1 2009-04-11 Andreas J. Koenig <andk@cpan.org>
3 * Bug: when the dirtymark changes upstream, then the downstream server
4 notices it immediately and mirrors the "1h" file but then leaves the
5 rmirror loop and continues with "6h" on the next round through. This
6 behaviour goes on until it reaches the Z file. Because we eat the Z file
7 piecemeal we leave the loop after a while but after that everybody seems
8 to have forgotten that there is some work left to be done. Bad, bad,
11 ETOOCONFUSING better name?
13 Recentfile::get_remote_recentfile_as_tempfile OK
14 Recentfile::resolve_recentfilename split_rfilename DONE
16 Recent::_principal_recentfile_object _principal_recentfile_fromremote
17 Recent::principal_recentfile OK
18 Recent::_resolve_rfilename _principal_recentfile_fromremote_resosymlink
20 some void, some not, some not void but called in void context.
22 2009-04-10 Andreas J. Koenig <andk@cpan.org>
24 * Todo: rename rrr-dirtyupdate to rrr-update and add an option --dirty
25 so we can use it for normal update with a single file.
27 * Bug? When the dirtymark changes then the second tier hosts quickly
28 reset their DONE state and restart mirroring. But then they mirror
29 potentially outdated index files with inconsistent dirtymark because the
30 first tier box may be a bit behind on the large recentfiles.
32 This means we need a brake when iterating over the recentfiles that
33 refuses to use a recentfile with the wrong dirtymark.
35 2009-04-08 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
37 * install the bin/* files and talk about rrr-overview somewhere.
39 * Ask B. Hansen asks for tempdir accessor such that tempfiles get
40 created outside the target tree.
42 * Barbie asks for better logging capabilities
44 2009-03-29 Andreas J. Koenig <andk@cpan.org>
46 * known bug: cannot configure the Recentfile objects to keep delete
47 events. How to call the accessor for that? keep_delete_objects_forever?
49 Sounds acceptable. FIXED (but untested).
51 * known bug: keeps temporary index files lying around.
53 2009-03-24 Andreas J. Koenig <andk@cpan.org>
57 for ( @{$rrr->recentfiles} ) { $_->verbose(0) }
61 2009-03-22 Andreas J. Koenig <andk@cpan.org>
63 * k81 is again client of k75 with different parameters.
65 * broken now: (1) seeding and unseeding: rmirror is seeding and talking
66 about it all the time and nobody reacts accordingly; (2) lots of temp
67 files get created and not removed; culprit the new call to
68 get_remote_recentfile_as_tempfile within Recent.pm; the manpage says the
69 caller has to remove the tempfile after use.
71 Need the drawing board.
73 PARTLY FIXED: retracting the idea to call
74 get_remote_recentfile_as_tempfile from Recent.pm and moving around seed
75 and unseed such that uptodate now means "we have mirrored this
76 recentfile's bunch of files" and seeded means "somebody believes the
77 index file for this rf needs be refreshed". This seems to work now.
79 Still collecting temp files that nobody cares to depollute.
81 2009-03-21 Andreas J. Koenig <andk@cpan.org>
83 * Bug?: should it be harder than it is atm to set the timestamp to the
86 * bug with native integers:
90 path: id/D/DE/DELTA/Crypt-Rijndael_PP-0.03.readme
94 path: id/L/LG/LGODDARD/Tk-Wizard-2.124.readme
97 Native integer broke when native math was turned off. FIXED
99 * Bug: something between id/P/PH/PHISH/CGI-XMLApplication-1.1.2.readme
100 and id/C/CH/CHOGAN/HTML-WWWTheme-1.06.readme. Mirroring the Z file loops
103 Yes, records out of order:
106 447004 epoch: 995885533
107 447005 path: id/P/PH/PHISH/CGI-XMLApplication_0.9.3.readme
110 447008 epoch: 995890358
111 447009 path: id/H/HD/HDIAS/Mail-Cclient-1.3.readme
114 447012 epoch: 995892221
115 447013 path: id/H/HD/HDIAS/Mail-Cclient-1.3.tar.gz
118 FIXED with sanity check and later with the integer fix.
120 * Bug: want the index files in a .recent directory
122 * Bug: lots of dot files are not deleted in time
124 * possible test case: can a delete change the timestamp? This would
125 probably break the order of events.
127 2009-03-20 Andreas J. Koenig <andk@cpan.org>
129 * 1233701831.34486 what's so special about this number/string? It
131 id/G/GR/GRODITI/MooseX-Emulate-Class-Accessor-Fast-0.00800.tar.gz and
132 atm lives in Y,Q, and Z.
134 It is the first entry after id/--skip-locking which has timestamp
135 1234164228.11325 which represents a file that doesn't exist anymore.
139 Sync 1237531537 (31547/33111/Z) id/J/JH/JHI/String-Approx-2.7.tar.gz ...
140 _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
141 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatcmp(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 131
142 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatlt(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/Done.pm line 110
143 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
144 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
145 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
146 at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 76
147 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatcmp(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 131
148 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatlt(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/Done.pm line 110
149 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
150 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
151 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
153 FIXED, it was the "--skip-locking" file where manual intervention was participating
155 * bug on the mirroring slave: when the dirtymark gets increased we
156 probably do not reset the done intervals. The mirrorer stays within
157 tight bounds where it tries to sync with upstream and never seems to
158 finish. In the debugging state file I see lots of identical intervals
159 that do not get collapsed. When I restart the mirrorer it dies with:
161 Sync 1237507989 (227/33111/Z) id/X/XI/XINMING/Catalyst-Plugin-Compress.tar.gz ...
162 _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
163 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatcmp(1237400817.94363, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 101
164 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatge(1237400817.94363, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/Done.pm line 226
165 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
166 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
167 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
168 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
169 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
170 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
172 and debugging stands at
178 epoch: 1237400802.5789
179 path: id/J/JO/JOHND/CHECKSUMS
182 epoch: 1237400807.97514
183 path: id/M/MI/MIYAGAWA/CHECKSUMS
186 epoch: 1237400817.94363
187 path: id/X/XI/XINMING/Catalyst-Plugin-Compress.tar.gz
195 and it is reproducable.
197 Why does the mirrorer not fetch a newer Z file? It is 18 hours old while
198 pause has a fresh one.
200 FIXED, it was the third anded term in each of the ifs in the IV block in
201 _register_one: with that we make sure that we do not stamp on valuable
204 2009-03-17 Andreas J. Koenig <andk@cpan.org>
206 * done: verified the existence of the floating point bug in bleadperl
207 and verified that switching from YAML::Syck to YAML::XS does not resolve
210 BTW, the switch was doable with
212 perl -i~ -pe 's/Syck/XS/g' lib/**/*.pm t/*.t
214 and should be considered as a separate TODO
216 * todo: integrate a dirty update with two aggregate calls before
217 unlocking for frictionless dirtying
219 * todo: start the second rsync daemon on pause
221 * todo: move index files to .recent: this cannot simply be done by
222 setting filenameroot to .recent/RECENT. Other parts of the modules rely
223 on the fact that dirname(recentfile) is the root of the mirrored tree.
225 2009-03-16 Andreas J. Koenig <andk@cpan.org>
227 * What was the resolution of the mirror.pl delete hook bug? Do we call
228 the delete hook when pause removes a file from MUIR?
230 * Today on pause: Updating 2a13fba..29f284d and installing it for
231 /usr/local/perl-5.10.0{,-RC2}
233 TURUGINA/Set-Intersection-0.01.tar.gz was the last upload before this
234 action and G/GW/GWILLIAMS/RDF-Query-2.100_01.tar.gz the first after it
236 2009-03-15 Andreas J. Koenig <andk@cpan.org>
238 * currently recent_events has the side effect of setting dirtymark
239 because it forces a read on the file. That should be transparent, so
240 that the dirtymark call always forces a cache-able(?) read.
242 * The bug below is -- after a lot of trying -- not reproducible on a
243 small script, only in the large test script. The closest to the output
249 my $x = "01237123229.8814";
254 ($l,$r) = ($1,$2) if $x =~ /(.)(.+)/;
256 $l = "1237123231.22458";
257 $r = "1237123231.22458";
259 Devel::Peek::Dump $l;
260 Devel::Peek::Dump $r;
261 Devel::Peek::Dump $x = $l <=> $r;
265 The checked in state at c404a85 fails the test with my
266 /usr/local/perl-5.10-uld/bin/perl on 64bit but curiously not with
267 /usr/local/perl-5.10-g/bin/perl. So it seems the behaviour is not even
268 in the test script always consistent.
270 * Todo: write a test that inserts a second dirty file with an already
271 existing timestamp. DONE
273 * Bug in perl 5.10 on my 64bit box:
275 DB<98> Devel::Peek::Dump $l
276 SV = PVMG(0x19e0450) at 0x142a550
278 FLAGS = (PADMY,NOK,POK,pNOK,pPOK)
280 NV = 1237123231.22458
281 PV = 0x194ce70 "1237123231.22458"\0
285 DB<99> Devel::Peek::Dump $r
286 SV = PVMG(0x19e0240) at 0x142a3e8
288 FLAGS = (PADMY,POK,pPOK)
291 PV = 0x19ff900 "1237123231.22458"\0
295 DB<100> Devel::Peek::Dump $l <=> $r
296 SV = IV(0x19ea6e8) at 0x19ea6f0
298 FLAGS = (PADTMP,IOK,pIOK)
301 DB<101> Devel::Peek::Dump $l
302 SV = PVMG(0x19e0450) at 0x142a550
304 FLAGS = (PADMY,NOK,POK,pNOK,pPOK)
306 NV = 1237123231.22458
307 PV = 0x194ce70 "1237123231.22458"\0
311 DB<102> Devel::Peek::Dump $r
312 SV = PVMG(0x19e0240) at 0x142a3e8
314 FLAGS = (PADMY,NOK,POK,pIOK,pNOK,pPOK)
316 NV = 1237123231.22458
317 PV = 0x19ff900 "1237123231.22458"\0
321 Retry with uselongdouble gives same effect. Not reproducable on 32bit box (k75).
323 * Todo: reset "done" or "covered" and "minmax" after a dirty operation?
325 2009-03-11 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
327 * $obj->merge ($other) needs to learn about equal epoch which may happen
328 since dirty_epoch intruded.
330 * Wontfix anytime soon: I think we currently do not support mkdir. Only
333 2009-01-01 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
335 * Todo: continue working on update(...,$dirty_epoch). It must be
336 followed by a fast_aggregate!
338 2008-12-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
340 * maybe we need a closest_entry or fitting_interval or something like
341 that. We want to merge an event into the middle of some recentfile.
342 First we do not know which file, then we do not know where to lock,
343 where to enter the new item, when and where to correct the dirtymark.
345 So my thought is we should first find which file.
347 Another part of my brain answers: what would happen if we would enter
348 the new file into the smallest file just like an ordinary new event,
349 just as an old event?
351 (1) we would write a duplicate timestamp? No, this would be easy to
354 (2) we would make the file large quickly? Yes, but so what? We are
355 changing the dirtymark, so are willing to disturb the downstream hosts.
357 2008-11-22 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
359 * 10705 root 17 0 725m 710m 1712 S 0.0 46.8 834:56.05 /home/src/perl/repoperls/installed-perls/perl/pVNtS9N/perl-5.8.0@32642/bin/perl -Ilib /home/k/sources/CPAN/GIT/trunk/bin/testing-rmirror.pl
363 https://rt.cpan.org/Ticket/Display.html?id=41199
365 * bzcat uploads.csv.bz2 | perl -F, -nale '$Seen{$F[-1]}++ and print'
367 Strangest output being HAKANARDO who managed to upload
369 Here is a better oneliner that includes also the first line of each
372 bzcat uploads.csv.bz2 | perl -MYAML::Syck -F, -nale '$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)}'
374 2008-10-31 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
376 * memory leak in the syncher? It currently weighs 100M.
380 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
383 2008-10-29 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
385 * lookup by epoch and by path and use this ability on the pause to never
386 again register a file twice that doesn't need it. Let's call it
389 * after the dirtymark is done: fill up recentfiles with fake (historic)
390 entries; fill up with individual corrections; algorithm maybe to be done
391 with bigfloat so that we can always place something in the middle
392 between two entries. Before we must switch to bigfloat we could try to
393 use Data::Float::nextup to get the.
395 * Inotify2 on an arbitrary tree and then play with that instead of PAUSE
398 * dirtymark now lives in Recentfile, needs to be used in rmirror.
400 * find out why the downloader died after a couple of hours without a net
401 connection. Write a test that survives the not-existence of the other
404 2008-10-15 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
406 * reconsider the HTTP epoch only. Not the whole thing over HTTP because
407 it makes less sense with tight coupling for secondary files. But asking
408 the server what the current epoch is might be cheaper on HTTP than on
409 rsync. (Needs to be evaluated)
411 * remove the 0.00 from the verbose overview in the Merged column in the
414 * write tests that expose the problems of the last few days: cascading
415 client/server roles, tight coupling for secondary RFs, deletes after
418 * Some day we might want to have policy options for the slave:
419 tight/loose/no coupling with upstream for secondary RFs. tight is what
420 we have now. loose would wait until a gap occurs that can be closed.
422 2008-10-14 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
424 * revisit all $rfs->[$i+1] places if they now make sense still
426 2008-10-11 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
428 * another bug is the fact that the mirror command deletes files before
429 it unhides the index file, thus confusing downstream slaves. We must not
430 delete before unhiding and must delete after unhiding. FIXED.
432 * new complication about the slave that is playing a server role.
433 Currently we mirror from newest to oldest with a hidden temporary file
434 as index. And when one file is finished, we unhide the index file.
435 Imagine the cascading server/slave is dead for a day. It then starts
436 mirroring again with the freshest thing and unhides the freshest index
437 file when it has worked through it. In that moment it exposes a time
438 hole. Because it now works on the second recentfile which is still
441 We currently do nothing special to converge after such a drop out. At
442 least not intentionally and robustly and thought through.
444 The algorithm we use to seed the next file needs quite a lot of more
445 robustness than it currently has. Something to do with looking at the
446 merged element of the next rf and when it has dropped off, we seed
447 immediately. And if it ramains dropped off, we seed again, of course.
449 Nope, looking from smaller to larger RFS we look at the merged element
450 of this RF and at the minmax/max element of the next RF. If that
451 $rf[next]->{minmax}{max} >= $rf[this]->{merged}{epoch}, then we can stop
454 And we need a public accessor seed and unseed or seeded. But not the mix
455 of public and private stuff that then is used behind the back.
457 And then the secondary* stuff must go.
459 And we must understand what the impact is on the DONE system. Can it go
460 unnoticed that there was a hole? And could the DONE system have decided
461 the hole is covered? This should be testable with three directories where
462 the middle stops working for a while. Done->merge is suspicious, we must
463 stop it from merging non-conflatable neighbors due to broken continuity.
467 2008-10-10 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
469 * Slaven suggests to have the current epoch or the whole current
470 recentfile available from the HTTP server and take it away with
471 keepalive. This direction goes the granularity down to subseconds.
473 We might want to rewrite everything to factor out transport and allow
474 the whole thing to run via HTTP.
476 2008-10-09 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
478 * are we sure we do NOT LEAVE DOT FILES around? Especially on the
481 * smoker on k81 fetching from k75 to verify cascading works. See
482 2008-07-17 in upgradexxx and rsync-over-recentfile-3.pl.
484 * maybe the loop should wait for CHECKSUMS file after every upload. And
485 CPAN.pm needs to deal with timestamps in the future.
487 * do not forget the dirtymark!
489 Text: have a new flag on recentfiles with the meaning: if this
490 changes, you're required to run a full rsync over all the files. The
491 reason why we set it would probably be: some foul happened. we injected
492 files in arbitrary places or didn't inject them although they changed.
