What's our policy on reporting bugs for TeX, LaTeX, XeTeX, ConTeXt, LuaTeX, and friends?

  • The OP should do it.
  • The answerer should do it.
  • The person discovering the bug should do it. He might in fact be more familiar with who to report a bug to because he might have a more direct line to the source. This answer motivated this meta question of mine.
  • Is there a repository for bug reports for LaTeX packages? Package authors don't always respond to my bug reports. (In fact, they rarely do. Probably because they're busy. But shouldn't the community be able to track such things? I know this is a question with wider scope ...)
  • Or should we assume that those in charge are beginning, these days, to spend time on this site anyways?

Related:

share
3  
suggestion: this (or a new question) be set up as community wiki so that participants in this forum who are responsible for reacting to bug reports can post the preferred mechanism for making such reports. as don knuth's "tex entomologist" and collector/initial analyst for reports about packages et al. distributed by ams, i'd be happy to have a good place to post this information where it's easy to find by users who need it. –  barbara beeton Feb 28 '13 at 18:43
    
@barbarabeeton I'm fine with that; if we make it this question, everything will be in one place. I can't convert it to CW, but I also won't flag it yet; let's wait for more opinion to trickle in or for others or a moderator to take action. –  Lover of Structure Feb 28 '13 at 19:18
    
Great question. If anyone finds bugs in or has feature requests for any of my LaTeX packages please report them athttps://bitbucket.org/martin_scharrer/<package>/issues. Even I may overlook related questions about my packages. –  Martin Scharrer Mar 1 '13 at 12:50
add comment

1 Answer

If you feel you have identified a bug in any software, the first step is to make sure (as far as you can) it is a bug, and the second is to report it. While several (La)TeX code authors are active on the site, it's not always possible to spot every bug report for your own code. As a package author, I am always pleased if people report bugs via the 'official' channel as well as e-mailing or raising it here. (For me, the 'official' channel is a bug database hosted on BitBucket. Each package author is different, and e-mailing them directly may or may not be encouraged.)

Who should report

In general, I'd encourage the original discovery to log a bug report. However, they do not always do that. So I'd suggest first commenting asking them to log a report, and if that does not happen taking action yourself even if you're 'just visiting' the question.

How bugs are tracked

Each (La)TeX code author is a 'law unto themselves' in terms of bug tracking. Having a single database would not help as simply logging bugs does not get them fixed, and might give a misleading impression. As such, you do have to be prepared to find out how each author wants the information passing on. It's also worth noting that not all (La)TeX code is actively maintained: sometimes there is no good way to report a bug. (Of course, raising the issue here may help as there may be a way to address this.)

share
4  
Addendum: Most package manuals include information on the prefered way of bug reporting, which is not always by mail to the maintainer. –  tohecz Feb 27 '13 at 22:42
3  
About your middle paragraph, here is a possible procedure: 1. Ask the discoverer to report the bug. 2. Wait for an answer. 3. Anyone who encounters such a comment may, if the comment has gone unanswered for at least a week (which we can reasonably define as 8*24h), report the bug himself/herself and then leave a note (as another comment) saying that he/she has used means X to report the bug to person/institution/system Y. What about that? What does our community think? –  Lover of Structure Feb 28 '13 at 0:56
4  
@LoverofStructure You are really lover of structure ;) –  tohecz Feb 28 '13 at 16:39
2  
@LoverofStructure One can overdo the formality: there's nothing wrong with a comment of the form I have logged this as [bug X](link) in the bug tracker for package Y or I have e-mailed the maintainer of package Y to let them know about this. –  Joseph Wright Feb 28 '13 at 20:57
    
Huh, what do you mean? How is that overdoing it? :-) My suggestion is consistent with what you wrote (earlier and here) and in fact an elaboration thereof and just meant to prevent (a) multiple people from sending the same bug report again and again and (b) giving the discoverer a chance to do it first (so that the maintainer will, if he gives credit, credit the right person). –  Lover of Structure Mar 1 '13 at 1:58
1  
Oh, I see. If your concern is my idea to wait for >1 week (because one might forget about it otherwise), the right procedure to handle this would imho be to (a) immediately file a bug report (no matter who) but (b) not forget to attribute the discovery to the actual discoverer (ideally with a link to the relevant TeX.SE post) and (c) leave a comment on TeX.SE telling the community that one has filed a bug report. –  Lover of Structure Mar 1 '13 at 2:07
1  
@LoverofStructure IMHO breif explanation in a bug report you file + a link to the question is enough. Most package authors know what TeX.SX is ;) –  tohecz Mar 1 '13 at 19:15
2  
And leave a comment saying that you filed a bug report :-) –  Lover of Structure Mar 2 '13 at 0:13
add comment

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .