1 2009-04-08 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
3 * install the bin/* files and talk about rrr-overview somewhere.
5 * Ask B. Hansen asks for tempdir accessor such that tempfiles get
6 created outside the target tree.
8 * Barbie asks for better logging capabilities
10 2009-03-29 Andreas J. Koenig <andk@cpan.org>
12 * known bug: cannot configure the Recentfile objects to keep delete
13 events. How to call the accessor for that? keep_delete_objects_forever?
15 Sounds acceptable. FIXED (but untested).
17 * known bug: keeps temporary index files lying around.
19 2009-03-24 Andreas J. Koenig <andk@cpan.org>
23 for ( @{$rrr->recentfiles} ) { $_->verbose(0) }
27 2009-03-22 Andreas J. Koenig <andk@cpan.org>
29 * k81 is again client of k75 with different parameters.
31 * broken now: (1) seeding and unseeding: rmirror is seeding and talking
32 about it all the time and nobody reacts accordingly; (2) lots of temp
33 files get created and not removed; culprit the new call to
34 get_remote_recentfile_as_tempfile within Recent.pm; the manpage says the
35 caller has to remove the tempfile after use.
37 Need the drawing board.
39 PARTLY FIXED: retracting the idea to call
40 get_remote_recentfile_as_tempfile from Recent.pm and moving around seed
41 and unseed such that uptodate now means "we have mirrored this
42 recentfile's bunch of files" and seeded means "somebody believes the
43 index file for this rf needs be refreshed". This seems to work now.
45 Still collecting temp files that nobody cares to depollute.
47 2009-03-21 Andreas J. Koenig <andk@cpan.org>
49 * Bug?: should it be harder than it is atm to set the timestamp to the
52 * bug with native integers:
56 path: id/D/DE/DELTA/Crypt-Rijndael_PP-0.03.readme
60 path: id/L/LG/LGODDARD/Tk-Wizard-2.124.readme
63 Native integer broke when native math was turned off. FIXED
65 * Bug: something between id/P/PH/PHISH/CGI-XMLApplication-1.1.2.readme
66 and id/C/CH/CHOGAN/HTML-WWWTheme-1.06.readme. Mirroring the Z file loops
69 Yes, records out of order:
72 447004 epoch: 995885533
73 447005 path: id/P/PH/PHISH/CGI-XMLApplication_0.9.3.readme
76 447008 epoch: 995890358
77 447009 path: id/H/HD/HDIAS/Mail-Cclient-1.3.readme
80 447012 epoch: 995892221
81 447013 path: id/H/HD/HDIAS/Mail-Cclient-1.3.tar.gz
84 FIXED with sanity check and later with the integer fix.
86 * Bug: want the index files in a .recent directory
88 * Bug: lots of dot files are not deleted in time
90 * possible test case: can a delete change the timestamp? This would
91 probably break the order of events.
93 2009-03-20 Andreas J. Koenig <andk@cpan.org>
95 * 1233701831.34486 what's so special about this number/string? It
97 id/G/GR/GRODITI/MooseX-Emulate-Class-Accessor-Fast-0.00800.tar.gz and
98 atm lives in Y,Q, and Z.
100 It is the first entry after id/--skip-locking which has timestamp
101 1234164228.11325 which represents a file that doesn't exist anymore.
105 Sync 1237531537 (31547/33111/Z) id/J/JH/JHI/String-Approx-2.7.tar.gz ...
106 _bigfloatcmp called with l[1237505213.21133]r[UNDEF]: but both must be defined at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 76
107 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatcmp(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 131
108 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatlt(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/Done.pm line 110
109 File::Rsync::Mirror::Recentfile::Done::covered('File::Rsync::Mirror::Recentfile::Done=HASH(0x8857fb4)', 1237505213.21133, 0.123456789) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile.pm line 2041
110 File::Rsync::Mirror::Recentfile::uptodate('File::Rsync::Mirror::Recentfile=HASH(0x8533a2c)') called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recent.pm line 536
111 File::Rsync::Mirror::Recent::rmirror('File::Rsync::Mirror::Recent=HASH(0x82ef3d0)', 'skip-deletes', 1) called at /home/k/sources/CPAN/GIT/trunk/bin/testing-rmirror.pl line 27
112 at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 76
113 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatcmp(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 131
114 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatlt(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/Done.pm line 110
115 File::Rsync::Mirror::Recentfile::Done::covered('File::Rsync::Mirror::Recentfile::Done=HASH(0x8857fb4)', 1237505213.21133, 0.123456789) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile.pm line 2041
116 File::Rsync::Mirror::Recentfile::uptodate('File::Rsync::Mirror::Recentfile=HASH(0x8533a2c)') called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recent.pm line 536
117 File::Rsync::Mirror::Recent::rmirror('File::Rsync::Mirror::Recent=HASH(0x82ef3d0)', 'skip-deletes', 1) called at /home/k/sources/CPAN/GIT/trunk/bin/testing-rmirror.pl line 27
119 FIXED, it was the "--skip-locking" file where manual intervention was participating
121 * bug on the mirroring slave: when the dirtymark gets increased we
122 probably do not reset the done intervals. The mirrorer stays within
123 tight bounds where it tries to sync with upstream and never seems to
124 finish. In the debugging state file I see lots of identical intervals
125 that do not get collapsed. When I restart the mirrorer it dies with:
127 Sync 1237507989 (227/33111/Z) id/X/XI/XINMING/Catalyst-Plugin-Compress.tar.gz ...
128 _bigfloatcmp called with l[1237400817.94363]r[UNDEF]: but both must be defined at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 76
129 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatcmp(1237400817.94363, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 101
130 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatge(1237400817.94363, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/Done.pm line 226
131 File::Rsync::Mirror::Recentfile::Done::_register_one('File::Rsync::Mirror::Recentfile::Done=HASH(0x84c6af8)', 'HASH(0xb693f2dc)') called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/Done.pm line 200
132 File::Rsync::Mirror::Recentfile::Done::register('File::Rsync::Mirror::Recentfile::Done=HASH(0x84c6af8)', 'ARRAY(0x8c54bfc)', 'ARRAY(0xb67e618c)') called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile.pm line 1044
133 File::Rsync::Mirror::Recentfile::_mirror_item('File::Rsync::Mirror::Recentfile=HASH(0x84abcb0)', 227, 'ARRAY(0x8c54bfc)', 33110, 'File::Rsync::Mirror::Recentfile::Done=HASH(0x84c6af8)', 'HASH(0x84abd8c)', 'ARRAY(0x839cb64)', 'HASH(0x839c95c)', 'HASH(0xb6a30f0c)', ...) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile.pm line 992
134 File::Rsync::Mirror::Recentfile::mirror('File::Rsync::Mirror::Recentfile=HASH(0x84abcb0)', 'piecemeal', 1, 'skip-deletes', 1) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recent.pm line 564
135 File::Rsync::Mirror::Recent::_rmirror_mirror('File::Rsync::Mirror::Recent=HASH(0x84ab6d4)', 7, 'HASH(0x8499488)') called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recent.pm line 532
136 File::Rsync::Mirror::Recent::rmirror('File::Rsync::Mirror::Recent=HASH(0x84ab6d4)', 'skip-deletes', 1) called at /home/k/sources/CPAN/GIT/trunk/bin/testing-rmirror.pl line 27
138 and debugging stands at
144 epoch: 1237400802.5789
145 path: id/J/JO/JOHND/CHECKSUMS
148 epoch: 1237400807.97514
149 path: id/M/MI/MIYAGAWA/CHECKSUMS
152 epoch: 1237400817.94363
153 path: id/X/XI/XINMING/Catalyst-Plugin-Compress.tar.gz
161 and it is reproducable.
163 Why does the mirrorer not fetch a newer Z file? It is 18 hours old while
164 pause has a fresh one.
166 FIXED, it was the third anded term in each of the ifs in the IV block in
167 _register_one: with that we make sure that we do not stamp on valuable
170 2009-03-17 Andreas J. Koenig <andk@cpan.org>
172 * done: verified the existence of the floating point bug in bleadperl
173 and verified that switching from YAML::Syck to YAML::XS does not resolve
176 BTW, the switch was doable with
178 perl -i~ -pe 's/Syck/XS/g' lib/**/*.pm t/*.t
180 and should be considered as a separate TODO
182 * todo: integrate a dirty update with two aggregate calls before
183 unlocking for frictionless dirtying
185 * todo: start the second rsync daemon on pause
187 * todo: move index files to .recent: this cannot simply be done by
188 setting filenameroot to .recent/RECENT. Other parts of the modules rely
189 on the fact that dirname(recentfile) is the root of the mirrored tree.
191 2009-03-16 Andreas J. Koenig <andk@cpan.org>
193 * What was the resolution of the mirror.pl delete hook bug? Do we call
194 the delete hook when pause removes a file from MUIR?
196 * Today on pause: Updating 2a13fba..29f284d and installing it for
197 /usr/local/perl-5.10.0{,-RC2}
199 TURUGINA/Set-Intersection-0.01.tar.gz was the last upload before this
200 action and G/GW/GWILLIAMS/RDF-Query-2.100_01.tar.gz the first after it
202 2009-03-15 Andreas J. Koenig <andk@cpan.org>
204 * currently recent_events has the side effect of setting dirtymark
205 because it forces a read on the file. That should be transparent, so
206 that the dirtymark call always forces a cache-able(?) read.
208 * The bug below is -- after a lot of trying -- not reproducible on a
209 small script, only in the large test script. The closest to the output
215 my $x = "01237123229.8814";
220 ($l,$r) = ($1,$2) if $x =~ /(.)(.+)/;
222 $l = "1237123231.22458";
223 $r = "1237123231.22458";
225 Devel::Peek::Dump $l;
226 Devel::Peek::Dump $r;
227 Devel::Peek::Dump $x = $l <=> $r;
231 The checked in state at c404a85 fails the test with my
232 /usr/local/perl-5.10-uld/bin/perl on 64bit but curiously not with
233 /usr/local/perl-5.10-g/bin/perl. So it seems the behaviour is not even
234 in the test script always consistent.
236 * Todo: write a test that inserts a second dirty file with an already
237 existing timestamp. DONE
239 * Bug in perl 5.10 on my 64bit box:
241 DB<98> Devel::Peek::Dump $l
242 SV = PVMG(0x19e0450) at 0x142a550
244 FLAGS = (PADMY,NOK,POK,pNOK,pPOK)
246 NV = 1237123231.22458
247 PV = 0x194ce70 "1237123231.22458"\0
251 DB<99> Devel::Peek::Dump $r
252 SV = PVMG(0x19e0240) at 0x142a3e8
254 FLAGS = (PADMY,POK,pPOK)
257 PV = 0x19ff900 "1237123231.22458"\0
261 DB<100> Devel::Peek::Dump $l <=> $r
262 SV = IV(0x19ea6e8) at 0x19ea6f0
264 FLAGS = (PADTMP,IOK,pIOK)
267 DB<101> Devel::Peek::Dump $l
268 SV = PVMG(0x19e0450) at 0x142a550
270 FLAGS = (PADMY,NOK,POK,pNOK,pPOK)
272 NV = 1237123231.22458
273 PV = 0x194ce70 "1237123231.22458"\0
277 DB<102> Devel::Peek::Dump $r
278 SV = PVMG(0x19e0240) at 0x142a3e8
280 FLAGS = (PADMY,NOK,POK,pIOK,pNOK,pPOK)
282 NV = 1237123231.22458
283 PV = 0x19ff900 "1237123231.22458"\0
287 Retry with uselongdouble gives same effect. Not reproducable on 32bit box (k75).
289 * Todo: reset "done" or "covered" and "minmax" after a dirty operation?
291 2009-03-11 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
293 * $obj->merge ($other) needs to learn about equal epoch which may happen
294 since dirty_epoch intruded.
296 * Wontfix anytime soon: I think we currently do not support mkdir. Only
299 2009-01-01 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
301 * Todo: continue working on update(...,$dirty_epoch). It must be
302 followed by a fast_aggregate!
304 2008-12-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
306 * maybe we need a closest_entry or fitting_interval or something like
307 that. We want to merge an event into the middle of some recentfile.
308 First we do not know which file, then we do not know where to lock,
309 where to enter the new item, when and where to correct the dirtymark.
311 So my thought is we should first find which file.
313 Another part of my brain answers: what would happen if we would enter
314 the new file into the smallest file just like an ordinary new event,
315 just as an old event?
317 (1) we would write a duplicate timestamp? No, this would be easy to
320 (2) we would make the file large quickly? Yes, but so what? We are
321 changing the dirtymark, so are willing to disturb the downstream hosts.
323 2008-11-22 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
325 * 10705 root 17 0 725m 710m 1712 S 0.0 46.8 834:56.05 /home/src/perl/repoperls/installed-perls/perl/pVNtS9N/perl-5.8.0@32642/bin/perl -Ilib /home/k/sources/CPAN/GIT/trunk/bin/testing-rmirror.pl
329 https://rt.cpan.org/Ticket/Display.html?id=41199
331 * bzcat uploads.csv.bz2 | perl -F, -nale '$Seen{$F[-1]}++ and print'
333 Strangest output being HAKANARDO who managed to upload
335 Here is a better oneliner that includes also the first line of each
338 bzcat uploads.csv.bz2 | perl -MYAML::Syck -F, -nale '$F[-1]=~s/\s+\z//; push @{$Seen{$F[-1]}||=[]},$_; END {for my $k (keys %Seen){ delete $Seen{$k} if @{$Seen{$k}}==1; } print YAML::Syck::Dump(\%Seen)}'
340 2008-10-31 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
342 * memory leak in the syncher? It currently weighs 100M.
346 root 10705 1.0 4.9 80192 76596 pts/32 S+ Nov02 24:05 /home/src/perl/repoperls/installed-perls/perl/pVNtS9N/perl-5.8.0@32642/bin/perl -Ilib /home/k/sources/CPAN/GIT/trunk/bin/testing-rmirror.pl
349 2008-10-29 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
351 * lookup by epoch and by path and use this ability on the pause to never
352 again register a file twice that doesn't need it. Let's call it
355 * after the dirtymark is done: fill up recentfiles with fake (historic)
356 entries; fill up with individual corrections; algorithm maybe to be done
357 with bigfloat so that we can always place something in the middle
358 between two entries. Before we must switch to bigfloat we could try to
359 use Data::Float::nextup to get the.
361 * Inotify2 on an arbitrary tree and then play with that instead of PAUSE
364 * dirtymark now lives in Recentfile, needs to be used in rmirror.
366 * find out why the downloader died after a couple of hours without a net
367 connection. Write a test that survives the not-existence of the other
370 2008-10-15 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
372 * reconsider the HTTP epoch only. Not the whole thing over HTTP because
373 it makes less sense with tight coupling for secondary files. But asking
374 the server what the current epoch is might be cheaper on HTTP than on
375 rsync. (Needs to be evaluated)
377 * remove the 0.00 from the verbose overview in the Merged column in the
380 * write tests that expose the problems of the last few days: cascading
381 client/server roles, tight coupling for secondary RFs, deletes after
384 * Some day we might want to have policy options for the slave:
385 tight/loose/no coupling with upstream for secondary RFs. tight is what
386 we have now. loose would wait until a gap occurs that can be closed.
388 2008-10-14 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
390 * revisit all $rfs->[$i+1] places if they now make sense still
392 2008-10-11 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
394 * another bug is the fact that the mirror command deletes files before
395 it unhides the index file, thus confusing downstream slaves. We must not
396 delete before unhiding and must delete after unhiding. FIXED.
398 * new complication about the slave that is playing a server role.
399 Currently we mirror from newest to oldest with a hidden temporary file
400 as index. And when one file is finished, we unhide the index file.
401 Imagine the cascading server/slave is dead for a day. It then starts
402 mirroring again with the freshest thing and unhides the freshest index
403 file when it has worked through it. In that moment it exposes a time
404 hole. Because it now works on the second recentfile which is still
407 We currently do nothing special to converge after such a drop out. At
408 least not intentionally and robustly and thought through.
410 The algorithm we use to seed the next file needs quite a lot of more
411 robustness than it currently has. Something to do with looking at the
412 merged element of the next rf and when it has dropped off, we seed
413 immediately. And if it ramains dropped off, we seed again, of course.
415 Nope, looking from smaller to larger RFS we look at the merged element
416 of this RF and at the minmax/max element of the next RF. If that
417 $rf[next]->{minmax}{max} >= $rf[this]->{merged}{epoch}, then we can stop
420 And we need a public accessor seed and unseed or seeded. But not the mix
421 of public and private stuff that then is used behind the back.