493 The content of the flag? Timestamp? The relation between the
494 recentfiles would have to be inheritance from the principal, because any
495 out of band changes would soon later propagate to the next recentfile.
497 By upping the flag often one can easily ruin the slaves.
499 last out of band change? dirtymark?
501 Anyway, this implies that we read a potentially existing recentfile
504 And it implies that we have an eventloop that keeps us busy in 2-3
505 cycles, one for current stuff (tight loop) and one for the recentfiles
506 (cascade when principal has changed), one for the old stuff after a
509 And it implies that the out-of-band change in any of the recentfiles
510 must have a lock on the principal file and there is the place to set the
513 * start a FAQ, especially quick start guide questions. Also to aid those
514 problematic areas where we have no good solution, like the "links"
517 * wish feedback when we are slow.
521 * Remove a few DEBUG statements.
523 * The multiple-rrr way of doing things needs a new option to rmirror,
524 like piecemeal or so. Not urgent because after the first pass through,
525 things run smoothely. It's only ugly during the first pass.
527 * I have the suspicion that the code is broken that decides if the
528 neighboring RF needs to be seeded. I fear when too much time has gone
529 between two calls (in our case more than one hour), it would not seed
530 the neighbor. Of course this will never be noticed, so we need a good
533 * local/localroot confusion: I currently pass both options but one must
536 * accounts for early birds on PAUSE rsync daemon.
538 * hardcoded 20 seconds
540 * who mirrors the index? DOING now.
542 * which CPAN mirrors offer rsync?
544 * visit all XXX, visit all _float places
546 * rename the pathdb stuff, it's too confusing. No idea how.
548 * rrr-inotify, backpan, rrr-register
550 2008-10-08 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
552 * current bugs: the pathdb seems to get no reset, the seeding of the
553 secondaryttl stuff seems not to have an effect. Have helped myself with
554 a rand(10), need to fix this back. So not checked in. Does the rand
557 The rand thing helps. The secondaryttl stuff was in the wrong line,
560 The pathdb stuff was because I called either _pathdb or __pathdb on the
561 wrong object. FIXED now.
563 * It's not so beautiful if we never fetch the recentfiles that are not
564 the principal, even if this is correct behaviour. We really do not need
565 them after we have fetched the whole content.
567 OK, we want a switch for that: secondaryttl DONE
569 2008-10-07 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
571 * bug: rrr-news --max does not count correctly. with "35" it shows me 35
572 lines but with 36 it shows 110. First it repeats 35, gives 70, and then
573 it lets 40 follow. FIXED
575 * See that the long running process really only updates the principal
576 file unless it has missed a timespan during which something happened. If
577 nothing happened, it must notice even when it misses the timespan. DONE
579 * we must throw away the pathdb when we have reached the end of Z. From
580 that moment we can have a very small pathdb because the only reason for
581 a pathdb is that we know to ignore old records in old files. We won't
582 need this pathdb again before the next full pass over the data is
583 necessary and then we will rebuild it as we go along. DONE
585 2008-10-06 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
587 * I think, Done::register_one is doing wrong in that it does not
588 conflate neighboring pieces. The covered() method cannot do this because
589 it has no recent_events array at hand. But register_one has it and could
590 do it and for some reason misses to do it (sometimes).
592 This means that the three tests I just wrote can probably not survive
593 because they test with an already broken Done structure.
595 The art now is to detect how it happens, then to reproduce, then write a
598 So from the logfile this is what happens: we have a good interval with
599 newest file being F1 at T1. Now remotely F1 gets a change and F2 goes on
600 top of it. Locally we now mirror F2 and open a new done interval for it.
601 Then we mirror F1 but this time with the timestamp T1b. And when we then
602 try to close the gap, we do not find T1 but instead something older. We
603 should gladly accept this older piece and this would fix this bug.
607 * bug to fix: when the 1h file changes while rmirror is running, we do
608 correctly sync the new files but never switch to the 6h file but rather
609 stay in a rather quick loop that fetches the 1h file again and again.
611 Is it possible that we initialize a new object? Or does
612 get_remote_recentfile_as_tempfile overwrite something in myself?
614 Want a new option: _runstatusfile => $file which frequently dumps the
615 state of all recentfiles to a file.
619 2008-10-04 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
621 * Todo: now teach update to verify the timestamp is about to write
622 against the previous and use _increase_a_bit if it doesn't comply with
623 strict monotony. DONE
625 * The problem of rounding. So far perl's default precision was
626 sufficient. One day it won't be. FakeFloat has an easy job when it is
627 only reading and other machines have written correctly. But when we want
628 to write a floating point number that is a bit larger than the other
629 one, then we need our own idea of precision.
631 Slaven said: just append a "1". This might be going towards the end of
632 usability too quickly. I'd like something that actually uses the decimal
633 system. Well, appending a 1 also does this but...
635 E.g. we have 1.0. nextup on this architecture is starting with
636 1.0000000000000004. So there is a gap to fill: 1,2,3. Now I have
637 taken the 1.0000000000000003 and the next user comes and the time tells
638 him 1.0 again. He has to beat my number without stepping over the
639 nextup. This is much less space than I had when I chose 1,2,3.
641 What is also irritating is that nextup is architecture dependent. The
642 128 bit guy must choose very long numbers to fit in between whereas the
643 other one with 16 bit uses larger steps. But then the algorithm is the
644 same for both, so that would be a nice thing.
