Hi Dave, Redirecting non-topical Savannah ticket comment to the general discussion list.
At 2026-04-29T19:59:57-0400, Dave wrote: > Follow-up Comment #14, bug #68207 (group groff): > > The following is a bug-maintenance discussion rather than anything to > do with this bug report, but I'm posting it here since the inciting > comment was posted here. Understandable, but I think "philosophical" discussions of bug tracker management are better carried out on the general discussion list, so that more people are informed and can share their views. > [comment #11 comment #11:] > > I use the tag labels to locate the relevant code in the source tree. > > Isn't that effectively a circular reference? Do you mean a redundant one? Not necessarily. For example, we have a "Category" called "Core"... https://savannah.gnu.org/bugs/admin/field_values.php?group_id=273&field=category_id&list_value=1 ...but I routinely prefix ticket summaries with "[troff]", "[nroff]", or "[libgroff]" for greater specificity. > The ticket identified a problem with pdfmom, but when you tagged it > with "[gropdf]", you already had to know (or look up) where in the > source tree pdfmom lived in order to add the correct tag, which (per > the above) then tells you where to find pdfmom. Yes. That curio supports Deri's proposal to migrate pdfmom elsewhere in the source tree. > I bring this up because when I open a bug, I'm sometimes unclear > whether to tag something, say, "[troff]" or "[libgroff]"; it sounds > now like the correct tag can't be definitively assigned until a bug is > root-caused, Sure, maybe not _definitively_. If someone lacks confidence in which tag to annotate a ticket with, that tag is best omitted! Doing so will imply a couple of possible truths: that a bug is multi-sited (in which case the best Category is likely "General", or that the work necessary to resolve the ticket is poorly understood. I think those are good bits of information to communicate to people. Further, neither a ticket's Category nor its Summary are immutable. If we guess wrong, we can update them. In my experience, re-Categorizing and re-tagging of tickets due to a wrong guess is uncommon, and thus not burdensome. We're humans. We get to change our minds. > at which point it seems less than useful as a source-tree pointer. See above for why I think my tag annotations can be _more_ clarifying. > It has seemed to me before that most of these tags are redundant with > the Category anyway, and comment #13 only reinforces that notion. Have you browsed our Savannah ticket tracker in its default configuration, using the default, "Basic" query form? The "Browse Items" pages that result lack columns for "Category". (They lack "Item Group", too.) As I recall, I started adding these tags to tickets before I devised the "Backlog" query form option that is my personal default _when I'm logged in to Savannah_. But I only log in to Savannah, as a rule, from my desktop machine. When I browse from my tablet or other device, I'm not logged in, so I get the default query form, and the Category datum does not show. It's been years since I started the tagging practice, but if I reconstruct my reasoning accurately, I wanted casual users of the Savannah ticket tracker to be able to get an idea of what our tickets "had to do with", since groff is a complex system (has several components)--the summary without such tagging can speak only in jargon that they might have trouble situating even if they understand it. For example, all of troff, pic, and eqn have independent macro expansion facilities. For that matter, as I recently noted in a thread with Karl Berry, mdoc kind of does too, albeit not one _wholly_ independent of troff's. Does the foregoing illuminate/justify things adequately? Regards, Branden
signature.asc
Description: PGP signature
