| blob | 53e2ef116fe18d23443b9fc75059bde08121d8a6 |
1 2008-10-04 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
3 * Todo: now teach update to verify the timestamp is about to write
4 against the previous and use _increase_a_bit if it doesn't comply with
5 strict monotony. DONE
7 * The problem of rounding. So far perl's default precision was
8 sufficient. One day it won't be. FakeFloat has an easy job when it is
9 only reading and other machines have written correctly. But when we want
10 to write a floating point number that is a bit larger than the other
11 one, then we need our own idea of precision.
13 Slaven said: just append a "1". This might be going towards the end of
14 usability too quickly. I'd like something that actually uses the decimal
15 system. Well, appending a 1 also does this but...
17 E.g. we have 1.0. nextup on this architecture is starting with
18 1.0000000000000004. So there is a gap to fill: 1,2,3. Now I have
19 taken the 1.0000000000000003 and the next user comes and the time tells
20 him 1.0 again. He has to beat my number without stepping over the
21 nextup. This is much less space than I had when I chose 1,2,3.
23 What is also irritating is that nextup is architecture dependent. The
24 128 bit guy must choose very long numbers to fit in between whereas the
25 other one with 16 bit uses larger steps. But then the algorithm is the
26 same for both, so that would be a nice thing.
28 I see two situation where we need this. One is when Time::HiRes returns
29 us a value that is <= the last entry in our recentfile. In this case
30 (let's call it the end-case) we must fill the region between that number
31 and the next higher native floating point number. The other is when we
32 inject an old file into an old recentfile (we would then also set a new
33 dirtymark). We find the integer value already taken and need a slightly
34 different one (let's call it the middle-case). The difference between
35 the two situations is that the next user will want to find something
36 higher than my number in the end-case and something lower than my number
37 in the middle case.
39 So I suggest we give the function both a value and an upper bound and it
40 calculates us a primitive middle. The upper bound in the middle-case is
41 the next integer. The upper bound on the end-case is the nextup floating
42 point number. But the latter poses another problem: if we have occupied
43 the middle m between x and nextup(x), then the nextup(m) will probably
44 not be the same as nextup(x) because some rounding will take place
45 before the nextup is calculated and when the rounding reaches the
46 nextup(x), we will end up at nextup(nextup(x)).
48 So we really need to consider the nextup and the nextdown from there and
49 then the middle and that's the number we may approach asymptotically.
50 Ugly. But DONE.
52 2008-10-03 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
54 * consider deprecating the use of RECENT.recent as a symlink. It turns
55 out to need extra hoops with the rsync options and just isn't worth it.
56 Or maybe these extra hoops are needed anyway for the rest of the tree?
57 Nope, can't be the case because not all filesystems support symlinks.
59 But before doing the large step, I'll deprecate the call of
60 get_remote_recentfile_as_tempfile with an argument. Rememberr this was
61 only introduced to resolve RECENT.recent and complicates the routine far
62 beyond what it deserves.
64 DONE. Won't deprecate RECENT.recent, just moved its handling to the
65 supervisor.
67 2008-10-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
69 * I think it's a bug that the rsync_option links must be set to true in
70 order to support RECENT.recent and that nobody cares to set it
71 automatically. Similar for ignore_link_stat_errors.
73 2008-09-27 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
75 * Todo: find all todos together and make a plan what is missing for a
76 release.
78 - verifytree or something like that. fsck maybe.
80 - rersyncrecent, the script itself
82 - a way to only mirror the recentfiles without mirroring the whole
83 remote system such that people can decide to mirror only partially see
84 also 2008-08-30. .shadow-xxx directory? this also needed for a
85 filesystem that is still incomplete and might need the mirrorfiles for
86 lookup(?)
88 - long living objects that mirror again and again. Inject something
89 into ta, see how it goes over to tb.
91 - how do we continue filling up the DONE system when we use an object
92 for the second time? "fully covered" and "uptodate" or new terminology.
94 - idea: have a new flag on recentfiles with the meaning: if this
95 changes, you're required to run a full rsync over all the files. The
96 reason why we set it would probably be: some foul happened. we injected
97 files in arbitrary places or didn't inject them although they changed.
98 The content of the flag? Timestamp? The relation between the
99 recentfiles would have to be inheritance from the principal, because any
100 out of band changes would soon later propagate to the next recentfile.
102 By upping the flag often one can easily ruin the slaves.
104 last out of band change? dirtymark?
106 Anyway, this implies that we read a potentially existing recentfile
107 before we write one.
109 And it implies that we have an eventloop that keeps us busy in 2-3
110 cycles, one for current stuff (tight loop) and one for the recentfiles
111 (cascade when principal has changed), one for the old stuff after a
112 dirtymark change.
114 - overview called on the wrong file should be understandable
116 - See that the long running process really only updates the principal
117 file unless it has missed a timespan during which something happened. If
118 nothing happened, it must notice even when it misses the timespan.
120 - fill up recentfiles with fake (historic) entries; fill up with
121 individual corrections; algorithm maybe to be done with bigfloat so that
122 we can always place something in the middle between two entries. Before
123 we must switch to bigfloat we could try to use Data::Float::nextup to
124 get the.
126 - the meta data field that must change when we fake something up so that
127 the downstream people know they have to re-fetch everything.
129 - how tolerant are we against missing files upstream? how do we keep
130 track? there are legitimate cases where we did read upstream index right
131 before a file got deleted there and then find that file as new and want
132 it. There are other cases that are not self healing and must be tracked
133 and bugreported.
135 - how, exactly, do we have to deal with deletes? With rsync errors?
137 rsync: link_stat "/id/K/KA/KARMAN/Rose-HTMLx-Form-Related-0.07.meta" (in
138 authors) failed: No such file or directory (2)
140 The file above is a delete in 1h and a new in file 1M and the
141 delete in the locally running rmirror did not get propagated to the 1M
142 object. Bug. And the consequence is a standstill.
144 It seems that a slave that works with a file below the principal needs
145 to merge things all the way up to get rid of later deletes. Or keep
146 track of all deletes and skip them later. So we need a trackdeletes.pm
147 similar to the done.pm?
149 see also 2008-08-20 about spurious deletes that really have no add
150 counterpart and yet they are not wrong.
152 - consider the effect when resyncing the recentfile takes longer than
153 the time per loop. Then we never rsync any file. We need to diagnose
154 that and force an increase of that loop time. But when we later are fast
155 enough again because the net has recovered, then we need to switch back
156 to original parameters. ERm, no, it's enough to keep syncing at least
157 one file before refetching an index file.
159 - remember to verify that no temp files are left lying around and the
160 signal handler
162 - we must throw away the pathdb when we have reached the end of Z. From
163 that moment we can have a very small pathdb because the only reason for
164 a pathdb is that we know to ignore old records in old files. We won't
165 need this pathdb again before the next full pass over the data is
166 necessary and then we will rebuild it as we go along.
168 - status file for not long running jobs that want to track upstream with
169 a, say, cronjob.
171 - revisit all XXX _float areas and study Sub::Exporter
173 - persistent DB even though we just said we do not need it. Just for
174 extended capabilities and time savings when, for example, upstream
175 announces a reset and we get new recentfiles and could then limit
176 ourselves to a subset of files (those that have a changed epoch) in a
177 first pass and would only then do the loop to verify the rest. Or
178 something.
180 - lookup by epoch and by path and use this ability on the pause to never
181 again register a file twice that doesn't need it.
183 * Todo: aggregate files should know their feed and finding the principal
184 should be done stepwise. (?)
186 * Todo: DESTROY thing that unlocks. Today when I left the debuggerr I
187 left locks around. DONE
189 2008-09-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
191 * maybe extend the _overview so that it always says if and where the
192 last file is in the next file and where the next event in the next rf
193 would lie. No, don't like this anymore. REJECT
195 * take the two new redundant tests out again, only the third must
196 survive. DONE
198 * Todo: add a sanity check if the merged structure is really pointing to
199 a different rf and that this different rf is larger. DONE
201 2008-09-25 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
203 * now test, if they are overlapping. And test if there is a file in the
204 next rf that would fit into this rf's interval.
206 1h 1222324012.8474 1222322541.7963 0.4086
207 6h 1222320411.2760 1222304207.6931 4.5010 missing overlap/gap!