646 I see two situation where we need this. One is when Time::HiRes returns
647 us a value that is <= the last entry in our recentfile. In this case
648 (let's call it the end-case) we must fill the region between that number
649 and the next higher native floating point number. The other is when we
650 inject an old file into an old recentfile (we would then also set a new
651 dirtymark). We find the integer value already taken and need a slightly
652 different one (let's call it the middle-case). The difference between
653 the two situations is that the next user will want to find something
654 higher than my number in the end-case and something lower than my number
657 So I suggest we give the function both a value and an upper bound and it
658 calculates us a primitive middle. The upper bound in the middle-case is
659 the next integer. The upper bound on the end-case is the nextup floating
660 point number. But the latter poses another problem: if we have occupied
661 the middle m between x and nextup(x), then the nextup(m) will probably
662 not be the same as nextup(x) because some rounding will take place
663 before the nextup is calculated and when the rounding reaches the
664 nextup(x), we will end up at nextup(nextup(x)).
666 So we really need to consider the nextup and the nextdown from there and
667 then the middle and that's the number we may approach asymptotically.
670 2008-10-03 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
672 * consider deprecating the use of RECENT.recent as a symlink. It turns
673 out to need extra hoops with the rsync options and just isn't worth it.
674 Or maybe these extra hoops are needed anyway for the rest of the tree?
675 Nope, can't be the case because not all filesystems support symlinks.
677 But before doing the large step, I'll deprecate the call of
678 get_remote_recentfile_as_tempfile with an argument. Rememberr this was
679 only introduced to resolve RECENT.recent and complicates the routine far
680 beyond what it deserves.
682 DONE. Won't deprecate RECENT.recent, just moved its handling to the
685 2008-10-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
687 * I think it's a bug that the rsync_option links must be set to true in
688 order to support RECENT.recent and that nobody cares to set it
689 automatically. Similar for ignore_link_stat_errors.
691 2008-09-27 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
693 * Todo: find all todos together and make a plan what is missing for a
696 - verifytree or something like that. fsck maybe.
698 - rersyncrecent, the script itself? What it do?
700 - a way to only mirror the recentfiles without mirroring the whole
701 remote system such that people can decide to mirror only partially see
702 also 2008-08-30. .shadow-xxx directory? this also needed for a
703 filesystem that is still incomplete and might need the mirrorfiles for
706 - long living objects that mirror again and again. Inject something
707 into ta, see how it goes over to tb.
709 - how do we continue filling up the DONE system when we use an object
710 for the second time? "fully covered" and "uptodate" or new terminology.
712 - overview called on the wrong file should be understandable
714 - the meta data field that must change when we fake something up so that
715 the downstream people know they have to re-fetch everything.
717 - how tolerant are we against missing files upstream? how do we keep
718 track? there are legitimate cases where we did read upstream index right
719 before a file got deleted there and then find that file as new and want
720 it. There are other cases that are not self healing and must be tracked
723 - how, exactly, do we have to deal with deletes? With rsync errors?
725 rsync: link_stat "/id/K/KA/KARMAN/Rose-HTMLx-Form-Related-0.07.meta" (in
726 authors) failed: No such file or directory (2)
728 The file above is a delete in 1h and a new in file 1M and the
729 delete in the locally running rmirror did not get propagated to the 1M
730 object. Bug. And the consequence is a standstill.
732 It seems that a slave that works with a file below the principal needs
733 to merge things all the way up to get rid of later deletes. Or keep
734 track of all deletes and skip them later. So we need a trackdeletes.pm
735 similar to the done.pm?
737 see also 2008-08-20 about spurious deletes that really have no add
738 counterpart and yet they are not wrong.
740 - consider the effect when resyncing the recentfile takes longer than
741 the time per loop. Then we never rsync any file. We need to diagnose
742 that and force an increase of that loop time. But when we later are fast
743 enough again because the net has recovered, then we need to switch back
744 to original parameters. ERm, no, it's enough to keep syncing at least
745 one file before refetching an index file.
747 - remember to verify that no temp files are left lying around and the
750 - status file for not long running jobs that want to track upstream with
753 - revisit all XXX _float areas and study Sub::Exporter DONE
755 - persistent DB even though we just said we do not need it. Just for
756 extended capabilities and time savings when, for example, upstream
757 announces a reset and we get new recentfiles and could then limit
758 ourselves to a subset of files (those that have a changed epoch) in a
759 first pass and would only then do the loop to verify the rest. Or
762 * Todo: aggregate files should know their feed and finding the principal
763 should be done stepwise. (?)
765 * Todo: DESTROY thing that unlocks. Today when I left the debuggerr I
766 left locks around. DONE
768 2008-09-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
770 * maybe extend the _overview so that it always says if and where the
771 last file is in the next file and where the next event in the next rf
772 would lie. No, don't like this anymore. REJECT
774 * take the two new redundant tests out again, only the third must
777 * Todo: add a sanity check if the merged structure is really pointing to
778 a different rf and that this different rf is larger. DONE
780 2008-09-25 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
782 * now test, if they are overlapping. And test if there is a file in the
783 next rf that would fit into this rf's interval.
785 1h 1222324012.8474 1222322541.7963 0.4086
786 6h 1222320411.2760 1222304207.6931 4.5010 missing overlap/gap!
787 1d 1222320411.2760 1222238750.5071 22.6835 large overlap
788 1W 1222313218.3626 1221708477.5829 167.9835
790 I suspect that somebody writes a merged timestamp without having merged
791 and then somebody else relies on it.