423 And then the secondary* stuff must go.
425 And we must understand what the impact is on the DONE system. Can it go
426 unnoticed that there was a hole? And could the DONE system have decided
427 the hole is covered? This should be testable with three directories where
428 the middle stops working for a while. Done->merge is suspicious, we must
429 stop it from merging non-conflatable neighbors due to broken continuity.
433 2008-10-10 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
435 * Slaven suggests to have the current epoch or the whole current
436 recentfile available from the HTTP server and take it away with
437 keepalive. This direction goes the granularity down to subseconds.
439 We might want to rewrite everything to factor out transport and allow
440 the whole thing to run via HTTP.
442 2008-10-09 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
444 * are we sure we do NOT LEAVE DOT FILES around? Especially on the
447 * smoker on k81 fetching from k75 to verify cascading works. See
448 2008-07-17 in upgradexxx and rsync-over-recentfile-3.pl.
450 * maybe the loop should wait for CHECKSUMS file after every upload. And
451 CPAN.pm needs to deal with timestamps in the future.
453 * do not forget the dirtymark!
455 Text: have a new flag on recentfiles with the meaning: if this
456 changes, you're required to run a full rsync over all the files. The
457 reason why we set it would probably be: some foul happened. we injected
458 files in arbitrary places or didn't inject them although they changed.
459 The content of the flag? Timestamp? The relation between the
460 recentfiles would have to be inheritance from the principal, because any
461 out of band changes would soon later propagate to the next recentfile.
463 By upping the flag often one can easily ruin the slaves.
465 last out of band change? dirtymark?
467 Anyway, this implies that we read a potentially existing recentfile
470 And it implies that we have an eventloop that keeps us busy in 2-3
471 cycles, one for current stuff (tight loop) and one for the recentfiles
472 (cascade when principal has changed), one for the old stuff after a
475 And it implies that the out-of-band change in any of the recentfiles
476 must have a lock on the principal file and there is the place to set the
479 * start a FAQ, especially quick start guide questions. Also to aid those
480 problematic areas where we have no good solution, like the "links"
483 * wish feedback when we are slow.
487 * Remove a few DEBUG statements.
489 * The multiple-rrr way of doing things needs a new option to rmirror,
490 like piecemeal or so. Not urgent because after the first pass through,
491 things run smoothely. It's only ugly during the first pass.
493 * I have the suspicion that the code is broken that decides if the
494 neighboring RF needs to be seeded. I fear when too much time has gone
495 between two calls (in our case more than one hour), it would not seed
496 the neighbor. Of course this will never be noticed, so we need a good
499 * local/localroot confusion: I currently pass both options but one must
502 * accounts for early birds on PAUSE rsync daemon.
504 * hardcoded 20 seconds
506 * who mirrors the index? DOING now.
508 * which CPAN mirrors offer rsync?
510 * visit all XXX, visit all _float places
512 * rename the pathdb stuff, it's too confusing. No idea how.
514 * rrr-inotify, backpan, rrr-register
516 2008-10-08 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
518 * current bugs: the pathdb seems to get no reset, the seeding of the
519 secondaryttl stuff seems not to have an effect. Have helped myself with
520 a rand(10), need to fix this back. So not checked in. Does the rand
523 The rand thing helps. The secondaryttl stuff was in the wrong line,
526 The pathdb stuff was because I called either _pathdb or __pathdb on the
527 wrong object. FIXED now.
529 * It's not so beautiful if we never fetch the recentfiles that are not
530 the principal, even if this is correct behaviour. We really do not need
531 them after we have fetched the whole content.
533 OK, we want a switch for that: secondaryttl DONE
535 2008-10-07 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
537 * bug: rrr-news --max does not count correctly. with "35" it shows me 35
538 lines but with 36 it shows 110. First it repeats 35, gives 70, and then
539 it lets 40 follow. FIXED
541 * See that the long running process really only updates the principal
542 file unless it has missed a timespan during which something happened. If
543 nothing happened, it must notice even when it misses the timespan. DONE
545 * we must throw away the pathdb when we have reached the end of Z. From
546 that moment we can have a very small pathdb because the only reason for
547 a pathdb is that we know to ignore old records in old files. We won't
548 need this pathdb again before the next full pass over the data is
549 necessary and then we will rebuild it as we go along. DONE
551 2008-10-06 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
553 * I think, Done::register_one is doing wrong in that it does not
554 conflate neighboring pieces. The covered() method cannot do this because
555 it has no recent_events array at hand. But register_one has it and could
556 do it and for some reason misses to do it (sometimes).
558 This means that the three tests I just wrote can probably not survive
559 because they test with an already broken Done structure.
561 The art now is to detect how it happens, then to reproduce, then write a
564 So from the logfile this is what happens: we have a good interval with
565 newest file being F1 at T1. Now remotely F1 gets a change and F2 goes on
566 top of it. Locally we now mirror F2 and open a new done interval for it.
567 Then we mirror F1 but this time with the timestamp T1b. And when we then
568 try to close the gap, we do not find T1 but instead something older. We
569 should gladly accept this older piece and this would fix this bug.
573 * bug to fix: when the 1h file changes while rmirror is running, we do
574 correctly sync the new files but never switch to the 6h file but rather
575 stay in a rather quick loop that fetches the 1h file again and again.
577 Is it possible that we initialize a new object? Or does
578 get_remote_recentfile_as_tempfile overwrite something in myself?
580 Want a new option: _runstatusfile => $file which frequently dumps the
581 state of all recentfiles to a file.
