3 The available [[ikiwiki/pagespec/sorting]] methods are currently hard-coded in
4 IkiWiki.pm, making it difficult to add any extra sorting mechanisms. I've
5 prepared a branch which adds 'sort' as a hook type and uses it to implement a
6 new `meta_title` sort type.
8 Someone could use this hook to make `\[[!inline sort=title]]` prefer the meta
9 title over the page name, but for compatibility, I'm not going to (I do wonder
10 whether it would be worth making sort=name an alias for the current sort=title,
11 and changing the meaning of sort=title in 4.0, though).
13 > What compatability concerns, exactly, are there that prevent making that
14 > change now? --[[Joey]]
16 *[sort-hooks branch now withdrawn in favour of sort-package --s]*
18 I briefly tried to turn *all* the current sort types into hook functions, and
19 have some of them pre-registered, but decided that probably wasn't a good idea.
20 That earlier version of the branch is also available for comparison:
22 *[also withdrawn in favour of sort-package --s]*
24 >> I wonder if IkiWiki would benefit from the concept of a "sortspec", like a [[ikiwiki/PageSpec]] but dedicated to sorting lists of pages rather than defining lists of pages? Rather than defining a sort-hook, define a SortSpec class, and enable people to add their own sort methods as functions defined inside that class, similarly to the way they can add their own pagespec definitions. --[[KathrynAndersen]]
26 >>> [[!template id=gitbranch branch=smcv/ready/sort-package author="[[Simon_McVittie|smcv]]"]]
27 >>> I'd be inclined to think that's overkill, but it wasn't very hard to
28 >>> implement, and in a way is more elegant. I set it up so sort mechanisms
29 >>> share the `IkiWiki::PageSpec` package, but with a `cmp_` prefix. Gitweb:
30 >>> <http://git.pseudorandom.co.uk/smcv/ikiwiki.git?a=shortlog;h=refs/heads/sort-package>
32 >>>> I agree it seems more elegant, so I have focused on it.
34 >>>> I don't know about reusing `IkiWiki::PageSpec` for this.
37 >>>>> Fair enough, `IkiWiki::SortSpec::cmp_foo` would be just
38 >>>>> as easy, or `IkiWiki::Sorting::cmp_foo` if you don't like
39 >>>>> introducing "sort spec" in the API. I took a cue from
40 >>>>> [[ikiwiki/pagespec/sorting]] being a subpage of
41 >>>>> [[ikiwiki/pagespec]], and decided that yes, sorting is
42 >>>>> a bit like a pagespec :-) Which name would you prefer? --s
44 >>>>>> `SortSpec` --[[Joey]]
48 >>>> I would be inclined to drop the `check_` stuff. --[[Joey]]
50 >>>>> It basically exists to support `title_natural`, to avoid
51 >>>>> firing up the whole import mechanism on every `cmp`
52 >>>>> (although I suppose that could just be a call to a
53 >>>>> memoized helper function). It also lets sort specs that
54 >>>>> *must* have a parameter, like
55 >>>>> [[field|plugins/contrib/field/discussion]], fail early
56 >>>>> (again, not so valuable).
58 >>>>>> AFAIK, `use foo` has very low overhead when the module is already
59 >>>>>> loaded. There could be some evalation overhead in `eval q{use foo}`,
60 >>>>>> if so it would be worth addressing across the whole codebase.
63 >>>>>>> check_cmp_foo now dropped. --s
65 >>>>> The former function could be achieved at a small
66 >>>>> compatibility cost by putting `title_natural` in a new
67 >>>>> `sortnatural` plugin (that fails to load if you don't
68 >>>>> have `title_natural`), if you'd prefer - that's what would
69 >>>>> have happened if `title_natural` was written after this
70 >>>>> code had been merged, I suspect. Would you prefer this? --s
72 >>>>>> Yes! (Assuming it does not make sense to support
73 >>>>>> natural order sort of other keys than the title, at least..)
76 >>>>>>> Done. I added some NEWS.Debian for it, too. --s
78 >>>> Wouldn't it make sense to have `meta(title)` instead
79 >>>> of `meta_title`? --[[Joey]]
81 >>>>> Yes, you're right. I added parameters to support `field`,
82 >>>>> and didn't think about making `meta` use them too.
83 >>>>> However, `title` does need a special case to make it
84 >>>>> default to the basename instead of the empty string.
86 >>>>> Another special case for `title` is to use `titlesort`
87 >>>>> first (the name `titlesort` is derived from Ogg/FLAC
88 >>>>> tags, which can have `titlesort` and `artistsort`).
89 >>>>> I could easily extend that to other metas, though;
90 >>>>> in fact, for e.g. book lists it would be nice for
91 >>>>> `field(bookauthor)` to behave similarly, so you can
92 >>>>> display "Douglas Adams" but sort by "Adams, Douglas".
