1 2009-03-29 Andreas J. Koenig <andk@cpan.org>
3 * known bug: cannot configure the Recentfile objects to keep delete
4 events. How to call the accessor for that? keep_delete_objects_forever?
6 Sounds acceptable. FIXED (but untested).
8 * known bug: keeps temporary index files lying around.
10 2009-03-24 Andreas J. Koenig <andk@cpan.org>
14 for ( @{$rrr->recentfiles} ) { $_->verbose(0) }
18 2009-03-22 Andreas J. Koenig <andk@cpan.org>
20 * k81 is again client of k75 with different parameters.
22 * broken now: (1) seeding and unseeding: rmirror is seeding and talking
23 about it all the time and nobody reacts accordingly; (2) lots of temp
24 files get created and not removed; culprit the new call to
25 get_remote_recentfile_as_tempfile within Recent.pm; the manpage says the
26 caller has to remove the tempfile after use.
28 Need the drawing board.
30 PARTLY FIXED: retracting the idea to call
31 get_remote_recentfile_as_tempfile from Recent.pm and moving around seed
32 and unseed such that uptodate now means "we have mirrored this
33 recentfile's bunch of files" and seeded means "somebody believes the
34 index file for this rf needs be refreshed". This seems to work now.
36 Still collecting temp files that nobody cares to depollute.
38 2009-03-21 Andreas J. Koenig <andk@cpan.org>
40 * Bug?: should it be harder than it is atm to set the timestamp to the
43 * bug with native integers:
47 path: id/D/DE/DELTA/Crypt-Rijndael_PP-0.03.readme
51 path: id/L/LG/LGODDARD/Tk-Wizard-2.124.readme
54 Native integer broke when native math was turned off. FIXED
56 * Bug: something between id/P/PH/PHISH/CGI-XMLApplication-1.1.2.readme
57 and id/C/CH/CHOGAN/HTML-WWWTheme-1.06.readme. Mirroring the Z file loops
60 Yes, records out of order:
63 447004 epoch: 995885533
64 447005 path: id/P/PH/PHISH/CGI-XMLApplication_0.9.3.readme
67 447008 epoch: 995890358
68 447009 path: id/H/HD/HDIAS/Mail-Cclient-1.3.readme
71 447012 epoch: 995892221
72 447013 path: id/H/HD/HDIAS/Mail-Cclient-1.3.tar.gz
75 FIXED with sanity check and later with the integer fix.
77 * Bug: want the index files in a .recent directory
79 * Bug: lots of dot files are not deleted in time
81 * possible test case: can a delete change the timestamp? This would
82 probably break the order of events.
84 2009-03-20 Andreas J. Koenig <andk@cpan.org>
86 * 1233701831.34486 what's so special about this number/string? It
88 id/G/GR/GRODITI/MooseX-Emulate-Class-Accessor-Fast-0.00800.tar.gz and
89 atm lives in Y,Q, and Z.
91 It is the first entry after id/--skip-locking which has timestamp
92 1234164228.11325 which represents a file that doesn't exist anymore.
96 Sync 1237531537 (31547/33111/Z) id/J/JH/JHI/String-Approx-2.7.tar.gz ...
97 _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
98 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatcmp(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 131
99 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatlt(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/Done.pm line 110
100 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
101 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
102 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
103 at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 76
104 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatcmp(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 131
105 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatlt(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/Done.pm line 110
106 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
107 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
108 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
110 FIXED, it was the "--skip-locking" file where manual intervention was participating
112 * bug on the mirroring slave: when the dirtymark gets increased we
113 probably do not reset the done intervals. The mirrorer stays within
114 tight bounds where it tries to sync with upstream and never seems to
115 finish. In the debugging state file I see lots of identical intervals
116 that do not get collapsed. When I restart the mirrorer it dies with:
118 Sync 1237507989 (227/33111/Z) id/X/XI/XINMING/Catalyst-Plugin-Compress.tar.gz ...
119 _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
120 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatcmp(1237400817.94363, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 101
121 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatge(1237400817.94363, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/Done.pm line 226
122 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
123 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
124 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
125 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
126 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
127 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
129 and debugging stands at
135 epoch: 1237400802.5789
136 path: id/J/JO/JOHND/CHECKSUMS
139 epoch: 1237400807.97514
140 path: id/M/MI/MIYAGAWA/CHECKSUMS
143 epoch: 1237400817.94363
144 path: id/X/XI/XINMING/Catalyst-Plugin-Compress.tar.gz
152 and it is reproducable.
154 Why does the mirrorer not fetch a newer Z file? It is 18 hours old while
155 pause has a fresh one.
157 FIXED, it was the third anded term in each of the ifs in the IV block in
158 _register_one: with that we make sure that we do not stamp on valuable
161 2009-03-17 Andreas J. Koenig <andk@cpan.org>
163 * done: verified the existence of the floating point bug in bleadperl
164 and verified that switching from YAML::Syck to YAML::XS does not resolve
167 BTW, the switch was doable with
169 perl -i~ -pe 's/Syck/XS/g' lib/**/*.pm t/*.t
171 and should be considered as a separate TODO
173 * todo: integrate a dirty update with two aggregate calls before
174 unlocking for frictionless dirtying
176 * todo: start the second rsync daemon on pause
178 * todo: move index files to .recent: this cannot simply be done by
179 setting filenameroot to .recent/RECENT. Other parts of the modules rely
180 on the fact that dirname(recentfile) is the root of the mirrored tree.
182 2009-03-16 Andreas J. Koenig <andk@cpan.org>
184 * What was the resolution of the mirror.pl delete hook bug? Do we call
185 the delete hook when pause removes a file from MUIR?
187 * Today on pause: Updating 2a13fba..29f284d and installing it for
188 /usr/local/perl-5.10.0{,-RC2}
190 TURUGINA/Set-Intersection-0.01.tar.gz was the last upload before this
191 action and G/GW/GWILLIAMS/RDF-Query-2.100_01.tar.gz the first after it
193 2009-03-15 Andreas J. Koenig <andk@cpan.org>
195 * currently recent_events has the side effect of setting dirtymark
196 because it forces a read on the file. That should be transparent, so
197 that the dirtymark call always forces a cache-able(?) read.
199 * The bug below is -- after a lot of trying -- not reproducible on a
200 small script, only in the large test script. The closest to the output
206 my $x = "01237123229.8814";
211 ($l,$r) = ($1,$2) if $x =~ /(.)(.+)/;
213 $l = "1237123231.22458";
214 $r = "1237123231.22458";
216 Devel::Peek::Dump $l;
217 Devel::Peek::Dump $r;
218 Devel::Peek::Dump $x = $l <=> $r;
222 The checked in state at c404a85 fails the test with my
223 /usr/local/perl-5.10-uld/bin/perl on 64bit but curiously not with
224 /usr/local/perl-5.10-g/bin/perl. So it seems the behaviour is not even
225 in the test script always consistent.
227 * Todo: write a test that inserts a second dirty file with an already
228 existing timestamp. DONE
230 * Bug in perl 5.10 on my 64bit box:
232 DB<98> Devel::Peek::Dump $l
233 SV = PVMG(0x19e0450) at 0x142a550
235 FLAGS = (PADMY,NOK,POK,pNOK,pPOK)
237 NV = 1237123231.22458
238 PV = 0x194ce70 "1237123231.22458"\0
242 DB<99> Devel::Peek::Dump $r
243 SV = PVMG(0x19e0240) at 0x142a3e8
245 FLAGS = (PADMY,POK,pPOK)
248 PV = 0x19ff900 "1237123231.22458"\0
252 DB<100> Devel::Peek::Dump $l <=> $r
253 SV = IV(0x19ea6e8) at 0x19ea6f0
255 FLAGS = (PADTMP,IOK,pIOK)
258 DB<101> Devel::Peek::Dump $l
259 SV = PVMG(0x19e0450) at 0x142a550
261 FLAGS = (PADMY,NOK,POK,pNOK,pPOK)
263 NV = 1237123231.22458
264 PV = 0x194ce70 "1237123231.22458"\0
268 DB<102> Devel::Peek::Dump $r
269 SV = PVMG(0x19e0240) at 0x142a3e8
271 FLAGS = (PADMY,NOK,POK,pIOK,pNOK,pPOK)
273 NV = 1237123231.22458
274 PV = 0x19ff900 "1237123231.22458"\0
278 Retry with uselongdouble gives same effect. Not reproducable on 32bit box (k75).
280 * Todo: reset "done" or "covered" and "minmax" after a dirty operation?
282 2009-03-11 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
284 * $obj->merge ($other) needs to learn about equal epoch which may happen
285 since dirty_epoch intruded.
287 * Wontfix anytime soon: I think we currently do not support mkdir. Only
290 2009-01-01 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
292 * Todo: continue working on update(...,$dirty_epoch). It must be
293 followed by a fast_aggregate!
295 2008-12-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
297 * maybe we need a closest_entry or fitting_interval or something like
298 that. We want to merge an event into the middle of some recentfile.
299 First we do not know which file, then we do not know where to lock,
300 where to enter the new item, when and where to correct the dirtymark.
302 So my thought is we should first find which file.
304 Another part of my brain answers: what would happen if we would enter
305 the new file into the smallest file just like an ordinary new event,
306 just as an old event?
308 (1) we would write a duplicate timestamp? No, this would be easy to
311 (2) we would make the file large quickly? Yes, but so what? We are
312 changing the dirtymark, so are willing to disturb the downstream hosts.
314 2008-11-22 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
316 * 10705 root 17 0 725m 710m 1712 S 0.0 46.8 834:56.05 /home/src/perl/repoperls/installed-perls/perl/pVNtS9N/perl-5.8.0@32642/bin/perl -Ilib /home/k/sources/CPAN/GIT/trunk/bin/testing-rmirror.pl
320 https://rt.cpan.org/Ticket/Display.html?id=41199
322 * bzcat uploads.csv.bz2 | perl -F, -nale '$Seen{$F[-1]}++ and print'
324 Strangest output being HAKANARDO who managed to upload
326 Here is a better oneliner that includes also the first line of each
329 bzcat uploads.csv.bz2 | perl -MYAML::Syck -F, -nale '$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)}'
331 2008-10-31 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
333 * memory leak in the syncher? It currently weighs 100M.