585 2008-10-04 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
587 * Todo: now teach update to verify the timestamp is about to write
588 against the previous and use _increase_a_bit if it doesn't comply with
589 strict monotony. DONE
591 * The problem of rounding. So far perl's default precision was
592 sufficient. One day it won't be. FakeFloat has an easy job when it is
593 only reading and other machines have written correctly. But when we want
594 to write a floating point number that is a bit larger than the other
595 one, then we need our own idea of precision.
597 Slaven said: just append a "1". This might be going towards the end of
598 usability too quickly. I'd like something that actually uses the decimal
599 system. Well, appending a 1 also does this but...
601 E.g. we have 1.0. nextup on this architecture is starting with
602 1.0000000000000004. So there is a gap to fill: 1,2,3. Now I have
603 taken the 1.0000000000000003 and the next user comes and the time tells
604 him 1.0 again. He has to beat my number without stepping over the
605 nextup. This is much less space than I had when I chose 1,2,3.
607 What is also irritating is that nextup is architecture dependent. The
608 128 bit guy must choose very long numbers to fit in between whereas the
609 other one with 16 bit uses larger steps. But then the algorithm is the
610 same for both, so that would be a nice thing.
612 I see two situation where we need this. One is when Time::HiRes returns
613 us a value that is <= the last entry in our recentfile. In this case
614 (let's call it the end-case) we must fill the region between that number
615 and the next higher native floating point number. The other is when we
616 inject an old file into an old recentfile (we would then also set a new
617 dirtymark). We find the integer value already taken and need a slightly
618 different one (let's call it the middle-case). The difference between
619 the two situations is that the next user will want to find something
620 higher than my number in the end-case and something lower than my number
623 So I suggest we give the function both a value and an upper bound and it
624 calculates us a primitive middle. The upper bound in the middle-case is
625 the next integer. The upper bound on the end-case is the nextup floating
626 point number. But the latter poses another problem: if we have occupied
627 the middle m between x and nextup(x), then the nextup(m) will probably
628 not be the same as nextup(x) because some rounding will take place
629 before the nextup is calculated and when the rounding reaches the
630 nextup(x), we will end up at nextup(nextup(x)).
632 So we really need to consider the nextup and the nextdown from there and
633 then the middle and that's the number we may approach asymptotically.
636 2008-10-03 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
638 * consider deprecating the use of RECENT.recent as a symlink. It turns
639 out to need extra hoops with the rsync options and just isn't worth it.
640 Or maybe these extra hoops are needed anyway for the rest of the tree?
641 Nope, can't be the case because not all filesystems support symlinks.
643 But before doing the large step, I'll deprecate the call of
644 get_remote_recentfile_as_tempfile with an argument. Rememberr this was
645 only introduced to resolve RECENT.recent and complicates the routine far
646 beyond what it deserves.
648 DONE. Won't deprecate RECENT.recent, just moved its handling to the
651 2008-10-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
653 * I think it's a bug that the rsync_option links must be set to true in
654 order to support RECENT.recent and that nobody cares to set it
655 automatically. Similar for ignore_link_stat_errors.
657 2008-09-27 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
659 * Todo: find all todos together and make a plan what is missing for a
662 - verifytree or something like that. fsck maybe.
664 - rersyncrecent, the script itself? What it do?
666 - a way to only mirror the recentfiles without mirroring the whole
667 remote system such that people can decide to mirror only partially see
668 also 2008-08-30. .shadow-xxx directory? this also needed for a
669 filesystem that is still incomplete and might need the mirrorfiles for
672 - long living objects that mirror again and again. Inject something
673 into ta, see how it goes over to tb.
675 - how do we continue filling up the DONE system when we use an object
676 for the second time? "fully covered" and "uptodate" or new terminology.
678 - overview called on the wrong file should be understandable
680 - the meta data field that must change when we fake something up so that
681 the downstream people know they have to re-fetch everything.
683 - how tolerant are we against missing files upstream? how do we keep
684 track? there are legitimate cases where we did read upstream index right
685 before a file got deleted there and then find that file as new and want
686 it. There are other cases that are not self healing and must be tracked
689 - how, exactly, do we have to deal with deletes? With rsync errors?
691 rsync: link_stat "/id/K/KA/KARMAN/Rose-HTMLx-Form-Related-0.07.meta" (in
692 authors) failed: No such file or directory (2)
694 The file above is a delete in 1h and a new in file 1M and the
695 delete in the locally running rmirror did not get propagated to the 1M
696 object. Bug. And the consequence is a standstill.
698 It seems that a slave that works with a file below the principal needs
699 to merge things all the way up to get rid of later deletes. Or keep
700 track of all deletes and skip them later. So we need a trackdeletes.pm
701 similar to the done.pm?
703 see also 2008-08-20 about spurious deletes that really have no add
704 counterpart and yet they are not wrong.
706 - consider the effect when resyncing the recentfile takes longer than
707 the time per loop. Then we never rsync any file. We need to diagnose
708 that and force an increase of that loop time. But when we later are fast
709 enough again because the net has recovered, then we need to switch back
710 to original parameters. ERm, no, it's enough to keep syncing at least
711 one file before refetching an index file.
713 - remember to verify that no temp files are left lying around and the
716 - status file for not long running jobs that want to track upstream with
719 - revisit all XXX _float areas and study Sub::Exporter DONE
721 - persistent DB even though we just said we do not need it. Just for
722 extended capabilities and time savings when, for example, upstream
723 announces a reset and we get new recentfiles and could then limit
724 ourselves to a subset of files (those that have a changed epoch) in a
725 first pass and would only then do the loop to verify the rest. Or
728 * Todo: aggregate files should know their feed and finding the principal
729 should be done stepwise. (?)