208 1d 1222320411.2760 1222238750.5071 22.6835 large overlap
209 1W 1222313218.3626 1221708477.5829 167.9835
211 I suspect that somebody writes a merged timestamp without having merged
212 and then somebody else relies on it.
214 If aggregate is running, the intervals must not be extravagated, if it
215 is not running, there must not be bounds, the total number of events in
216 the system must be counted and must be controlled throughout the tests.
217 That the test required the additional update was probably nonsense,
218 because aggregate can cut pieces too. FIXED & DONE
220 2008-09-23 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
222 * rrr-aggregate seems to rewrite the RECENT file even if nothing has
223 changed. FIXED
225 2008-09-21 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
227 * Most apparent bug at the moment is that the recentfiles are fetched
228 too often. Only the principal should be fetched and if it has not
229 changed, the others should not be refetched. ATM I must admit that I'm
230 happy that we refetch more often than needed because I can more easily
231 fix bugs while the thing is running.
233 * Let's say, 1220474966.19501 is a timestamp of a file that is already
234 done but the done system does not know about it. The reason for the
235 failure is not known and we never reach the status uptodate because of
236 this. We must get over it.
238 Later it turns out that the origin server had a bug somewhere.
239 1220474966.19042 came after 1220474966.19501. Or better: it was in the
240 array of the recentfile one position above. The bug was my own.
242 2008-09-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
244 * There is the race condition where the server does a delete and the
245 slave does not yet know and then tries to download it because he sees
246 the new. So for this time window we must be more tolerant against
247 failure. If we cannot download a file, we should just skip it and should
248 not retry immediately. The whole system should discover the lost thing
249 later. Keeping track with the DONE system should really be a no brainer.
251 But there is something more: the whole filesystem is a database and the
252 recentfiles are one possible representation of it. It's a pretty useful
253 representation I think that's why I have implemented something around
254 it. But for strictly local operation it has little value. For local
255 operation we would much rather have a database. So we would enter every
256 recentfile reading and every rsync operation and for every file the last
257 state change and what it leads to. Then we would always ignore older
258 records without the efforts involved with recentfiles.
260 The database would have: path,recentepoch,rsyncedon,deletedon
262 Oh well, not yet clear where this leads to.
264 2008-09-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
266 * Bug: the bigloop ran into a funny endless loop after EWILHELM uploaded
267 Module-Build. It *only* rsynced the "1h" recentfile from that moment on.
269 * statusfile, maybe only on demand, alone to have a sharp debugging
270 tool. It is locked and all recentfiles dump themselves into it and we
271 can build a viewer that lets us know where we stand and what's inside.
273 * remember: only the principal recentfile needs expiration, all others
274 shall be expired by principal if it discovers that something has move
275 upstream.
277 2008-09-18 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
279 * Always check if we stringify to a higher value than in the entry
280 before. DONE
282 * And in covered make an additional check if we would be able to see a
283 numerical difference between the two numbers and if we can't then switch
284 to a different, more expensive algorithm. Do not want to be caught by
285 floating surprises. DONE
287 2008-09-17 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
289 * caching has several aspects here: we can cache the interval of the
290 recentfile which only will change when the mtime of the file changes. We
291 must re-mirror the recentfile when its ttl has expired. Does have_read
292 tell you anything? It counts nothing at all. Only the mtime is
293 interesting. The ntuple mtime, low-epoch, high-epoch. And as a separate
294 thing the have_mirrored because it is unrelated to the mtime.
296 * Robustness of floating point calculations! I always thought that the
297 string calculated by the origin server for the floating representation
298 of the epoch time is just a string. When we convert it to a number and
299 later back to a string, the other computer might come to a different
300 conclusion. This must not happen, we want to preserve it under any
301 circumstances. I will have to write tests with overlong sequences that
302 get lost in arithmetic and must see if all still works well. DONE
304 But one fragile point remains: if one host considers a>b and the other
305 one considers them == but no eq. To prevent this, we must probably do
306 some extra homework. DONE
308 2008-09-16 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
310 * the concept of tracking DONE needs an object per recentfile that has
311 something like these methods:
313 do_we_have(xxx), we_have(xxx), do_we_have_all(xxx,yyy), reset()
315 covered() register() covered()
317 The unclear thing is how we translate points in time into intervals. We
318 could pass a reference to the current recent_events array when running
319 we_have(xxx) and let the DONE object iterate over it such that it only
320 has to store a list of intervals that can melt into each other. Ah, even
321 passing the list together with a list of indexes seems feasiable.