337 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
340 2008-10-29 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
342 * lookup by epoch and by path and use this ability on the pause to never
343 again register a file twice that doesn't need it. Let's call it
346 * after the dirtymark is done: fill up recentfiles with fake (historic)
347 entries; fill up with individual corrections; algorithm maybe to be done
348 with bigfloat so that we can always place something in the middle
349 between two entries. Before we must switch to bigfloat we could try to
350 use Data::Float::nextup to get the.
352 * Inotify2 on an arbitrary tree and then play with that instead of PAUSE
355 * dirtymark now lives in Recentfile, needs to be used in rmirror.
357 * find out why the downloader died after a couple of hours without a net
358 connection. Write a test that survives the not-existence of the other
361 2008-10-15 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
363 * reconsider the HTTP epoch only. Not the whole thing over HTTP because
364 it makes less sense with tight coupling for secondary files. But asking
365 the server what the current epoch is might be cheaper on HTTP than on
366 rsync. (Needs to be evaluated)
368 * remove the 0.00 from the verbose overview in the Merged column in the
371 * write tests that expose the problems of the last few days: cascading
372 client/server roles, tight coupling for secondary RFs, deletes after
375 * Some day we might want to have policy options for the slave:
376 tight/loose/no coupling with upstream for secondary RFs. tight is what
377 we have now. loose would wait until a gap occurs that can be closed.
379 2008-10-14 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
381 * revisit all $rfs->[$i+1] places if they now make sense still
383 2008-10-11 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
385 * another bug is the fact that the mirror command deletes files before
386 it unhides the index file, thus confusing downstream slaves. We must not
387 delete before unhiding and must delete after unhiding. FIXED.
389 * new complication about the slave that is playing a server role.
390 Currently we mirror from newest to oldest with a hidden temporary file
391 as index. And when one file is finished, we unhide the index file.
392 Imagine the cascading server/slave is dead for a day. It then starts
393 mirroring again with the freshest thing and unhides the freshest index
394 file when it has worked through it. In that moment it exposes a time
395 hole. Because it now works on the second recentfile which is still
398 We currently do nothing special to converge after such a drop out. At
399 least not intentionally and robustly and thought through.
401 The algorithm we use to seed the next file needs quite a lot of more
402 robustness than it currently has. Something to do with looking at the
403 merged element of the next rf and when it has dropped off, we seed
404 immediately. And if it ramains dropped off, we seed again, of course.
406 Nope, looking from smaller to larger RFS we look at the merged element
407 of this RF and at the minmax/max element of the next RF. If that
408 $rf[next]->{minmax}{max} >= $rf[this]->{merged}{epoch}, then we can stop
411 And we need a public accessor seed and unseed or seeded. But not the mix
412 of public and private stuff that then is used behind the back.
414 And then the secondary* stuff must go.
416 And we must understand what the impact is on the DONE system. Can it go
417 unnoticed that there was a hole? And could the DONE system have decided
418 the hole is covered? This should be testable with three directories where
419 the middle stops working for a while. Done->merge is suspicious, we must
420 stop it from merging non-conflatable neighbors due to broken continuity.
424 2008-10-10 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
426 * Slaven suggests to have the current epoch or the whole current
427 recentfile available from the HTTP server and take it away with
428 keepalive. This direction goes the granularity down to subseconds.
430 We might want to rewrite everything to factor out transport and allow
431 the whole thing to run via HTTP.
433 2008-10-09 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
435 * are we sure we do NOT LEAVE DOT FILES around? Especially on the
438 * smoker on k81 fetching from k75 to verify cascading works. See
439 2008-07-17 in upgradexxx and rsync-over-recentfile-3.pl.
441 * maybe the loop should wait for CHECKSUMS file after every upload. And
442 CPAN.pm needs to deal with timestamps in the future.
444 * do not forget the dirtymark!
446 Text: have a new flag on recentfiles with the meaning: if this
447 changes, you're required to run a full rsync over all the files. The
448 reason why we set it would probably be: some foul happened. we injected
449 files in arbitrary places or didn't inject them although they changed.
450 The content of the flag? Timestamp? The relation between the
451 recentfiles would have to be inheritance from the principal, because any
452 out of band changes would soon later propagate to the next recentfile.
454 By upping the flag often one can easily ruin the slaves.
456 last out of band change? dirtymark?
458 Anyway, this implies that we read a potentially existing recentfile
461 And it implies that we have an eventloop that keeps us busy in 2-3
462 cycles, one for current stuff (tight loop) and one for the recentfiles
463 (cascade when principal has changed), one for the old stuff after a
466 And it implies that the out-of-band change in any of the recentfiles
467 must have a lock on the principal file and there is the place to set the
470 * start a FAQ, especially quick start guide questions. Also to aid those
471 problematic areas where we have no good solution, like the "links"
474 * wish feedback when we are slow.
478 * Remove a few DEBUG statements.
480 * The multiple-rrr way of doing things needs a new option to rmirror,
481 like piecemeal or so. Not urgent because after the first pass through,
482 things run smoothely. It's only ugly during the first pass.
