is a high probability that someone else already had this problem or idea. If so, an appropriate bug report may already exist. In this case, please do not create a duplicate report; see #Following up on bugs instead.

Search thoroughly for existing information, including:

  • Arch Linux Forums: The forums are often the first stop for users looking for help or sharing ideas. While a solution may not yet exist, additional background information and discussion can steer you in the right direction.
  • Arch Linux bug tracker: Your problem may have already been reported to Arch Linux developers. Duplicate bug reports are unhelpful and promptly closed. Search both closed bugs as well as open reports by choosing 'All Statuses' under Advanced. Remember to mark 'search details' and/or 'search comments' under Advanced, the bug title may not contain the text you are searching for.
  • Google or your favorite search engine: Search using the program's name, version, and a relevant part of the error message, if any.
  • Upstream forum, mailing list and bug tracker: If Arch Linux is not responsible for a bug, it should be reported upstream rather than the Arch Linux bug tracker. Search both recently closed bugs as well as open reports. Bugs may have already been fixed in the program's development version.
  • Other distribution forums: The free software community is vast; Archers are not the only users with ideas! Consider searching the Gentoo Forums, FedoraForum.org, and Ubuntu Forums, for example.

Upstream or Arch?

Arch Linux is a GNU/Linux distribution. Arch developers and Trusted Users are responsible for compiling, packaging, and distributing software from a wide range of sources. Upstream refers to these sources – the original authors or maintainers of software that is distributed in Arch Linux. For example, the popular Firefox web browser is developed by the Mozilla Project.

If Arch is not responsible for a bug, the problem will not be solved by reporting the bug to Arch developers. Responsibility for a bug is said to lie upstream when it is not caused through the distribution's porting and integration efforts.

By reporting bugs upstream, you will not only help Archers and Arch developers, but you will also help other people in the free software community as the solution will trickle down to all distributions.

Once you have reported a bug upstream or have found other relevant information from upstream, it might be helpful to post this in the Arch bug tracker, so both Arch developers and users are made aware of it.

So what is Arch Linux responsible for?

  • Arch Linux Projects: pacman, AUR, mkinitcpio, Arch Websites. If you have a doubt about if the project belongs to Arch or not, display the package information (pacman -Qi package_name or using the website) and look at the upstream URL.
    Note: Report issues for Arch Linux software projects in the project's upstream issue tracker on gitlab.archlinux.org or GitHub instead of the respective package issues.
  • Packaging: Packaging basically consists of fetching the source code from upstream, compiling it with relevant options, making sure that it will be correctly installed on an Arch system, and checking that the main functionality works. Packaging at Arch does not consist of adding new functionality or patches for existing bugs; this is the job of the upstream developer.

If a bug/feature is not under Arch's responsibility, report it upstream. See also #Reasons for not being a bug.

Bug or feature?

bug
something that should work but does not work, contrary to the developer's intentions.
feature
something which software does or would do if somebody coded it.

Reasons for not being a bug

  • Something you would like a piece of software to do, which is not implemented by the upstream developers. In short: a new feature.
  • A bug already reported upstream.
  • A bug already fixed upstream but not in Arch because the package is not up-to-date.
  • A package which is not-up-to-date. Use the Flag Package Out-of-Date feature on Arch's packages website.
  • A package which does not use Fedora, Ubuntu or some other community patch. Patches should be submitted upstream.
  • A package where non essential function X or function Y is not activated. This is a feature request.
  • A package which does not include a .desktop file or icons or other freedesktop stuff. This is not a bug if such files are not included in the source tarball, and this must be requested as a feature request upstream. If such files are provided by upstream but not used in the package then this is a bug.

Reasons for not being a feature

  • When it is a bug...
  • When it is not under Arch responsibility to implement the feature, i.e. an upstream feature.
  • A package is not up-to-date. Use the Flag Package Out-of-Date feature on Arch's packages website.
  • A package which does not use Fedora, Ubuntu or some other community patch. Patches should be submitted upstream.

Gather useful information

Here is a list of useful information that should be mentioned in your bug report:

  • Version of the package being used. Always specify package version. Saying "the latest", "todays", or "the package in extra" have absolutely no meaning. Especially if the bug is not about to get fixed right away.
  • Version of the main libraries used by the package (available in the depends variable in the PKGBUILD), when they are involved in the problem. If you do not know exactly what information to provide then wait for a bug hunter to ask you for it...
  • Version of the kernel used if you are having hardware related problems.
  • Indicate whether or not the functionality worked at one time or not. If so, indicate since when it stopped working.
  • Indicate your hardware brand when you are having hardware related problems
  • Add relevant log information when any is available. This can be obtained in the following places depending on the problem:
    • The systemd journal. If using syslog-ng, /var/log/messages contains logs related to kernel and hardware related issues.
    • ~/.local/share/xorg/Xorg.0.log or /var/log/Xorg.0.log or /var/log/Xorg.2.log or any Xorg like log files if video related problems (nvidia, ati, xorg...)
    • Run your program in a console and use verbose and/or debug mode if available (see your program documentation), and copy the output in a file. When running an application in a terminal make sure relevant information will be displayed in English so that many people can understand it. This can be done by using export LC_ALL=C.UTF-8. Example with a software named foobar from which you would like to have relevant information at runtime and provided that foobar has a --verbose option:
      $ export LC_ALL=C.UTF-8
      $ foobar --verbose
      
      This will affect only the current terminal and will stop taking effect when the terminal is closed.

If you have to paste a lot of text, like the output of dmesg, or an Xorg log, is it preferred to save it as a file on your computer and attach it to your bug report. Attaching a file rather than using a pastebin to present relevant information is preferable in general due to the fact that pastebined content may suffer by expired links or any other potential problems. Attaching a file guarantees the provided information will always be available.

  • Indicate how to reproduce the bug. This is very important, it will help people test the bug and potential patches on their own computer.
  • The stack trace. It is a list of calls made by the program during its execution, and helps in finding part of the program where the bug is located, especially if bug involves the program crashing. You can obtain a stack trace using gdb (The GNU Debugger) as explained in Debugging/Getting traces#Getting the trace.

Opening a bug

When you are sure it is a bug or a feature and you gathered all useful information, then you are ready to open a bug report or feature request.

Creating an account

You have to create an account on Arch's GitLab, which manages its login via Arch Linux SSO.

Note: Due to an influx of spam, we have had to temporarily disable account registrations there. Please send an email to accountsupport@archlinux.org, with your desired username, if you want to get access.

Where to open the bug

Once you have determined your feature or bug is related to Arch and not an upstream issue, you will need to file your problem in the correct project. This is most easily done via Add new bug on the respective packages page on archlinux.org.

You can alternatively also get to the correct page using pkgctl from devtools:

$ pkgctl repo web pkgname

Problems with packages in the AUR are not reported in the bug tracker. The AUR allows you to add comments to a package, which you should use to report bugs to the package maintainer.

Title

Please write a concise and descriptive title for your bug/feature request.

Here is a list of recommendations:

  • Do not name your report "pkgname is broken after the last update" - it is non-descriptive and "after last update" has no meaning in Arch.
  • Do not write too much text in the title. Excessive text will not be visible in reports list.

Severity

This article or section is out of date.

Reason: Users themselves can't assign labels to GitLab issues. (Discuss in Talk:Bug reporting guidelines)

Choosing a critical severity will not help to solve the bug faster. It will only make truly critical problems less visible and probably make the developer assigned to your bug a bit less open to fixing it.

Here is a general usage of severities:

  • Critical - System crash, severe boot failure that is likely to affect more than just you, or an exploitable security issue in either a core or outward-facing service package.
  • High - The main functionality of the application does not work, less critical security issues, etc.
  • Medium - A non-essential functionality does not work, UNIX standards not respected, etc.
  • Low - An optional functionality (plugin or compilation activated) does not work.
  • Very Low - Translation or documentation mistake. Something that really does not matter much but should be noticed for a future release.