731 * Todo: DESTROY thing that unlocks. Today when I left the debuggerr I
732 left locks around. DONE
734 2008-09-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
736 * maybe extend the _overview so that it always says if and where the
737 last file is in the next file and where the next event in the next rf
738 would lie. No, don't like this anymore. REJECT
740 * take the two new redundant tests out again, only the third must
743 * Todo: add a sanity check if the merged structure is really pointing to
744 a different rf and that this different rf is larger. DONE
746 2008-09-25 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
748 * now test, if they are overlapping. And test if there is a file in the
749 next rf that would fit into this rf's interval.
751 1h 1222324012.8474 1222322541.7963 0.4086
752 6h 1222320411.2760 1222304207.6931 4.5010 missing overlap/gap!
753 1d 1222320411.2760 1222238750.5071 22.6835 large overlap
754 1W 1222313218.3626 1221708477.5829 167.9835
756 I suspect that somebody writes a merged timestamp without having merged
757 and then somebody else relies on it.
759 If aggregate is running, the intervals must not be extravagated, if it
760 is not running, there must not be bounds, the total number of events in
761 the system must be counted and must be controlled throughout the tests.
762 That the test required the additional update was probably nonsense,
763 because aggregate can cut pieces too. FIXED & DONE
765 2008-09-23 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
767 * rrr-aggregate seems to rewrite the RECENT file even if nothing has
770 2008-09-21 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
772 * Most apparent bug at the moment is that the recentfiles are fetched
773 too often. Only the principal should be fetched and if it has not
774 changed, the others should not be refetched. ATM I must admit that I'm
775 happy that we refetch more often than needed because I can more easily
776 fix bugs while the thing is running.
778 * Let's say, 1220474966.19501 is a timestamp of a file that is already
779 done but the done system does not know about it. The reason for the
780 failure is not known and we never reach the status uptodate because of
781 this. We must get over it.
783 Later it turns out that the origin server had a bug somewhere.
784 1220474966.19042 came after 1220474966.19501. Or better: it was in the
785 array of the recentfile one position above. The bug was my own.
787 2008-09-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
789 * There is the race condition where the server does a delete and the
790 slave does not yet know and then tries to download it because he sees
791 the new. So for this time window we must be more tolerant against
792 failure. If we cannot download a file, we should just skip it and should
793 not retry immediately. The whole system should discover the lost thing
794 later. Keeping track with the DONE system should really be a no brainer.
796 But there is something more: the whole filesystem is a database and the
797 recentfiles are one possible representation of it. It's a pretty useful
798 representation I think that's why I have implemented something around
799 it. But for strictly local operation it has little value. For local
800 operation we would much rather have a database. So we would enter every
801 recentfile reading and every rsync operation and for every file the last
802 state change and what it leads to. Then we would always ignore older
803 records without the efforts involved with recentfiles.
805 The database would have: path,recentepoch,rsyncedon,deletedon
807 Oh well, not yet clear where this leads to.
809 2008-09-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
811 * Bug: the bigloop ran into a funny endless loop after EWILHELM uploaded
812 Module-Build. It *only* rsynced the "1h" recentfile from that moment on.
814 * statusfile, maybe only on demand, alone to have a sharp debugging
815 tool. It is locked and all recentfiles dump themselves into it and we
816 can build a viewer that lets us know where we stand and what's inside.
818 * remember: only the principal recentfile needs expiration, all others
819 shall be expired by principal if it discovers that something has move
822 2008-09-18 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
824 * Always check if we stringify to a higher value than in the entry
827 * And in covered make an additional check if we would be able to see a
828 numerical difference between the two numbers and if we can't then switch
829 to a different, more expensive algorithm. Do not want to be caught by
830 floating surprises. DONE
832 2008-09-17 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
834 * caching has several aspects here: we can cache the interval of the
835 recentfile which only will change when the mtime of the file changes. We
836 must re-mirror the recentfile when its ttl has expired. Does have_read
837 tell you anything? It counts nothing at all. Only the mtime is
838 interesting. The ntuple mtime, low-epoch, high-epoch. And as a separate
839 thing the have_mirrored because it is unrelated to the mtime.
841 * Robustness of floating point calculations! I always thought that the
842 string calculated by the origin server for the floating representation
843 of the epoch time is just a string. When we convert it to a number and
844 later back to a string, the other computer might come to a different
845 conclusion. This must not happen, we want to preserve it under any
846 circumstances. I will have to write tests with overlong sequences that
847 get lost in arithmetic and must see if all still works well. DONE
849 But one fragile point remains: if one host considers a>b and the other
850 one considers them == but no eq. To prevent this, we must probably do
851 some extra homework. DONE
853 2008-09-16 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
855 * the concept of tracking DONE needs an object per recentfile that has
856 something like these methods:
858 do_we_have(xxx), we_have(xxx), do_we_have_all(xxx,yyy), reset()
860 covered() register() covered()
862 The unclear thing is how we translate points in time into intervals. We
863 could pass a reference to the current recent_events array when running
864 we_have(xxx) and let the DONE object iterate over it such that it only
865 has to store a list of intervals that can melt into each other. Ah, even
866 passing the list together with a list of indexes seems feasiable.
868 Or maybe ask for the inverted list?
870 Whenever the complete array is covered by the interval we say we are
871 fully covered and if the recentfile is not expired, we are uptodate.
873 2008-09-07 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
875 2008-09-05 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
877 * need a way to "return" the next entry after the end of a list. When
878 the caller says "before" or "after" we would like to know if he could
879 cover that interval/threshold or not because this influences the effect
880 of a newer timestamp of that recentfile. DONE with $opt{info}.
882 2008-09-04 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
884 * one of the next things to tackle: the equivalent of csync2 -TIXU.