793 If aggregate is running, the intervals must not be extravagated, if it
794 is not running, there must not be bounds, the total number of events in
795 the system must be counted and must be controlled throughout the tests.
796 That the test required the additional update was probably nonsense,
797 because aggregate can cut pieces too. FIXED & DONE
799 2008-09-23 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
801 * rrr-aggregate seems to rewrite the RECENT file even if nothing has
804 2008-09-21 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
806 * Most apparent bug at the moment is that the recentfiles are fetched
807 too often. Only the principal should be fetched and if it has not
808 changed, the others should not be refetched. ATM I must admit that I'm
809 happy that we refetch more often than needed because I can more easily
810 fix bugs while the thing is running.
812 * Let's say, 1220474966.19501 is a timestamp of a file that is already
813 done but the done system does not know about it. The reason for the
814 failure is not known and we never reach the status uptodate because of
815 this. We must get over it.
817 Later it turns out that the origin server had a bug somewhere.
818 1220474966.19042 came after 1220474966.19501. Or better: it was in the
819 array of the recentfile one position above. The bug was my own.
821 2008-09-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
823 * There is the race condition where the server does a delete and the
824 slave does not yet know and then tries to download it because he sees
825 the new. So for this time window we must be more tolerant against
826 failure. If we cannot download a file, we should just skip it and should
827 not retry immediately. The whole system should discover the lost thing
828 later. Keeping track with the DONE system should really be a no brainer.
830 But there is something more: the whole filesystem is a database and the
831 recentfiles are one possible representation of it. It's a pretty useful
832 representation I think that's why I have implemented something around
833 it. But for strictly local operation it has little value. For local
834 operation we would much rather have a database. So we would enter every
835 recentfile reading and every rsync operation and for every file the last
836 state change and what it leads to. Then we would always ignore older
837 records without the efforts involved with recentfiles.
839 The database would have: path,recentepoch,rsyncedon,deletedon
841 Oh well, not yet clear where this leads to.
843 2008-09-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
845 * Bug: the bigloop ran into a funny endless loop after EWILHELM uploaded
846 Module-Build. It *only* rsynced the "1h" recentfile from that moment on.
848 * statusfile, maybe only on demand, alone to have a sharp debugging
849 tool. It is locked and all recentfiles dump themselves into it and we
850 can build a viewer that lets us know where we stand and what's inside.
852 * remember: only the principal recentfile needs expiration, all others
853 shall be expired by principal if it discovers that something has move
856 2008-09-18 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
858 * Always check if we stringify to a higher value than in the entry
861 * And in covered make an additional check if we would be able to see a
862 numerical difference between the two numbers and if we can't then switch
863 to a different, more expensive algorithm. Do not want to be caught by
864 floating surprises. DONE
866 2008-09-17 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
868 * caching has several aspects here: we can cache the interval of the
869 recentfile which only will change when the mtime of the file changes. We
870 must re-mirror the recentfile when its ttl has expired. Does have_read
871 tell you anything? It counts nothing at all. Only the mtime is
872 interesting. The ntuple mtime, low-epoch, high-epoch. And as a separate
873 thing the have_mirrored because it is unrelated to the mtime.
875 * Robustness of floating point calculations! I always thought that the
876 string calculated by the origin server for the floating representation
877 of the epoch time is just a string. When we convert it to a number and
878 later back to a string, the other computer might come to a different
879 conclusion. This must not happen, we want to preserve it under any
880 circumstances. I will have to write tests with overlong sequences that
881 get lost in arithmetic and must see if all still works well. DONE
883 But one fragile point remains: if one host considers a>b and the other
884 one considers them == but no eq. To prevent this, we must probably do
885 some extra homework. DONE
887 2008-09-16 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
889 * the concept of tracking DONE needs an object per recentfile that has
890 something like these methods:
892 do_we_have(xxx), we_have(xxx), do_we_have_all(xxx,yyy), reset()
894 covered() register() covered()
896 The unclear thing is how we translate points in time into intervals. We
897 could pass a reference to the current recent_events array when running
898 we_have(xxx) and let the DONE object iterate over it such that it only
899 has to store a list of intervals that can melt into each other. Ah, even
900 passing the list together with a list of indexes seems feasiable.
902 Or maybe ask for the inverted list?
904 Whenever the complete array is covered by the interval we say we are
905 fully covered and if the recentfile is not expired, we are uptodate.
907 2008-09-07 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
909 2008-09-05 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
911 * need a way to "return" the next entry after the end of a list. When
912 the caller says "before" or "after" we would like to know if he could
913 cover that interval/threshold or not because this influences the effect
914 of a newer timestamp of that recentfile. DONE with $opt{info}.
916 2008-09-04 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
918 * one of the next things to tackle: the equivalent of csync2 -TIXU.
920 loop implies tixu (?). Nope, something like --statefile decides. Per
923 T test, I init, X including removals, U nodirtymark
925 So we have no concept of dirtymarks, we only trust that since we are
926 running we have observed everything steadily. But people will not let
927 this program run forever so we must consider both startup penalty and
928 book keeping for later runs. We keep this for later. For now we write a
929 long running mirror that merges several intervals.
931 2008-09-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
933 * need to speed up the 02 test, it's not clever to sleep so much. Reduce
936 * rersyncrecent, the script: default to one week. The name of the switch
937 is --after. Other switches? --loop!
939 2008-08-30 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
941 * need a switch --skip-deletes (?)
943 * need a switch --enduser that tells us that the whole tempfile
944 discipline is not needed when there is no downstream user. (?)
946 Without this switch we cannot have a reasonable recent.pl that just
947 displays the recent additions. Either we accept to download everything.
948 Or we download temporary files without the typical rsync protocol
951 Or maybe the switch is --tmpdir? If --tmpdir would mean: do not use
952 File::Temp::tempdir, this might be a win.
954 2008-08-29 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
956 * apropos missing: we have no push, we never know the downstream
957 servers. People who know their downstream hosts and want to ascertain
958 something will want additional methods we have never thought about, like
959 update or delete a certain file.
961 2008-08-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
963 * tempted to refactor rmirror into resolve_symlink, localize, etc.
964 Curious if rsync_options=links equal 0 vs. 1 will make the expected
967 * rsync options: it's a bit of a pain that we usually need several rsync
968 options, like compress, links, times, checksum and that there is no
969 reasonable default except the original rsync default. I think wee can
970 safely assume that the rsync options are shared between all recentfile
971 instances within one recent tree.
973 2008-08-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
975 * deletes: if a delete follows an add quickly enough it may happen that
976 a downstream mirror did not see the add at all! It seems this needs to
977 be mentioned somewhere. The point here is that even if the downstream is
978 never missing the principal timeframe it may encounter a "delete" that
979 has no complimentary "add" anywhere.
981 2008-08-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
983 * I suspect the treat of metadata is incorrect during read or something.
984 The bug that I am watching is that between 06:08 and 06:09 the 6h file
985 contained more than 6 hours worth of data. At 06:08 we merged into the
986 1d file. We need to take snapshots of the 6h file over the course of an
987 hour or maybe only between XX:08 and XX:09? Nope, the latter is not
990 Much worse: watching the 1h file: right at the moment (at 06:35) it
991 covers 1218867584-1219120397 which is 70 hours.
993 Something terribly broken. BTW, 1218867584 corresponds to Sat Aug 16
994 08:19:44 2008, that is when I checked out last time, so it seems to be
995 aggregating and never truncating?
997 No, correct is: it is never truncating; but wrong is: it is aggregating.
998 It does receive a lot of events from time to time from a larger file.
999 Somehow a large file gets merged into the small one and because the
1000 "meta/merged" attribute is missing, nobody is paying attention. I
1001 believe that I can fix this by making sure that metadata are honoured
1002 during read. DONE and test adjusted.
1004 2008-08-17 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1006 * grand renaming plan
1008 remotebase => remoteroot to fit well with localroot DONE
1009 local_path() => localroot seems to me should already work DONE
1010 recentfile_basename => rfilename no need to stress it has no slash DONE
1012 filenameroot??? Doesn't seem too bad to me today. Maybe something like
1013 kern? It would anyway need a deprecation cycle because it is an
1014 important constructor.
1016 * I like the portability that Data::Serializer brings us but the price
1017 is that some day we might find out that it is slowing us a bit. We'll
1020 2008-08-16 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1022 * should we not enter the interval of the principal (or the interval of
1023 the merging file?) in every aggregated/merged file?
1025 * we should aim at a first release and give up on thinking about
1026 sanitizing stuff and zloop. Let's just admit that a full traditional
1027 rsync is the only available sanitizer ATM. Otherwise it's complicated
1028 stuff: sanitizing on the origin server, sanitizing on the slaves,
1029 sanitizing forgotten files, broken timestamps, etc. Let's delay it and
1030 get the basics out before this becomes a major cause for mess.
1032 2008-08-13 Andreas Koenig <k@andreas-koenigs-computer.local>
1034 * On OSes not supporting symlinks we expect that RECENT.recent contains
1035 the contents of the principal recentfile. Actually this is identical on
1036 systems supporting symlinks. Simple, what follows from that is that we
1037 need to keep the serializer in the metadata because we cannot read it
1038 from the filename, doesn't it? Of course not. It's a chicken and egg
1039 problem. This leaves us with the problem to actually parse the
1040 serialized data to find out in which format it is. So who can do the 4
1041 or 5 magics we wanted to support? File::LibMagic?
1043 2008-08-09 Andreas Koenig <k@andreas-koenigs-computer.local>
1045 * remotebase and recentfile_basename are ugly names. Now that we need a
1046 word for the shortest/principal/driving recentfile too we should do
1049 localroot is good. rfile is good. local_path() is bad, local_path($path)
1050 is medium, filenameroot() is bad, remotebase is bad, recentfile is
1053 Up to now remotebase was the string that described the remote root
1054 directory in rsync notation, like pause.perl.org::authors. And
1055 recentfile_basename was "RECENT-1h.yaml".
1057 2008-08-08 Andreas Koenig <k@andreas-koenigs-computer.local>
1059 * The test that was added in today's checkin is a good start for a test
1060 of rmirror. We should have more methods in Recent.pm: verify,
1061 addmissingfiles. We should verify the current tree, then rmirror it and
1062 then verifytree the copy. We could then add some arbitrary file and let
1063 it be discovered by addmissingfiles, then rmirror again and then
1064 verifytree the copy again.
1066 Then we could start stealing from csync2 sqlite database [no port to
1067 OSX!] and fill a local DB. And methods to compare the database with the
1068 recentfiles. Our strength is that in principle we could maintain state
1069 with a single float. We have synced up to 1234567890.123456. If the Z
1070 file does not add new files all we have to do is mirror the new ones and
1073 This makes it clear that we should extend current protocol and declare
1074 that we cheat when we add files too late, just to help the other end
1075 keeping track. Ah yes, that's what was meant when zloop was mentioned
1078 Maybe need to revisit File::Mirror to help me with this task.
1080 2008-08-07 Andreas Koenig <k@andreas-koenigs-computer.local>
1082 * There must be an allow-me-to-truncate flag in every recentfile.
1083 Without it one could construct a sequence of updates winning the locking
1084 battle against the aggregator. Only if an aggregator has managed to
1085 merge data over to the next level, truncating can be allowed. DONE with
1088 2008-08-06 Andreas Koenig <k@andreas-koenigs-computer.local>
1090 * We should probably guarantee that no duplicates enter the aggregator
1093 2008-08-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1095 * To get merge operation faster would need a good benchmark test. What
1096 02 spits out isn't reliable enough and is dominated by many other
1099 commit 10176bf6b79865d4fe9f46e3857a3b8669fa7961
1100 Author: Andreas J. Koenig <k@k75.(none)>
1101 Date: Sat Aug 2 07:58:04 2008 +0200
1105 commit 3243120a0c120aaddcd9b1f4db6689ff12ed2523
1106 Author: Andreas J. Koenig <k@k75.(none)>
1107 Date: Sat Aug 2 11:40:29 2008 +0200
1109 there was a lot of trying but the effect is hardly measurable with
1112 * overhead of connecting seems high. When setting
1113 max_files_per_connection to 1 we see that.
1115 2008-08-01 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1117 * 1217622571.0889 - 1217597432.86734 = 25138.2215600014
1119 25138.2215600014/3600 = 6.98283932222261
1121 It jumps into the eye that this is ~ 7 hours, not ~6, so there seems to
1122 be a bug in the aggregator. FIXED
1124 2008-07-27 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1126 * e.g. id/Y/YE/YEWENBIN/Emacs-PDE-0.2.16.tar.gz: Do we have it, should
1127 we have it, can we mirror it, mirror it!
1129 I fear this needs a new class which might be called
1130 File::Rsync::Mirror::Recent. It would collect all recentfiles of a kind
1131 and treat them as an entity. I realize that a single recentfile may be
1132 sufficient for certain tasks and that it is handy for the low level
1133 programmer but it is not nice to use. If there is a delete in the 1h
1134 file then the 6h file still contains it. Seekers of the best information
1135 need to combine at least some of the recentfiles most of the time.
1137 There is the place for the Z loop!
1139 But the combination is something to collect in a database, isn't it. Did
1140 csync2 just harrumph?
1142 2008-07-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1144 * it just occurred to me that hosts in the same mirroring pool could
1145 help out each other even without rewriting the recentfile. Just fetch
1146 the stuff to mirror from several places, bingo. But that's something
1147 that should rather live in a separate package or in rsync directly.
1149 * cronjobs are unsuited because with ntp they would all come at the full
1150 minute and disturb each other. Besides that I'd hate to have a backbone
1151 with more than a few seconds latency.
1153 2008-07-25 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1155 * a second rsync server with access control for PAUSE. Port? 873 is the
1156 standard port, let's take 8873.
1158 * if there were a filesystem based on this, it would have a slow access
1159 to inexistent files. It would probably provide wrong readdir (only based
1160 on current content) or also a slow one (based on a recentfile written
1161 after the call). But it would provide fast access to existing files. Or
1162 one would deliberately allow slightly blurred answers based on some
1163 sqlite reflection of the recentfiles.
1165 * todo: write a variant of mirror() that combines two or more
1166 recentfiles and treats them like one
1168 * todo: signal handler to remove the tempfile
1170 2008-07-24 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1172 * now that we have the symlink I forgot how it should be used in
1175 * the z loop: add missing files to Z file. Just append them (instead of
1176 prepending). So one guy prepends something from the Y file from time to
1177 time and another guy appends something rather frequently. Collecting
1178 pond. When Y merges into Z, things get epoch and the collecting pond
1179 gets smaller. What exactly are "missing files"?
1181 take note of current epoch of the alpha file, let's call it the
1184 find all files on disk
1186 remove all files registered in the recentworld up to recent-ts
1188 remove all files that have been deleted after recent-ts according to
1191 2008-07-23 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1193 * rersyncrecent might be a cronjob with a (locked) state file which
1194 contains things like after and maybe last z sync or such?
1196 rrr-mirror might be an alternative name but how would we justify the
1197 three Rs when there is no Re-Rsync-Recent?
1199 With the --loop parameter it is an endless loop, without it is no loop.
1200 At least this is simple.
1202 * todo: new accssor z-interval specifies how often the Z file is updated
1203 against the filesystem. We probably want no epoch stamp on these
1204 entries. And we want to be able to filter the entries (e.g. no
1205 by-modules and by-category tree)
1207 2008-07-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1209 * Fill the Z file. gc or fsck or both. Somehow we must get the old files
1210 into Z. We do not need the other files filled up with filesystem
1213 * need interface to query for a file in order to NOT call update on
1214 PAUSE a second time within a short time.
1216 2008-07-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1218 * recommended update interval? Makes no sense, is different for
1225 change-log-default-name: "Todo"