323 Or maybe ask for the inverted list?
325 Whenever the complete array is covered by the interval we say we are
326 fully covered and if the recentfile is not expired, we are uptodate.
328 2008-09-07 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
330 2008-09-05 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
332 * need a way to "return" the next entry after the end of a list. When
333 the caller says "before" or "after" we would like to know if he could
334 cover that interval/threshold or not because this influences the effect
335 of a newer timestamp of that recentfile. DONE with $opt{info}.
337 2008-09-04 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
339 * one of the next things to tackle: the equivalent of csync2 -TIXU.
341 loop implies tixu (?). Nope, something like --statefile decides. Per
342 default we do ...?
344 T test, I init, X including removals, U nodirtymark
346 So we have no concept of dirtymarks, we only trust that since we are
347 running we have observed everything steadily. But people will not let
348 this program run forever so we must consider both startup penalty and
349 book keeping for later runs. We keep this for later. For now we write a
350 long running mirror that merges several intervals.
352 2008-09-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
354 * need to speed up the 02 test, it's not clever to sleep so much. Reduce
355 the intervals!
357 * rersyncrecent, the script: default to one week. The name of the switch
358 is --after. Other switches? --loop!
360 2008-08-30 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
362 * need a switch --skip-deletes (?)
364 * need a switch --enduser that tells us that the whole tempfile
365 discipline is not needed when there is no downstream user. (?)
367 Without this switch we cannot have a reasonable recent.pl that just
368 displays the recent additions. Either we accept to download everything.
369 Or we download temporary files without the typical rsync protocol
370 advantages.
372 Or maybe the switch is --tmpdir? If --tmpdir would mean: do not use
373 File::Temp::tempdir, this might be a win.
375 2008-08-29 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
377 * apropos missing: we have no push, we never know the downstream
378 servers. People who know their downstream hosts and want to ascertain
379 something will want additional methods we have never thought about, like
380 update or delete a certain file.
382 2008-08-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
384 * tempted to refactor rmirror into resolve_symlink, localize, etc.
385 Curious if rsync_options=links equal 0 vs. 1 will make the expected
386 difference.
388 * rsync options: it's a bit of a pain that we usually need several rsync
389 options, like compress, links, times, checksum and that there is no
390 reasonable default except the original rsync default. I think wee can
391 safely assume that the rsync options are shared between all recentfile
392 instances within one recent tree.
394 2008-08-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
396 * deletes: if a delete follows an add quickly enough it may happen that
397 a downstream mirror did not see the add at all! It seems this needs to
398 be mentioned somewhere. The point here is that even if the downstream is
399 never missing the principal timeframe it may encounter a "delete" that
400 has no complimentary "add" anywhere.
402 2008-08-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
404 * I suspect the treat of metadata is incorrect during read or something.
405 The bug that I am watching is that between 06:08 and 06:09 the 6h file
406 contained more than 6 hours worth of data. At 06:08 we merged into the
407 1d file. We need to take snapshots of the 6h file over the course of an
408 hour or maybe only between XX:08 and XX:09? Nope, the latter is not
409 enough.
411 Much worse: watching the 1h file: right at the moment (at 06:35) it
412 covers 1218867584-1219120397 which is 70 hours.
414 Something terribly broken. BTW, 1218867584 corresponds to Sat Aug 16
415 08:19:44 2008, that is when I checked out last time, so it seems to be
416 aggregating and never truncating?
418 No, correct is: it is never truncating; but wrong is: it is aggregating.
419 It does receive a lot of events from time to time from a larger file.
420 Somehow a large file gets merged into the small one and because the
421 "meta/merged" attribute is missing, nobody is paying attention. I
422 believe that I can fix this by making sure that metadata are honoured
423 during read. DONE and test adjusted.