Including relevant information

This is maybe one of the most difficult parts of bug reporting. You have to choose from the section #Gather useful information which information you will add to your bug report. This will depend on which your problem is. If you do not know what the relevant pieces of information are, do not be shy: it is better to give more information than needed than not enough.

A good tutorial on reporting bugs can be found at https://www.chiark.greenend.org.uk/~sgtatham/bugs.html.

However, developers or bug hunters will ask you for more information if needed. Fortunately after a few bug reports you will know what information should be given.

Short information can be inlined in your bug report, whereas long information (logs, screenshots...) should be attached.

Following up on bugs

Do not think the work is done once you have reported a bug!

Developers or other users will probably ask you for more details and give you some potential fixes to try. Without continued feedback, bugs cannot be closed and the end result is both angry users and developers.

Voting and Watching

You can vote for your favourite bugs via reactions on the issue. The number of reactions indicates to the developers how many people are impacted by the bug without creating too much noise. However, this is not a very effective way of getting the bug solved. Much more important would be posting any additional information you know about the bug if you were not the original reporter.

Watching a bug is important: you will receive an email when new comments are added or the bug status has changed. This can be done via the "..." menu in the upper right corner by toggling the Notification switch there. If you opened a issue or commented on it you will automatically be subscribed to changes.

Answering additional information requests

People will take the time to look at your bug report and will try to help you. You need to continue to help them resolve your bug. Not answering to their questions will keep your bug unresolved and likely hamper enthusiasm to fix it.

Please take the time to give people more information if requested and try the solutions proposed.

Developers or bug hunters will close your bug if you do not answer questions after a few weeks or a month.

Updating the bug report when a new package version is out

Sometimes, a bug is only present in certain version(s) of a given package, and the bug is fixed in a new version of the package. If this is the case, say so in the bug report comments, and request that the bug be closed.

Closing when solved

Sometimes people report a bug but do not notify when they have solved it on their own, leaving people searching for a solution that has already been found. Please close the bug if you found a solution, and give the solution in the bug report comments.

Bug status

During its life, a bug may go through several states:

  • Unconfirmed - This is the default state. You have just reported it and nobody managed to reproduce the problem or confirmed that it is actually a bug.
  • New - The bug is confirmed but it has not been assigned to the developer responsible for the related software. It is usually the case when more investigation is needed to determine which software is responsible for the bug.
  • Assigned - The bug has been assigned to a developer responsible for the software involved in the bug. It does not mean that the developer will be the one who will fix the bug. It does not even mean that the developer will work on a solution. It just means that the developer will take care of the life cycle of the bug, including reviewing patches if any, releasing a fix and closing the bug when required. It is unwise to contact a developer directly and to ask them to fix it more quickly; they will certainly not like this.
  • Researching - Somebody is looking for a solution. This status is rarely used at Arch, for good reason. The researching status could make people believe they do not need to get interested in the bug report. But usually we need more than one person to fix a bug: having several experienced people on a bug helps a lot.
  • Waiting on Response and Requires testing - The one who reported a bug has been asked to provide more information or to try a proposed solution, but they did not reply yet. These statuses are rarely used at Arch, but should be used more often. Still, it is important that you watch the bug (see #Voting and Watching), as developers or bug hunters usually ask questions in the comments.
  • A task closure has been requested - This is not exactly a status, but you may find some bug reports with such a notification. This indicates that somebody requested a closure for the bug. A reason is added to the request most of the time. It is upon the assignee developer to decide whether they will accept the closure or not.
  • Closed - Either this is not a valid bug (see #Reasons for not being a bug) or a solution has been found and released.

The Bug Wranglers are responsible for dispatching bugs and setting their status together with the help of the respective maintainers of the package the bug was opened against.