1 On [[todo/auto-create_tag_pages_according_to_a_template]], [[chrysn]]
4 > Instead of creating a file that gets checked in into the RCS, the
5 > source files could be left out and the output files be written as
6 > long as there is no physical source file (think of a virtual underlay).
7 > Something similar would be required to implement alias directive,
8 > which couldn't be easily done by writing to the RCS as the page's
9 > contents can change depending on which other pages claim it as an alias.
11 `add_autofile` could be adapted to do this, or a similar API could be
14 This would also be useful for autoindex, as suggested on
15 [[plugins/autoindex/discussion]]. I'd also like to use it for
16 [[plugins/contrib/album]].
18 It could also be used for an [[todo/alias_directive]].
22 --------------------------
24 [[!template id=gitbranch branch=smcv/ready/transient author="[[smcv]]"]]
25 [[!template id=gitbranch branch=smcv/ready/transient-recentchanges author="[[smcv]]"]]
26 [[!template id=gitbranch branch=smcv/ready/tag-test author="[[smcv]]"]]
27 [[!template id=gitbranch branch=smcv/transient-tag author="[[smcv]]"]]
30 I think this branch is now enough to be useful. It adds the following:
32 If the `transient` plugin is loaded, `$srcdir/.ikiwiki/transient` is added
33 as an underlay. I'm not sure whether this should be a plugin or core, so
34 I erred on the side of more plugins; I think it's "on the edge of the core",
37 Pages in the transient underlay are automatically
38 deleted if a page of the same name is created in the srcdir (or an underlay
39 closer to the srcdir in stacking order).
41 With the additional `ready/transient-tag` branch,
42 `tag` enables `transient`, and if `tag_autocreate_commit` is set to 0
43 (default 1), autocreated tags are written to the transient underlay.
44 There is a regression test.
46 With the additional `transient-autoindex` branch,
47 `autoindex` uses autofiles. It also enables `transient`, and if
48 `autoindex_commit` is set to 0 (default 1), autoindexes are written to
49 the transient underlay. There is a regression test. However, this branch
50 is blocked by working out what the desired behaviour is, on
51 [[todo/autoindex_should_use_add__95__autofile]].
53 > I wonder why this needs to be configurable? I suppose that gets back to
54 > whether it makes sense to check these files in or not. The benefits of
57 > * You can edit them from the VCS, don't have to go into the web
58 > interface. Of course, files from the underlays have a similar issue,
59 > but does it make sense to make that wart larger?
60 > * You can know you can build the same site with nothing missing
61 > even if you don't there enable autoindex or whatever. (Edge case.)
63 >> I'm not sure that that's a huge wart; you can always "edit by
64 >> overwriting". If you're running a local clone of the wiki on your laptop
65 >> or whatever, you have the underlays already, and can copy from there.
66 >> Tag and autoindex pages have rather simple source code anyway. --s
68 > The benefit of using transient pages seems to just be avoiding commit
69 > clutter? For files that are never committed, transient pages are a clear
70 > win, but I wonder if adding configuration clutter just to avoid some
71 > commit clutter is really worth it.
73 >> According to the last section of
74 >> [[todo/auto-create_tag_pages_according_to_a_template]], [[chrysn]] and
75 >> Eric both feel rather strongly that it should be possible to
76 >> not commit any tags; in [[plugins/autoindex/discussion]],
77 >> lollipopman and [[JoeRayhawk]] both requested the same for autoindex.
78 >> I made it configurable because, as you point out,
79 >> there are also reasons why it makes sense to check these
80 >> automatically-created files in. I'm neutral on this, personally.
82 >> If this is a point of contention, would you accept a branch that
83 >> just adds `transient` and uses it for [[plugins/recentchanges]],
84 >> which aren't checked in and never have been? I've split the
85 >> branch up in the hope that *some* of it can get merged.
87 >>> I will be happy to merge transient-recentchanges when it's ready.
88 >>> I see no obstacle to merging transient-tag either, and am not
89 >>> really against using it for autoindex or aggregate either
90 >>> once they get completed.
91 >>> I just wanted to think through why configurability is needed.
94 >> One potentially relevant point is that configuration clutter only
95 >> affects the site admin whereas commit clutter is part of the whole
96 >> wiki's history. --[[smcv]]
98 > Anyway, the configurability
99 > appears subtly broken; the default is only 1 if a new setup file is
100 > generated. (Correction: It was not even the default then --[[Joey]])
101 > With an existing setup file, the 'default' values in
102 > `getsetup` don't take effect, so it will default to undef, which
103 > is treated the same as 0. --[[Joey]]
105 >> Fixed in the branches, hopefully. (How disruptive would it be to have
106 >> defaults take effect whenever the setup file doesn't set a value, btw?
107 >> It seems pretty astonishing to have them work as they do at the moment.) --s
109 >>> Well, note that default is not actually a documented field in
110 >>> getsetup hooks at all! (It is used in IkiWiki.pm's own `getsetup()`, and
111 >>> the concept may have leaked out into one or two plugins (comments,
114 >>> Running getsetup at plugin load time is something I have considered
115 >>> doing. It would simplify some checkconfig hooks that just set hardcoded
116 >>> defaults. Although since dying is part of the getsetup hook's API, it
117 >>> could be problimaric.
120 autoindex ignores pages in the transient underlay when deciding whether
121 to generate an index.
123 With the additional `ready/transient-recentchanges` branch, new recent
124 changes go in the transient underlay; I tested this manually.
126 Not done yet (in that branch, at least):
128 * `remove` can't remove transient pages: this turns out to be harder than
129 I'd hoped, because I don't want to introduce a vulnerability in the
130 non-regular-file detection, so I'd rather defer that.
132 > Hmm, I'd at least want that to be dealt with before this was used
133 > by default for autoindex or tag. --[[Joey]]
135 >> I'll try to work out which of the checks are required for security
136 >> and which are just nice-to-have, but I'd appreciate any pointers
137 >> you could give. Note that my branch wasn't meant to enable either
138 >> by default, and now hopefully doesn't. --[[smcv]]
140 * Transient tags that don't match any pages aren't deleted: I'm not sure
141 that that's a good idea anyway, though. Similarly, transient autoindexes
142 of directories that become empty aren't deleted.
144 > Doesn't seem necessary, or really desirable to do that. --[[Joey]]
146 >> Good, that was my inclination too. --s
148 * In my `untested/transient` branch, new aggregated files go in the
149 transient underlay too (they'll naturally migrate over time). I haven't
150 tested this yet, it's just a proof-of-concept.
152 > I can confirm that the behavior of autoindex, at least, is excellent.
153 > Haven't tried tag. Joey, can you merge transient and autoindex? --JoeRayhawk
155 >> Here are some other things I'd like to think about first: --[[Joey]]
157 >> * There's a FIXME in autoindex.
159 >> > Right, the extra logic for preventing autoindex pages from being
160 >> > re-created. This is taking a while, so I'm going to leave out the
161 >> > autoindex part for the moment. The FIXME is only relevant
162 >> > because I tried to solve
163 >> > [[todo/autoindex should use add__95__autofile]] first, but
164 >> > strictly speaking, that's an orthogonal change. --s
166 >> * Suggest making recentchanges unlink the transient page
167 >> first, and only unlink from the old location if it wasn't
168 >> in the transient location. Ok, it only saves 1 syscall :)
170 >> > Is an unlink() really that expensive? But, OK, fixed in the
171 >> > `ready/transient-recentchanges` branch. --s
173 >> >> It's not, but it's easy. :) --[[Joey]]
175 >> * Similarly it's a bit worrying for performance that it
176 >> needs to pull in and use `Cwd` on every ikiwiki startup now.
177 >> I really don't see the need; `wikistatedir` should
178 >> mostly be absolute, and ikiwiki should not chdir in ways
179 >> that break it anyway.
181 >> > The reason to make it absolute is that relative underlays
182 >> > are interpreted as relative to the base underlay directory,
183 >> > not the cwd, by `add_underlay`.
185 >> > The updated `ready/transient-only` branch only loads `Cwd` if
186 >> > the path is relative; an extra commit on branch
187 >> > `smcv/transient-relative` goes behind `add_underlay`'s
188 >> > back to allow use of a cwd-relative underlay. Which direction
189 >> > would you prefer?
191 >> > I note in passing that [[plugins/autoindex]] and `IkiWiki::Render`
192 >> > both need to use `Cwd` and `File::Find` on every refresh, so
193 >> > there's only any point in avoiding `Cwd` for runs that don't
194 >> > actually refresh, like simple uses of the CGI. --s
196 >> >> Oh, right, I'd forgotten about the horrificness of File::Find
197 >> >> that required a chdir for security. Ugh. Can we just avoid
198 >> >> it for those simple cases then? (demand-calculate wikistatedir)
201 >> >>> The reason that transientdir needs to be absolute is that it's
202 >> >>> added as an underlay.
204 >> >>> We could avoid using `Cwd` by taking the extra commit from either
205 >> >>> `smcv/transient-relative` or `smcv/transient-relative-api`;
206 >> >>> your choice. I'd personally go for the latter.
208 >> >>> According to git grep, [[plugins/po]] already wants to look at
209 >> >>> the underlaydirs in its checkconfig hook, so I don't think
210 >> >>> delaying calculation of the underlaydir is viable. (I also noticed
212 >> >>> [[bugs/po:_might_not_add_translated_versions_of_all_underlays]].)
214 >> >>> `underlaydirs` certainly needs to have been calculated by the
215 >> >>> time `refresh` hooks finish, so `find_src_files` can use it. --s
217 >> * Unsure about the use of `default_pageext` in the `change`
218 >> hook. Is everything in the transientdir really going
219 >> to use that pageext? Would it be better to look up the
220 >> complete source filename?
222 >> > I've updated `ready/transient` to do a more thorough GC by
223 >> > using File::Find on the transient directory. This does
224 >> > require `File::Find` and `Cwd`, but only when pages change,
225 >> > and `refresh` loads both of those in that situation anyway.
227 >> > At the moment everything in the transientdir will either
228 >> > have the `default_pageext` or be internal, although I
229 >> > did wonder whether to make [[plugins/contrib/album]]
230 >> > viewer pages optionally be `html`, for better performance
231 >> > when there's a very large number of photos. --s
233 >> >> Oh, ugh, more File::Find... Couldn't it just assume that the
234 >> >> transient page has the same extension as its replacement?
237 >> >>> Good idea, that'll be true for web edits at least.
238 >> >>> Commit added. --s
240 --------------------------
242 ## An earlier version
244 I had a look at implementing this. It turns out to be harder than I thought
245 to have purely in-memory pages (several plugins want to be able to access the
246 source file as a file), but I did get this proof-of-concept branch
247 to write tag and autoindex pages into an underlay.
249 This loses the ability to delete the auto-created pages (although they don't
250 clutter up git this way, at least), and a lot of the code in autoindex is
251 probably now redundant, so this is probably not quite ready for merge, but
252 I'd welcome opinions.
254 Usage: set `tag_underlay` and/or `autoindex_underlay` to an absolute path,
255 which you must create beforehand. I suggest *srcdir* + `/.ikiwiki/transient`.
257 Refinements that could be made if this approach seems reasonable:
259 * make these options boolean, and have the path always be `.ikiwiki/transient`
260 * improve the `remove` plugin so it also deletes from this special underlay
262 >> Perhaps it should be something more generic, so that other plugins could use it (such as "album" mentioned above).
263 >> The `.ikiwiki/transient` would suit this, but instead of saying "tag_underlay" or "autoindex_underlay" have "use_transient_underlay" or something like that?
264 >> Or to make it more flexible, have just one option "transient_underlay" which is set to an absolute path, and if it is set, then one is using a transient-underlay.
265 >> --[[KathrynAndersen]]
267 >>> What I had in mind was more like `tag_autocreate_transient => 1` or
268 >>> `autoindex_transient => 1`; you might conceivably want tags to be
269 >>> checked in but autoindices to be transient, and it's fine for each
270 >>> plugin to make its own decision. Going from that to one boolean
271 >>> (or just always-transient if people don't think that's too
272 >>> astonishing) would be trivial, though.
274 >>> I don't think relocating the transient underlay really makes sense,
275 >>> except for prototyping: you only want one, and `.ikiwiki` is as good
276 >>> a place as any (ikiwiki already needs to be able to write there).
278 >>> For [[plugins/contrib/album]] I think I'd just make the photo viewer
279 >>> pages always-transient - you can always make a transient page
280 >>> permanent by editing it, after all.
282 >>> Do you think this approach has enough potential that I should
283 >>> continue to hack on it? Any thoughts on the implementation? --[[smcv]]
285 >>>> Ah, now I understand what you're getting at. Yes, it makes sense to put transient pages under `.ikiwiki`.
286 >>>> I haven't looked at the code, but I'd be interested in seeing whether it's generic enough to be used by other plugins (such as `album`) without too much fuss.
287 >>>> The idea of a transient underlay gives us a desirable feature for free: that if someone edits the transient page, it is made permanent and added to the repository.
289 >>>> I think the tricky thing with removing these transient underlay pages is the question of how to prevent whatever auto-generated the pages in the first place from generating them again - or, conversely, how to force whatever auto-generated those pages to regenerate them if you've changed your mind.
290 >>>> I think you'd need something similar to `will_render` so that transient pages would be automatically removed if whatever auto-generated them is no longer around.
291 >>>> -- [[KathrynAndersen]]