Closed Bug 143604 Opened 23 years ago Closed 23 years ago

format=html not supported

Categories

(Bugzilla :: Bugzilla-General, defect, P3)

2.15
x86
Linux
defect

Tracking

()

RESOLVED FIXED
Bugzilla 2.18

People

(Reporter: bbaetz, Assigned: myk)

References

()

Details

Since our files aren't called foo-html.html.tmpl, a format of 'html' fails, giving the error: "The html output format is not supported by this script. Supported formats (besides the default HTML format) are simple, rdf." This error is silly, as we do. If html really is the default, then we should check for html earlier.
This error is correct, but misleading. When it says "the default HTML format" it should really say "the DEFAULT (html) format", if you see what I mean. In other words, the point is that it's default, not that it's HTML - the default format isn't always HTML - or rather, it shouldn't be. That's the problem. I'm probably having difficulty explaining this coherently because it's 20 to 1 in the morning... Gerv
Yes, but my point is that if the default format is html, then format=html should produce html. If the default was foobar, then format=foobar should produce foobar. I will, however, point out that globals.pl does assume that the default format is html, with the different style names: # If the format name is undefined or the default format is specified, # do not do any validation but instead return the default format. if (!defined($format) || $format eq "default") { return { 'template' => "$script.html.tmpl" , 'extension' => "html" , 'contenttype' => "text/html" , }; }
The default format (for all templates implemented with multiple formats) is not "html"; it is unnamed. "foo.html.tmpl" does not implement the "html" format, it implements an unnamed default format whose content is of the "text/html" type. Nevertheless, it is true that not naming our default formats causes problems. In particular, it forces us to make filename changes when changing the default format for a given script, which is likely to happen as we redesign pages in the future while keeping older versions of those pages around, as well as when multiple equivalent formats exist for installations to choose from. I recommend you morph this bug into "default formats are nameless".
The error message is ambiguous, then, though. Besides, if I ask for a format of html, then I should get it. The fact that this format happens to be the default one is irrelevent. I want html, we have html, so....
>The error message is ambiguous, then, though. Agreed. The error message should explain the difference between format name and content type (but see my suggestion below). >Besides, if I ask for a format of html, then I should get it. Not if it doesn't exist, but a case could be made that if a format with the given name doesn't exist, but at least one format exists whose file extension matches the given name, then one of those formats (not sure how to pick which one) should be used. I think this is maybe what you are asking for. > The fact that this >format happens to be the default one is irrelevent. Agreed, and the solution to that is to name all formats, even the default one, so users can request default formats by name. This bug, as currently defined, is "worksforme", so it should be morphed into one of these other bugs (I recommend the third as the bug closest to what you want): "Format not found" error message does not adequately explain difference between format and content type Default formats are unnamed Format-determining code should fall back on file extension when no format with given name is found
Isn't empty string a valid name?
Priority: -- → P3
Target Milestone: --- → Bugzilla 2.18
>Isn't empty string a valid name? Not in any naming scheme I've ever heard of.
> Agreed, and the solution to that is to name all formats, even the default one, > so users can request default formats by name. I'd rather not do this - it would make all our template names 5 chars longer, and some of them are already verbose. Also, we'd have a lot of templates called sometemplate-html.html.tmpl , and while we know this isn't redundant, it does look it. I prefer the other solution: if someone asks for "format=foo", and we have no sometemplate-foo.contenttype.tmpl, then we try sometemplate.foo.tmpl before giving up. Gerv
Take a look at bug 146945. Not only does it remove the redundancy, it resolves the format/content-type ambiguity and provides the requested format name -> file extension gracefulness.
Blocks: 150049
Fixed by bug 146945 - after the split of formats and ctypes, ctype=html now does a no-op. Gerv
So it's fixed, then :-) Gerv
Status: NEW → RESOLVED
Closed: 23 years ago
Resolution: --- → FIXED
QA Contact: matty_is_a_geek → default-qa
You need to log in before you can comment on or make changes to this bug.