886 loop implies tixu (?). Nope, something like --statefile decides. Per
889 T test, I init, X including removals, U nodirtymark
891 So we have no concept of dirtymarks, we only trust that since we are
892 running we have observed everything steadily. But people will not let
893 this program run forever so we must consider both startup penalty and
894 book keeping for later runs. We keep this for later. For now we write a
895 long running mirror that merges several intervals.
897 2008-09-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
899 * need to speed up the 02 test, it's not clever to sleep so much. Reduce
902 * rersyncrecent, the script: default to one week. The name of the switch
903 is --after. Other switches? --loop!
905 2008-08-30 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
907 * need a switch --skip-deletes (?)
909 * need a switch --enduser that tells us that the whole tempfile
910 discipline is not needed when there is no downstream user. (?)
912 Without this switch we cannot have a reasonable recent.pl that just
913 displays the recent additions. Either we accept to download everything.
914 Or we download temporary files without the typical rsync protocol
917 Or maybe the switch is --tmpdir? If --tmpdir would mean: do not use
918 File::Temp::tempdir, this might be a win.
920 2008-08-29 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
922 * apropos missing: we have no push, we never know the downstream
923 servers. People who know their downstream hosts and want to ascertain
924 something will want additional methods we have never thought about, like
925 update or delete a certain file.
927 2008-08-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
929 * tempted to refactor rmirror into resolve_symlink, localize, etc.
930 Curious if rsync_options=links equal 0 vs. 1 will make the expected
933 * rsync options: it's a bit of a pain that we usually need several rsync
934 options, like compress, links, times, checksum and that there is no
935 reasonable default except the original rsync default. I think wee can
936 safely assume that the rsync options are shared between all recentfile
937 instances within one recent tree.
939 2008-08-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
941 * deletes: if a delete follows an add quickly enough it may happen that
942 a downstream mirror did not see the add at all! It seems this needs to
943 be mentioned somewhere. The point here is that even if the downstream is
944 never missing the principal timeframe it may encounter a "delete" that
945 has no complimentary "add" anywhere.
947 2008-08-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
949 * I suspect the treat of metadata is incorrect during read or something.
950 The bug that I am watching is that between 06:08 and 06:09 the 6h file
951 contained more than 6 hours worth of data. At 06:08 we merged into the
952 1d file. We need to take snapshots of the 6h file over the course of an
953 hour or maybe only between XX:08 and XX:09? Nope, the latter is not
956 Much worse: watching the 1h file: right at the moment (at 06:35) it
957 covers 1218867584-1219120397 which is 70 hours.
959 Something terribly broken. BTW, 1218867584 corresponds to Sat Aug 16
960 08:19:44 2008, that is when I checked out last time, so it seems to be
961 aggregating and never truncating?
963 No, correct is: it is never truncating; but wrong is: it is aggregating.
964 It does receive a lot of events from time to time from a larger file.
965 Somehow a large file gets merged into the small one and because the
966 "meta/merged" attribute is missing, nobody is paying attention. I
967 believe that I can fix this by making sure that metadata are honoured
968 during read. DONE and test adjusted.
970 2008-08-17 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
972 * grand renaming plan
974 remotebase => remoteroot to fit well with localroot DONE
975 local_path() => localroot seems to me should already work DONE
976 recentfile_basename => rfilename no need to stress it has no slash DONE
978 filenameroot??? Doesn't seem too bad to me today. Maybe something like
979 kern? It would anyway need a deprecation cycle because it is an
980 important constructor.
982 * I like the portability that Data::Serializer brings us but the price
983 is that some day we might find out that it is slowing us a bit. We'll
986 2008-08-16 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
988 * should we not enter the interval of the principal (or the interval of
989 the merging file?) in every aggregated/merged file?
991 * we should aim at a first release and give up on thinking about
992 sanitizing stuff and zloop. Let's just admit that a full traditional
993 rsync is the only available sanitizer ATM. Otherwise it's complicated
994 stuff: sanitizing on the origin server, sanitizing on the slaves,
995 sanitizing forgotten files, broken timestamps, etc. Let's delay it and
996 get the basics out before this becomes a major cause for mess.
998 2008-08-13 Andreas Koenig <k@andreas-koenigs-computer.local>
1000 * On OSes not supporting symlinks we expect that RECENT.recent contains
1001 the contents of the principal recentfile. Actually this is identical on
1002 systems supporting symlinks. Simple, what follows from that is that we
1003 need to keep the serializer in the metadata because we cannot read it
1004 from the filename, doesn't it? Of course not. It's a chicken and egg
1005 problem. This leaves us with the problem to actually parse the
1006 serialized data to find out in which format it is. So who can do the 4
1007 or 5 magics we wanted to support? File::LibMagic?
1009 2008-08-09 Andreas Koenig <k@andreas-koenigs-computer.local>
1011 * remotebase and recentfile_basename are ugly names. Now that we need a
1012 word for the shortest/principal/driving recentfile too we should do
1015 localroot is good. rfile is good. local_path() is bad, local_path($path)
1016 is medium, filenameroot() is bad, remotebase is bad, recentfile is
1019 Up to now remotebase was the string that described the remote root
1020 directory in rsync notation, like pause.perl.org::authors. And
1021 recentfile_basename was "RECENT-1h.yaml".
1023 2008-08-08 Andreas Koenig <k@andreas-koenigs-computer.local>
1025 * The test that was added in today's checkin is a good start for a test
1026 of rmirror. We should have more methods in Recent.pm: verify,
1027 addmissingfiles. We should verify the current tree, then rmirror it and
1028 then verifytree the copy. We could then add some arbitrary file and let
1029 it be discovered by addmissingfiles, then rmirror again and then
1030 verifytree the copy again.