94 >>>>> `meta_title` is also meant to be a prototype of how
95 >>>>> `sort=title` could behave in 4.0 or something - sorting
96 >>>>> by page name (which usually sorts in approximately the
97 >>>>> same place as the meta-title, but occasionally not), while
98 >>>>> displaying meta-titles, does look quite odd. --s
100 >>>>>> Agreed. --[[Joey]]
102 >>>>>>> I've implemented meta(title). meta(author) also has the
103 >>>>>>> `sortas` special case; meta(updated) and meta(date)
104 >>>>>>> should also work how you'd expect them to (but they're
105 >>>>>>> earliest-first, unlike age). --s
107 >>>> As I read the regexp in `cmpspec_translate`, the "command"
108 >>>> is required to have params. They should be optional,
109 >>>> to match the documentation and because most sort methods
110 >>>> do not need parameters. --[[Joey]]
112 >>>>> No, `$2` is either `\w+\([^\)]*\)` or `[^\s]+` (with the
113 >>>>> latter causing an error later if it doesn't also match `\w+`).
114 >>>>> This branch doesn't add any parameterized sort methods,
115 >>>>> in fact, although I did provide one on
116 >>>>> [[field's_discussion_page|plugins/contrib/report/discussion]]. --s
118 >>>> I wonder if it would make sense to add some combining keywords, so
119 >>>> a sortspec reads like `sort="age then ascending title"`
120 >>>> In a way, this reduces the amount of syntax that needs to be learned.
121 >>>> I like the "then" (and it could allow other operations than
122 >>>> simple combination, if any others make sense). Not so sure about the
123 >>>> "ascending", which could be "reverse" instead, but "descending age" and
124 >>>> "ascending age" both seem useful to be able to explicitly specify.
127 >>>>> Perhaps. I do like the simplicity of [[KathrynAndersen]]'s syntax
128 >>>>> from [[plugins/contrib/report]] (which I copied verbatim, except for
129 >>>>> turning sort-by-`field` into a parameterized spec).
131 >>>>> If we're getting into English-like (or at least SQL-like) queries,
132 >>>>> it might make sense to change the signature of the hook function
133 >>>>> so it's a function to return a key, e.g.
134 >>>>> `sub key_age { return -%pagemtime{$_[0]) }`. Then we could sort like
137 >>>>> field(artistsort) or field(artist) or constant(Various Artists) then meta(titlesort) or meta(title) or title
139 >>>>> with "or" binding more closely than "then". Does this seem valuable?
140 >>>>> I think the implementation would be somewhat more difficult. and
141 >>>>> it's probably getting too complicated to be worthwhile, though?
142 >>>>> (The keys that actually benefit from this could just
143 >>>>> have smarter cmp functions, I think.)
145 >>>>> If the hooks return keys rather than cmp results, then we could even
146 >>>>> have "lowercase" as an adjective used like "ascending"... maybe.
147 >>>>> However, there are two types of adjective here: "lowercase"
148 >>>>> really applies to the keys, whereas "ascending" applies to the "cmp"
149 >>>>> result. Again, I think this is getting too complex, and could just
150 >>>>> be solved with smarter cmp functions.
152 >>>>>> I agree. (Also, I think returning keys may make it harder to write
153 >>>>>> smarter cmp functions.) --[[Joey]]
155 >>>>> Unfortunately, `sort="ascending mtime"` actually sorts by *descending*
156 >>>>> timestamp (but`sort=age` is fine, because `age` could be defined as
157 >>>>> now minus `ctime`). `sort=freshness` isn't right either, because
158 >>>>> "sort by freshness" seems as though it ought to mean freshest first,
159 >>>>> but "sort by ascending freshness" means put the least fresh first. If
160 >>>>> we have ascending and descending keywords which are optional, I don't
161 >>>>> think we really want different sort types to have different default
162 >>>>> directions - it seems clearer to have `ascending` always be a no-op,
163 >>>>> and `descending` always negate.
165 >>>>>> I think you've convinced me that ascending/descending impose too
166 >>>>>> much semantics on it, so "-" is better. --[[Joey]]
168 >>>>>>> I've kept the semantics from `report` as-is, then:
169 >>>>>>> e.g. `sort="age -title"`. --s
171 >>>>> Perhaps we could borrow from `meta updated` and use `update_age`?
172 >>>>> `updateage` would perhaps be a more normal IkiWiki style - but that
173 >>>>> makes me think that updateage is a quantity analagous to tonnage or
174 >>>>> voltage, with more or less recently updated pages being said to have
175 >>>>> more or less updateage. I don't know whether that's good or bad :-)
177 >>>>> I'm sure there's a much better word, but I can't see it. Do you have
178 >>>>> a better idea? --s
180 [Regarding the `meta title=foo sort=bar` special case]
182 > I feel it sould be clearer to call that "sortas", since "sort=" is used
183 > to specify a sort method in other directives. --[[Joey]]
188 I notice the implementation does not use the magic `$a` and `$b` globals.
189 That nasty perl optimisation is still worthwhile:
191 perl -e 'use warnings; use strict; use Benchmark; sub a { $a <=> $b } sub b ($$) { $_[0] <=> $_[1] }; my @list=reverse(1..9999); timethese(10000, {a => sub {my @f=sort a @list}, b => sub {my @f=sort b @list}, c => => sub {my @f=sort { b($a,$b) } @list}})'