484 * I have the suspicion that the code is broken that decides if the
485 neighboring RF needs to be seeded. I fear when too much time has gone
486 between two calls (in our case more than one hour), it would not seed
487 the neighbor. Of course this will never be noticed, so we need a good
490 * local/localroot confusion: I currently pass both options but one must
493 * accounts for early birds on PAUSE rsync daemon.
495 * hardcoded 20 seconds
497 * who mirrors the index? DOING now.
499 * which CPAN mirrors offer rsync?
501 * visit all XXX, visit all _float places
503 * rename the pathdb stuff, it's too confusing. No idea how.
505 * rrr-inotify, backpan, rrr-register
507 2008-10-08 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
509 * current bugs: the pathdb seems to get no reset, the seeding of the
510 secondaryttl stuff seems not to have an effect. Have helped myself with
511 a rand(10), need to fix this back. So not checked in. Does the rand
514 The rand thing helps. The secondaryttl stuff was in the wrong line,
517 The pathdb stuff was because I called either _pathdb or __pathdb on the
518 wrong object. FIXED now.
520 * It's not so beautiful if we never fetch the recentfiles that are not
521 the principal, even if this is correct behaviour. We really do not need
522 them after we have fetched the whole content.
524 OK, we want a switch for that: secondaryttl DONE
526 2008-10-07 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
528 * bug: rrr-news --max does not count correctly. with "35" it shows me 35
529 lines but with 36 it shows 110. First it repeats 35, gives 70, and then
530 it lets 40 follow. FIXED
532 * See that the long running process really only updates the principal
533 file unless it has missed a timespan during which something happened. If
534 nothing happened, it must notice even when it misses the timespan. DONE
536 * we must throw away the pathdb when we have reached the end of Z. From
537 that moment we can have a very small pathdb because the only reason for
538 a pathdb is that we know to ignore old records in old files. We won't
539 need this pathdb again before the next full pass over the data is
540 necessary and then we will rebuild it as we go along. DONE
542 2008-10-06 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
544 * I think, Done::register_one is doing wrong in that it does not
545 conflate neighboring pieces. The covered() method cannot do this because
546 it has no recent_events array at hand. But register_one has it and could
547 do it and for some reason misses to do it (sometimes).
549 This means that the three tests I just wrote can probably not survive
550 because they test with an already broken Done structure.
552 The art now is to detect how it happens, then to reproduce, then write a
555 So from the logfile this is what happens: we have a good interval with
556 newest file being F1 at T1. Now remotely F1 gets a change and F2 goes on
557 top of it. Locally we now mirror F2 and open a new done interval for it.
558 Then we mirror F1 but this time with the timestamp T1b. And when we then
559 try to close the gap, we do not find T1 but instead something older. We
560 should gladly accept this older piece and this would fix this bug.
564 * bug to fix: when the 1h file changes while rmirror is running, we do
565 correctly sync the new files but never switch to the 6h file but rather
566 stay in a rather quick loop that fetches the 1h file again and again.
568 Is it possible that we initialize a new object? Or does
569 get_remote_recentfile_as_tempfile overwrite something in myself?
571 Want a new option: _runstatusfile => $file which frequently dumps the
572 state of all recentfiles to a file.
576 2008-10-04 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
578 * Todo: now teach update to verify the timestamp is about to write
579 against the previous and use _increase_a_bit if it doesn't comply with
580 strict monotony. DONE
582 * The problem of rounding. So far perl's default precision was
583 sufficient. One day it won't be. FakeFloat has an easy job when it is
584 only reading and other machines have written correctly. But when we want
585 to write a floating point number that is a bit larger than the other
586 one, then we need our own idea of precision.
588 Slaven said: just append a "1". This might be going towards the end of
589 usability too quickly. I'd like something that actually uses the decimal
590 system. Well, appending a 1 also does this but...
592 E.g. we have 1.0. nextup on this architecture is starting with
593 1.0000000000000004. So there is a gap to fill: 1,2,3. Now I have
594 taken the 1.0000000000000003 and the next user comes and the time tells
595 him 1.0 again. He has to beat my number without stepping over the
596 nextup. This is much less space than I had when I chose 1,2,3.
598 What is also irritating is that nextup is architecture dependent. The
599 128 bit guy must choose very long numbers to fit in between whereas the
600 other one with 16 bit uses larger steps. But then the algorithm is the
601 same for both, so that would be a nice thing.
603 I see two situation where we need this. One is when Time::HiRes returns
604 us a value that is <= the last entry in our recentfile. In this case
605 (let's call it the end-case) we must fill the region between that number
606 and the next higher native floating point number. The other is when we
607 inject an old file into an old recentfile (we would then also set a new
608 dirtymark). We find the integer value already taken and need a slightly
609 different one (let's call it the middle-case). The difference between
610 the two situations is that the next user will want to find something
611 higher than my number in the end-case and something lower than my number
614 So I suggest we give the function both a value and an upper bound and it
615 calculates us a primitive middle. The upper bound in the middle-case is
616 the next integer. The upper bound on the end-case is the nextup floating
617 point number. But the latter poses another problem: if we have occupied
618 the middle m between x and nextup(x), then the nextup(m) will probably
619 not be the same as nextup(x) because some rounding will take place
620 before the nextup is calculated and when the rounding reaches the
621 nextup(x), we will end up at nextup(nextup(x)).
623 So we really need to consider the nextup and the nextdown from there and
624 then the middle and that's the number we may approach asymptotically.
627 2008-10-03 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
629 * consider deprecating the use of RECENT.recent as a symlink. It turns
630 out to need extra hoops with the rsync options and just isn't worth it.
631 Or maybe these extra hoops are needed anyway for the rest of the tree?
632 Nope, can't be the case because not all filesystems support symlinks.
634 But before doing the large step, I'll deprecate the call of
635 get_remote_recentfile_as_tempfile with an argument. Rememberr this was
636 only introduced to resolve RECENT.recent and complicates the routine far
637 beyond what it deserves.
639 DONE. Won't deprecate RECENT.recent, just moved its handling to the
642 2008-10-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
644 * I think it's a bug that the rsync_option links must be set to true in
645 order to support RECENT.recent and that nobody cares to set it
646 automatically. Similar for ignore_link_stat_errors.
648 2008-09-27 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
650 * Todo: find all todos together and make a plan what is missing for a
653 - verifytree or something like that. fsck maybe.
655 - rersyncrecent, the script itself? What it do?
657 - a way to only mirror the recentfiles without mirroring the whole
658 remote system such that people can decide to mirror only partially see
659 also 2008-08-30. .shadow-xxx directory? this also needed for a
660 filesystem that is still incomplete and might need the mirrorfiles for
663 - long living objects that mirror again and again. Inject something
664 into ta, see how it goes over to tb.
666 - how do we continue filling up the DONE system when we use an object
667 for the second time? "fully covered" and "uptodate" or new terminology.
669 - overview called on the wrong file should be understandable
671 - the meta data field that must change when we fake something up so that
672 the downstream people know they have to re-fetch everything.
674 - how tolerant are we against missing files upstream? how do we keep
675 track? there are legitimate cases where we did read upstream index right
676 before a file got deleted there and then find that file as new and want
677 it. There are other cases that are not self healing and must be tracked
680 - how, exactly, do we have to deal with deletes? With rsync errors?
682 rsync: link_stat "/id/K/KA/KARMAN/Rose-HTMLx-Form-Related-0.07.meta" (in
683 authors) failed: No such file or directory (2)
685 The file above is a delete in 1h and a new in file 1M and the
686 delete in the locally running rmirror did not get propagated to the 1M
687 object. Bug. And the consequence is a standstill.
689 It seems that a slave that works with a file below the principal needs
690 to merge things all the way up to get rid of later deletes. Or keep
691 track of all deletes and skip them later. So we need a trackdeletes.pm
692 similar to the done.pm?
694 see also 2008-08-20 about spurious deletes that really have no add
695 counterpart and yet they are not wrong.
697 - consider the effect when resyncing the recentfile takes longer than
698 the time per loop. Then we never rsync any file. We need to diagnose
699 that and force an increase of that loop time. But when we later are fast
700 enough again because the net has recovered, then we need to switch back
701 to original parameters. ERm, no, it's enough to keep syncing at least
702 one file before refetching an index file.
704 - remember to verify that no temp files are left lying around and the
707 - status file for not long running jobs that want to track upstream with
710 - revisit all XXX _float areas and study Sub::Exporter DONE
712 - persistent DB even though we just said we do not need it. Just for
713 extended capabilities and time savings when, for example, upstream
714 announces a reset and we get new recentfiles and could then limit
715 ourselves to a subset of files (those that have a changed epoch) in a
716 first pass and would only then do the loop to verify the rest. Or
719 * Todo: aggregate files should know their feed and finding the principal
720 should be done stepwise. (?)
722 * Todo: DESTROY thing that unlocks. Today when I left the debuggerr I
723 left locks around. DONE
725 2008-09-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
727 * maybe extend the _overview so that it always says if and where the
728 last file is in the next file and where the next event in the next rf
729 would lie. No, don't like this anymore. REJECT
731 * take the two new redundant tests out again, only the third must
734 * Todo: add a sanity check if the merged structure is really pointing to
735 a different rf and that this different rf is larger. DONE
737 2008-09-25 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
739 * now test, if they are overlapping. And test if there is a file in the
740 next rf that would fit into this rf's interval.
742 1h 1222324012.8474 1222322541.7963 0.4086
743 6h 1222320411.2760 1222304207.6931 4.5010 missing overlap/gap!
744 1d 1222320411.2760 1222238750.5071 22.6835 large overlap
745 1W 1222313218.3626 1221708477.5829 167.9835
747 I suspect that somebody writes a merged timestamp without having merged
748 and then somebody else relies on it.
750 If aggregate is running, the intervals must not be extravagated, if it
751 is not running, there must not be bounds, the total number of events in
752 the system must be counted and must be controlled throughout the tests.
753 That the test required the additional update was probably nonsense,
754 because aggregate can cut pieces too. FIXED & DONE
756 2008-09-23 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
758 * rrr-aggregate seems to rewrite the RECENT file even if nothing has
761 2008-09-21 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
763 * Most apparent bug at the moment is that the recentfiles are fetched
764 too often. Only the principal should be fetched and if it has not
765 changed, the others should not be refetched. ATM I must admit that I'm
766 happy that we refetch more often than needed because I can more easily
767 fix bugs while the thing is running.
769 * Let's say, 1220474966.19501 is a timestamp of a file that is already
770 done but the done system does not know about it. The reason for the
771 failure is not known and we never reach the status uptodate because of
772 this. We must get over it.
774 Later it turns out that the origin server had a bug somewhere.
775 1220474966.19042 came after 1220474966.19501. Or better: it was in the
776 array of the recentfile one position above. The bug was my own.
778 2008-09-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
780 * There is the race condition where the server does a delete and the
781 slave does not yet know and then tries to download it because he sees
782 the new. So for this time window we must be more tolerant against
783 failure. If we cannot download a file, we should just skip it and should
784 not retry immediately. The whole system should discover the lost thing
785 later. Keeping track with the DONE system should really be a no brainer.
787 But there is something more: the whole filesystem is a database and the
788 recentfiles are one possible representation of it. It's a pretty useful
789 representation I think that's why I have implemented something around
790 it. But for strictly local operation it has little value. For local
791 operation we would much rather have a database. So we would enter every
792 recentfile reading and every rsync operation and for every file the last
793 state change and what it leads to. Then we would always ignore older
794 records without the efforts involved with recentfiles.
796 The database would have: path,recentepoch,rsyncedon,deletedon
798 Oh well, not yet clear where this leads to.
800 2008-09-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
802 * Bug: the bigloop ran into a funny endless loop after EWILHELM uploaded
803 Module-Build. It *only* rsynced the "1h" recentfile from that moment on.
805 * statusfile, maybe only on demand, alone to have a sharp debugging
806 tool. It is locked and all recentfiles dump themselves into it and we
807 can build a viewer that lets us know where we stand and what's inside.
809 * remember: only the principal recentfile needs expiration, all others
810 shall be expired by principal if it discovers that something has move
813 2008-09-18 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
815 * Always check if we stringify to a higher value than in the entry
818 * And in covered make an additional check if we would be able to see a
819 numerical difference between the two numbers and if we can't then switch
820 to a different, more expensive algorithm. Do not want to be caught by
821 floating surprises. DONE
823 2008-09-17 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
825 * caching has several aspects here: we can cache the interval of the
826 recentfile which only will change when the mtime of the file changes. We
827 must re-mirror the recentfile when its ttl has expired. Does have_read
828 tell you anything? It counts nothing at all. Only the mtime is
829 interesting. The ntuple mtime, low-epoch, high-epoch. And as a separate
830 thing the have_mirrored because it is unrelated to the mtime.
832 * Robustness of floating point calculations! I always thought that the
833 string calculated by the origin server for the floating representation
834 of the epoch time is just a string. When we convert it to a number and
835 later back to a string, the other computer might come to a different
836 conclusion. This must not happen, we want to preserve it under any
837 circumstances. I will have to write tests with overlong sequences that
838 get lost in arithmetic and must see if all still works well. DONE
840 But one fragile point remains: if one host considers a>b and the other
841 one considers them == but no eq. To prevent this, we must probably do
842 some extra homework. DONE
844 2008-09-16 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
846 * the concept of tracking DONE needs an object per recentfile that has
847 something like these methods:
849 do_we_have(xxx), we_have(xxx), do_we_have_all(xxx,yyy), reset()
851 covered() register() covered()
853 The unclear thing is how we translate points in time into intervals. We
854 could pass a reference to the current recent_events array when running
855 we_have(xxx) and let the DONE object iterate over it such that it only
856 has to store a list of intervals that can melt into each other. Ah, even
857 passing the list together with a list of indexes seems feasiable.
859 Or maybe ask for the inverted list?
861 Whenever the complete array is covered by the interval we say we are
862 fully covered and if the recentfile is not expired, we are uptodate.
864 2008-09-07 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
866 2008-09-05 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
868 * need a way to "return" the next entry after the end of a list. When
869 the caller says "before" or "after" we would like to know if he could
870 cover that interval/threshold or not because this influences the effect
871 of a newer timestamp of that recentfile. DONE with $opt{info}.
873 2008-09-04 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
875 * one of the next things to tackle: the equivalent of csync2 -TIXU.
877 loop implies tixu (?). Nope, something like --statefile decides. Per
880 T test, I init, X including removals, U nodirtymark
882 So we have no concept of dirtymarks, we only trust that since we are
883 running we have observed everything steadily. But people will not let
884 this program run forever so we must consider both startup penalty and
885 book keeping for later runs. We keep this for later. For now we write a
886 long running mirror that merges several intervals.
888 2008-09-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
890 * need to speed up the 02 test, it's not clever to sleep so much. Reduce
893 * rersyncrecent, the script: default to one week. The name of the switch
894 is --after. Other switches? --loop!
896 2008-08-30 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
898 * need a switch --skip-deletes (?)
900 * need a switch --enduser that tells us that the whole tempfile
901 discipline is not needed when there is no downstream user. (?)
903 Without this switch we cannot have a reasonable recent.pl that just
904 displays the recent additions. Either we accept to download everything.
905 Or we download temporary files without the typical rsync protocol
908 Or maybe the switch is --tmpdir? If --tmpdir would mean: do not use
909 File::Temp::tempdir, this might be a win.
911 2008-08-29 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
913 * apropos missing: we have no push, we never know the downstream
914 servers. People who know their downstream hosts and want to ascertain
915 something will want additional methods we have never thought about, like
916 update or delete a certain file.
918 2008-08-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
920 * tempted to refactor rmirror into resolve_symlink, localize, etc.
921 Curious if rsync_options=links equal 0 vs. 1 will make the expected
924 * rsync options: it's a bit of a pain that we usually need several rsync
925 options, like compress, links, times, checksum and that there is no
926 reasonable default except the original rsync default. I think wee can
927 safely assume that the rsync options are shared between all recentfile
928 instances within one recent tree.
930 2008-08-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
932 * deletes: if a delete follows an add quickly enough it may happen that
933 a downstream mirror did not see the add at all! It seems this needs to
934 be mentioned somewhere. The point here is that even if the downstream is
935 never missing the principal timeframe it may encounter a "delete" that
936 has no complimentary "add" anywhere.
938 2008-08-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
940 * I suspect the treat of metadata is incorrect during read or something.
941 The bug that I am watching is that between 06:08 and 06:09 the 6h file
942 contained more than 6 hours worth of data. At 06:08 we merged into the
943 1d file. We need to take snapshots of the 6h file over the course of an
944 hour or maybe only between XX:08 and XX:09? Nope, the latter is not
947 Much worse: watching the 1h file: right at the moment (at 06:35) it
948 covers 1218867584-1219120397 which is 70 hours.
950 Something terribly broken. BTW, 1218867584 corresponds to Sat Aug 16
951 08:19:44 2008, that is when I checked out last time, so it seems to be
952 aggregating and never truncating?
954 No, correct is: it is never truncating; but wrong is: it is aggregating.
955 It does receive a lot of events from time to time from a larger file.
956 Somehow a large file gets merged into the small one and because the
957 "meta/merged" attribute is missing, nobody is paying attention. I
958 believe that I can fix this by making sure that metadata are honoured
959 during read. DONE and test adjusted.
961 2008-08-17 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
963 * grand renaming plan
965 remotebase => remoteroot to fit well with localroot DONE
966 local_path() => localroot seems to me should already work DONE
967 recentfile_basename => rfilename no need to stress it has no slash DONE
969 filenameroot??? Doesn't seem too bad to me today. Maybe something like
970 kern? It would anyway need a deprecation cycle because it is an
971 important constructor.
973 * I like the portability that Data::Serializer brings us but the price
974 is that some day we might find out that it is slowing us a bit. We'll
977 2008-08-16 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
979 * should we not enter the interval of the principal (or the interval of
980 the merging file?) in every aggregated/merged file?
982 * we should aim at a first release and give up on thinking about
983 sanitizing stuff and zloop. Let's just admit that a full traditional
984 rsync is the only available sanitizer ATM. Otherwise it's complicated
985 stuff: sanitizing on the origin server, sanitizing on the slaves,
986 sanitizing forgotten files, broken timestamps, etc. Let's delay it and
987 get the basics out before this becomes a major cause for mess.
989 2008-08-13 Andreas Koenig <k@andreas-koenigs-computer.local>
991 * On OSes not supporting symlinks we expect that RECENT.recent contains
992 the contents of the principal recentfile. Actually this is identical on
993 systems supporting symlinks. Simple, what follows from that is that we
994 need to keep the serializer in the metadata because we cannot read it
995 from the filename, doesn't it? Of course not. It's a chicken and egg
996 problem. This leaves us with the problem to actually parse the
997 serialized data to find out in which format it is. So who can do the 4
998 or 5 magics we wanted to support? File::LibMagic?
1000 2008-08-09 Andreas Koenig <k@andreas-koenigs-computer.local>
1002 * remotebase and recentfile_basename are ugly names. Now that we need a
1003 word for the shortest/principal/driving recentfile too we should do
1006 localroot is good. rfile is good. local_path() is bad, local_path($path)
1007 is medium, filenameroot() is bad, remotebase is bad, recentfile is
1010 Up to now remotebase was the string that described the remote root
1011 directory in rsync notation, like pause.perl.org::authors. And
1012 recentfile_basename was "RECENT-1h.yaml".
1014 2008-08-08 Andreas Koenig <k@andreas-koenigs-computer.local>
1016 * The test that was added in today's checkin is a good start for a test
1017 of rmirror. We should have more methods in Recent.pm: verify,
1018 addmissingfiles. We should verify the current tree, then rmirror it and
1019 then verifytree the copy. We could then add some arbitrary file and let
1020 it be discovered by addmissingfiles, then rmirror again and then
1021 verifytree the copy again.
1023 Then we could start stealing from csync2 sqlite database [no port to
1024 OSX!] and fill a local DB. And methods to compare the database with the
1025 recentfiles. Our strength is that in principle we could maintain state
1026 with a single float. We have synced up to 1234567890.123456. If the Z
1027 file does not add new files all we have to do is mirror the new ones and
1030 This makes it clear that we should extend current protocol and declare
1031 that we cheat when we add files too late, just to help the other end
1032 keeping track. Ah yes, that's what was meant when zloop was mentioned
1035 Maybe need to revisit File::Mirror to help me with this task.
1037 2008-08-07 Andreas Koenig <k@andreas-koenigs-computer.local>
1039 * There must be an allow-me-to-truncate flag in every recentfile.
1040 Without it one could construct a sequence of updates winning the locking
1041 battle against the aggregator. Only if an aggregator has managed to
1042 merge data over to the next level, truncating can be allowed. DONE with
1045 2008-08-06 Andreas Koenig <k@andreas-koenigs-computer.local>
1047 * We should probably guarantee that no duplicates enter the aggregator
1050 2008-08-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1052 * To get merge operation faster would need a good benchmark test. What
1053 02 spits out isn't reliable enough and is dominated by many other
1056 commit 10176bf6b79865d4fe9f46e3857a3b8669fa7961
1057 Author: Andreas J. Koenig <k@k75.(none)>
1058 Date: Sat Aug 2 07:58:04 2008 +0200
1062 commit 3243120a0c120aaddcd9b1f4db6689ff12ed2523
1063 Author: Andreas J. Koenig <k@k75.(none)>
1064 Date: Sat Aug 2 11:40:29 2008 +0200
1066 there was a lot of trying but the effect is hardly measurable with
1069 * overhead of connecting seems high. When setting
1070 max_files_per_connection to 1 we see that.
1072 2008-08-01 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1074 * 1217622571.0889 - 1217597432.86734 = 25138.2215600014
1076 25138.2215600014/3600 = 6.98283932222261
1078 It jumps into the eye that this is ~ 7 hours, not ~6, so there seems to
1079 be a bug in the aggregator. FIXED
1081 2008-07-27 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1083 * e.g. id/Y/YE/YEWENBIN/Emacs-PDE-0.2.16.tar.gz: Do we have it, should
1084 we have it, can we mirror it, mirror it!
1086 I fear this needs a new class which might be called
1087 File::Rsync::Mirror::Recent. It would collect all recentfiles of a kind
1088 and treat them as an entity. I realize that a single recentfile may be
1089 sufficient for certain tasks and that it is handy for the low level
1090 programmer but it is not nice to use. If there is a delete in the 1h
1091 file then the 6h file still contains it. Seekers of the best information
1092 need to combine at least some of the recentfiles most of the time.
1094 There is the place for the Z loop!
1096 But the combination is something to collect in a database, isn't it. Did
1097 csync2 just harrumph?
1099 2008-07-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1101 * it just occurred to me that hosts in the same mirroring pool could
1102 help out each other even without rewriting the recentfile. Just fetch
1103 the stuff to mirror from several places, bingo. But that's something
1104 that should rather live in a separate package or in rsync directly.
1106 * cronjobs are unsuited because with ntp they would all come at the full
1107 minute and disturb each other. Besides that I'd hate to have a backbone
1108 with more than a few seconds latency.
1110 2008-07-25 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1112 * a second rsync server with access control for PAUSE. Port? 873 is the
1113 standard port, let's take 8873.
1115 * if there were a filesystem based on this, it would have a slow access
1116 to inexistent files. It would probably provide wrong readdir (only based
1117 on current content) or also a slow one (based on a recentfile written
1118 after the call). But it would provide fast access to existing files. Or
1119 one would deliberately allow slightly blurred answers based on some
1120 sqlite reflection of the recentfiles.
1122 * todo: write a variant of mirror() that combines two or more
1123 recentfiles and treats them like one
1125 * todo: signal handler to remove the tempfile
1127 2008-07-24 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1129 * now that we have the symlink I forgot how it should be used in
1132 * the z loop: add missing files to Z file. Just append them (instead of
1133 prepending). So one guy prepends something from the Y file from time to
1134 time and another guy appends something rather frequently. Collecting
1135 pond. When Y merges into Z, things get epoch and the collecting pond
1136 gets smaller. What exactly are "missing files"?
1138 take note of current epoch of the alpha file, let's call it the
1141 find all files on disk
1143 remove all files registered in the recentworld up to recent-ts
1145 remove all files that have been deleted after recent-ts according to
1148 2008-07-23 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1150 * rersyncrecent might be a cronjob with a (locked) state file which
1151 contains things like after and maybe last z sync or such?
1153 rrr-mirror might be an alternative name but how would we justify the
1154 three Rs when there is no Re-Rsync-Recent?
1156 With the --loop parameter it is an endless loop, without it is no loop.
1157 At least this is simple.
1159 * todo: new accssor z-interval specifies how often the Z file is updated
1160 against the filesystem. We probably want no epoch stamp on these
1161 entries. And we want to be able to filter the entries (e.g. no
1162 by-modules and by-category tree)
1164 2008-07-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1166 * Fill the Z file. gc or fsck or both. Somehow we must get the old files
1167 into Z. We do not need the other files filled up with filesystem
1170 * need interface to query for a file in order to NOT call update on
1171 PAUSE a second time within a short time.
1173 2008-07-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1175 * recommended update interval? Makes no sense, is different for
1182 change-log-default-name: "Todo"