425 2008-08-17 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
427 * grand renaming plan
429 remotebase => remoteroot to fit well with localroot DONE
430 local_path() => localroot seems to me should already work DONE
431 recentfile_basename => rfilename no need to stress it has no slash DONE
433 filenameroot??? Doesn't seem too bad to me today. Maybe something like
434 kern? It would anyway need a deprecation cycle because it is an
435 important constructor.
437 * I like the portability that Data::Serializer brings us but the price
438 is that some day we might find out that it is slowing us a bit. We'll
439 see.
441 2008-08-16 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
443 * should we not enter the interval of the principal (or the interval of
444 the merging file?) in every aggregated/merged file?
446 * we should aim at a first release and give up on thinking about
447 sanitizing stuff and zloop. Let's just admit that a full traditional
448 rsync is the only available sanitizer ATM. Otherwise it's complicated
449 stuff: sanitizing on the origin server, sanitizing on the slaves,
450 sanitizing forgotten files, broken timestamps, etc. Let's delay it and
451 get the basics out before this becomes a major cause for mess.
453 2008-08-13 Andreas Koenig <k@andreas-koenigs-computer.local>
455 * On OSes not supporting symlinks we expect that RECENT.recent contains
456 the contents of the principal recentfile. Actually this is identical on
457 systems supporting symlinks. Simple, what follows from that is that we
458 need to keep the serializer in the metadata because we cannot read it
459 from the filename, doesn't it? Of course not. It's a chicken and egg
460 problem. This leaves us with the problem to actually parse the
461 serialized data to find out in which format it is. So who can do the 4
462 or 5 magics we wanted to support? File::LibMagic?
464 2008-08-09 Andreas Koenig <k@andreas-koenigs-computer.local>
466 * remotebase and recentfile_basename are ugly names. Now that we need a
467 word for the shortest/principal/driving recentfile too we should do
468 something about it.
470 localroot is good. rfile is good. local_path() is bad, local_path($path)
471 is medium, filenameroot() is bad, remotebase is bad, recentfile is
472 already deprecated.
474 Up to now remotebase was the string that described the remote root
475 directory in rsync notation, like pause.perl.org::authors. And
476 recentfile_basename was "RECENT-1h.yaml".
478 2008-08-08 Andreas Koenig <k@andreas-koenigs-computer.local>
480 * The test that was added in today's checkin is a good start for a test
481 of rmirror. We should have more methods in Recent.pm: verify,
482 addmissingfiles. We should verify the current tree, then rmirror it and
483 then verifytree the copy. We could then add some arbitrary file and let
484 it be discovered by addmissingfiles, then rmirror again and then
485 verifytree the copy again.
487 Then we could start stealing from csync2 sqlite database [no port to
488 OSX!] and fill a local DB. And methods to compare the database with the
489 recentfiles. Our strength is that in principle we could maintain state
490 with a single float. We have synced up to 1234567890.123456. If the Z
491 file does not add new files all we have to do is mirror the new ones and
492 delete the goners.
494 This makes it clear that we should extend current protocol and declare
495 that we cheat when we add files too late, just to help the other end
496 keeping track. Ah yes, that's what was meant when zloop was mentioned
497 earlier.
499 Maybe need to revisit File::Mirror to help me with this task.
501 2008-08-07 Andreas Koenig <k@andreas-koenigs-computer.local>
503 * There must be an allow-me-to-truncate flag in every recentfile.
504 Without it one could construct a sequence of updates winning the locking
505 battle against the aggregator. Only if an aggregator has managed to
506 merge data over to the next level, truncating can be allowed. DONE with
507 accessor merged.
509 2008-08-06 Andreas Koenig <k@andreas-koenigs-computer.local>
511 * We should probably guarantee that no duplicates enter the aggregator
512 array.
514 2008-08-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
516 * To get merge operation faster would need a good benchmark test. What
517 02 spits out isn't reliable enough and is dominated by many other
518 things. Between
520 commit 10176bf6b79865d4fe9f46e3857a3b8669fa7961
521 Author: Andreas J. Koenig <k@k75.(none)>
522 Date: Sat Aug 2 07:58:04 2008 +0200
524 and
526 commit 3243120a0c120aaddcd9b1f4db6689ff12ed2523
527 Author: Andreas J. Koenig <k@k75.(none)>
528 Date: Sat Aug 2 11:40:29 2008 +0200
530 there was a lot of trying but the effect is hardly measurable with
531 current tests.