192 Benchmark: timing 10000 iterations of a, b, c...
193 a: 80 wallclock secs (76.74 usr + 0.05 sys = 76.79 CPU) @ 130.23/s (n=10000)
194 b: 112 wallclock secs (106.14 usr + 0.20 sys = 106.34 CPU) @ 94.04/s (n=10000)
195 c: 330 wallclock secs (320.25 usr + 0.17 sys = 320.42 CPU) @ 31.21/s (n=10000)
197 Unfortunatly, I think that c is closest to the new implementation.
200 > Unfortunately, `$a` isn't always `$main::a` - it's `$Package::a` where
201 > `Package` is the call site of the sort call. This was a showstopper when
202 > `sort` was a hook implemented in many packages, but now that it's a
203 > `SortSpec`, I may be able to fix this by putting a `sort` wrapper in the
204 > `SortSpec` namespace, so it's like this:
209 > return sort $cmp @_;
212 > which would mean that the comparison used `$IkiWiki::SortSpec::a`.
215 >> I've now done this. On a wiki with many [[plugins/contrib/album]]s
216 >> (a full rebuild takes half an hour!), I tested a refresh after
217 >> `touch tags/*.mdwn` (my tag pages contain inlines of the form
218 >> `tagged(foo)` sorted by date, so they exercise sorting).
219 >> I also tried removing sorting from `pagespec_match_list`
220 >> altogether, as an upper bound for how fast we can possibly make it.
222 >> * `master` at branch point: 63.72user 0.29system
223 >> * `master` at branch point: 63.91user 0.37system
224 >> * my branch, with `@_`: 65.28user 0.29system
225 >> * my branch, with `@_`: 65.21user 0.28system
226 >> * my branch, with `$a`: 64.09user 0.28system
227 >> * my branch, with `$a`: 63.83user 0.36system
228 >> * not sorted at all: 58.99user 0.29system
229 >> * not sorted at all: 58.92user 0.29system
233 > I do notice that `pagespec_match_list` performs the sort before the
234 > filter by pagespec. Is this a deliberate design choice, or
235 > coincidence? I can see that when `limit` is used, this could be
236 > used to only run the pagespec match function until `limit` pages
237 > have been selected, but the cost is that every page in the wiki
238 > is sorted. Or, it might be useful to do the filtering first, then
239 > sort the sub-list thus produced, then finally apply the limit? --s
241 >> Yes, it was deliberate, pagespec matching can be expensive enough that
242 >> needing to sort a lot of pages seems likely to be less work. (I don't
243 >> remember what benchmarking was done though.) --[[Joey]]
245 >>> We discussed this on IRC and Joey pointed out that this also affects
246 >>> dependency calculation, so I'm not going to get into this now... --s
248 Joey pointed out on IRC that the `titlesort` feature duplicates all the
249 meta titles. I did that in order to sort by the unescaped version, but
250 I've now changed the branch to only store that if it makes a difference.
253 ## Documentation from sort-package branch
255 ### advanced sort orders (conditionally added to [[ikiwiki/pagespec/sorting]])
257 * `title_natural` - Orders by title, but numbers in the title are treated
258 as such, ("1 2 9 10 20" instead of "1 10 2 20 9")
259 * `meta(title)` - Order according to the `\[[!meta title="foo" sortas="bar"]]`
260 or `\[[!meta title="foo"]]` [[ikiwiki/directive]], or the page name if no
261 full title was set. `meta(author)`, `meta(date)`, `meta(updated)`, etc.
264 ### Multiple sort orders (added to [[ikiwiki/pagespec/sorting]])
266 In addition, you can combine several sort orders and/or reverse the order of
267 sorting, with a string like `age -title` (which would sort by age, then by
268 title in reverse order if two pages have the same age).
270 ### meta sortas parameter (added to [[ikiwiki/directive/meta]])
274 An optional `sort` parameter will be used preferentially when
275 [[ikiwiki/pagespec/sorting]] by `meta(title)`:
277 \[[!meta title="The Beatles" sort="Beatles, The"]]
279 \[[!meta title="David Bowie" sort="Bowie, David"]]
283 An optional `sortas` parameter will be used preferentially when
284 [[ikiwiki/pagespec/sorting]] by `meta(author)`:
286 \[[!meta author="Joey Hess" sortas="Hess, Joey"]]
288 ### Sorting plugins (added to [[plugins/write]])
290 Similarly, it's possible to write plugins that add new functions as
291 [[ikiwiki/pagespec/sorting]] methods. To achieve this, add a function to
292 the IkiWiki::SortSpec package named `cmp_foo`, which will be used when sorting
293 by `foo` or `foo(...)` is requested.
295 The names of pages to be compared are in the global variables `$a` and `$b`
296 in the IkiWiki::SortSpec package. The function should return the same thing
297 as Perl's `cmp` and `<=>` operators: negative if `$a` is less than `$b`,
298 positive if `$a` is greater, or zero if they are considered equal. It may
299 also raise an error using `error`, for instance if it needs a parameter but
302 The function will also be passed one or more parameters. The first is
303 `undef` if invoked as `foo`, or the parameter `"bar"` if invoked as `foo(bar)`;
304 it may also be passed additional, named parameters.