1032 Then we could start stealing from csync2 sqlite database [no port to
1033 OSX!] and fill a local DB. And methods to compare the database with the
1034 recentfiles. Our strength is that in principle we could maintain state
1035 with a single float. We have synced up to 1234567890.123456. If the Z
1036 file does not add new files all we have to do is mirror the new ones and
1039 This makes it clear that we should extend current protocol and declare
1040 that we cheat when we add files too late, just to help the other end
1041 keeping track. Ah yes, that's what was meant when zloop was mentioned
1044 Maybe need to revisit File::Mirror to help me with this task.
1046 2008-08-07 Andreas Koenig <k@andreas-koenigs-computer.local>
1048 * There must be an allow-me-to-truncate flag in every recentfile.
1049 Without it one could construct a sequence of updates winning the locking
1050 battle against the aggregator. Only if an aggregator has managed to
1051 merge data over to the next level, truncating can be allowed. DONE with
1054 2008-08-06 Andreas Koenig <k@andreas-koenigs-computer.local>
1056 * We should probably guarantee that no duplicates enter the aggregator
1059 2008-08-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1061 * To get merge operation faster would need a good benchmark test. What
1062 02 spits out isn't reliable enough and is dominated by many other
1065 commit 10176bf6b79865d4fe9f46e3857a3b8669fa7961
1066 Author: Andreas J. Koenig <k@k75.(none)>
1067 Date: Sat Aug 2 07:58:04 2008 +0200
1071 commit 3243120a0c120aaddcd9b1f4db6689ff12ed2523
1072 Author: Andreas J. Koenig <k@k75.(none)>
1073 Date: Sat Aug 2 11:40:29 2008 +0200
1075 there was a lot of trying but the effect is hardly measurable with
1078 * overhead of connecting seems high. When setting
1079 max_files_per_connection to 1 we see that.
1081 2008-08-01 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1083 * 1217622571.0889 - 1217597432.86734 = 25138.2215600014
1085 25138.2215600014/3600 = 6.98283932222261
1087 It jumps into the eye that this is ~ 7 hours, not ~6, so there seems to
1088 be a bug in the aggregator. FIXED
1090 2008-07-27 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1092 * e.g. id/Y/YE/YEWENBIN/Emacs-PDE-0.2.16.tar.gz: Do we have it, should
1093 we have it, can we mirror it, mirror it!
1095 I fear this needs a new class which might be called
1096 File::Rsync::Mirror::Recent. It would collect all recentfiles of a kind
1097 and treat them as an entity. I realize that a single recentfile may be
1098 sufficient for certain tasks and that it is handy for the low level
1099 programmer but it is not nice to use. If there is a delete in the 1h
1100 file then the 6h file still contains it. Seekers of the best information
1101 need to combine at least some of the recentfiles most of the time.
1103 There is the place for the Z loop!
1105 But the combination is something to collect in a database, isn't it. Did
1106 csync2 just harrumph?
1108 2008-07-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1110 * it just occurred to me that hosts in the same mirroring pool could
1111 help out each other even without rewriting the recentfile. Just fetch
1112 the stuff to mirror from several places, bingo. But that's something
1113 that should rather live in a separate package or in rsync directly.
1115 * cronjobs are unsuited because with ntp they would all come at the full
1116 minute and disturb each other. Besides that I'd hate to have a backbone
1117 with more than a few seconds latency.
1119 2008-07-25 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1121 * a second rsync server with access control for PAUSE. Port? 873 is the
1122 standard port, let's take 8873.
1124 * if there were a filesystem based on this, it would have a slow access
1125 to inexistent files. It would probably provide wrong readdir (only based
1126 on current content) or also a slow one (based on a recentfile written
1127 after the call). But it would provide fast access to existing files. Or
1128 one would deliberately allow slightly blurred answers based on some
1129 sqlite reflection of the recentfiles.
1131 * todo: write a variant of mirror() that combines two or more
1132 recentfiles and treats them like one
1134 * todo: signal handler to remove the tempfile
1136 2008-07-24 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1138 * now that we have the symlink I forgot how it should be used in
1141 * the z loop: add missing files to Z file. Just append them (instead of
1142 prepending). So one guy prepends something from the Y file from time to
1143 time and another guy appends something rather frequently. Collecting
1144 pond. When Y merges into Z, things get epoch and the collecting pond
1145 gets smaller. What exactly are "missing files"?
1147 take note of current epoch of the alpha file, let's call it the
1150 find all files on disk
1152 remove all files registered in the recentworld up to recent-ts
1154 remove all files that have been deleted after recent-ts according to
1157 2008-07-23 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1159 * rersyncrecent might be a cronjob with a (locked) state file which
1160 contains things like after and maybe last z sync or such?
1162 rrr-mirror might be an alternative name but how would we justify the
1163 three Rs when there is no Re-Rsync-Recent?
1165 With the --loop parameter it is an endless loop, without it is no loop.
1166 At least this is simple.
1168 * todo: new accssor z-interval specifies how often the Z file is updated
1169 against the filesystem. We probably want no epoch stamp on these
1170 entries. And we want to be able to filter the entries (e.g. no
1171 by-modules and by-category tree)
1173 2008-07-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1175 * Fill the Z file. gc or fsck or both. Somehow we must get the old files
1176 into Z. We do not need the other files filled up with filesystem
1179 * need interface to query for a file in order to NOT call update on
1180 PAUSE a second time within a short time.
1182 2008-07-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1184 * recommended update interval? Makes no sense, is different for
1191 change-log-default-name: "Todo"