533 * overhead of connecting seems high. When setting
534 max_files_per_connection to 1 we see that.
536 2008-08-01 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
538 * 1217622571.0889 - 1217597432.86734 = 25138.2215600014
540 25138.2215600014/3600 = 6.98283932222261
542 It jumps into the eye that this is ~ 7 hours, not ~6, so there seems to
543 be a bug in the aggregator. FIXED
545 2008-07-27 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
547 * e.g. id/Y/YE/YEWENBIN/Emacs-PDE-0.2.16.tar.gz: Do we have it, should
548 we have it, can we mirror it, mirror it!
550 I fear this needs a new class which might be called
551 File::Rsync::Mirror::Recent. It would collect all recentfiles of a kind
552 and treat them as an entity. I realize that a single recentfile may be
553 sufficient for certain tasks and that it is handy for the low level
554 programmer but it is not nice to use. If there is a delete in the 1h
555 file then the 6h file still contains it. Seekers of the best information
556 need to combine at least some of the recentfiles most of the time.
558 There is the place for the Z loop!
560 But the combination is something to collect in a database, isn't it. Did
561 csync2 just harrumph?
563 2008-07-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
565 * it just occurred to me that hosts in the same mirroring pool could
566 help out each other even without rewriting the recentfile. Just fetch
567 the stuff to mirror from several places, bingo. But that's something
568 that should rather live in a separate package or in rsync directly.
570 * cronjobs are unsuited because with ntp they would all come at the full
571 minute and disturb each other. Besides that I'd hate to have a backbone
572 with more than a few seconds latency.
574 2008-07-25 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
576 * a second rsync server with access control for PAUSE. Port? 873 is the
577 standard port, let's take 8873.
579 * if there were a filesystem based on this, it would have a slow access
580 to inexistent files. It would probably provide wrong readdir (only based
581 on current content) or also a slow one (based on a recentfile written
582 after the call). But it would provide fast access to existing files. Or
583 one would deliberately allow slightly blurred answers based on some
584 sqlite reflection of the recentfiles.
586 * todo: write a variant of mirror() that combines two or more
587 recentfiles and treats them like one
589 * todo: signal handler to remove the tempfile
591 2008-07-24 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
593 * now that we have the symlink I forgot how it should be used in
594 practice.
596 * the z loop: add missing files to Z file. Just append them (instead of
597 prepending). So one guy prepends something from the Y file from time to
598 time and another guy appends something rather frequently. Collecting
599 pond. When Y merges into Z, things get epoch and the collecting pond
600 gets smaller. What exactly are "missing files"?
602 take note of current epoch of the alpha file, let's call it the
603 recent-ts
605 find all files on disk
607 remove all files registered in the recentworld up to recent-ts
609 remove all files that have been deleted after recent-ts according to
610 recentworld
612 2008-07-23 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
614 * rersyncrecent might be a cronjob with a (locked) state file which
615 contains things like after and maybe last z sync or such?
617 rrr-mirror might be an alternative name but how would we justify the
618 three Rs when there is no Re-Rsync-Recent?
620 With the --loop parameter it is an endless loop, without it is no loop.
621 At least this is simple.
623 * todo: new accssor z-interval specifies how often the Z file is updated
624 against the filesystem. We probably want no epoch stamp on these
625 entries. And we want to be able to filter the entries (e.g. no
626 by-modules and by-category tree)
628 2008-07-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
630 * Fill the Z file. gc or fsck or both. Somehow we must get the old files
631 into Z. We do not need the other files filled up with filesystem
632 contents though.
634 * need interface to query for a file in order to NOT call update on
635 PAUSE a second time within a short time.
637 2008-07-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
639 * recommended update interval? Makes no sense, is different for
640 different users.
642 * Moosify
644 Local Variables:
645 mode: change-log
646 change-log-default-name: "Todo"
647 tab-width: 2
648 left-margin: 2
649 End: