Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 8:52 AM, Stephan Beal sgb...@googlemail.com wrote: On Fri, Oct 4, 2013 at 6:27 AM, Ron Aaron r...@ronware.org wrote: I have a large-ish repo, and after cloning it and updating some files, I try to commit ... and get this message. I did already try fossil rebuild to no effect. eeek! It should never be possible that that happens _after_ a commit (if it's going to happen it should fail during the before-commit checking). Spoke too soon: it DID happen while committing, which means nothing is permanently broken :-D But that also means that the command i sent you won't work because rid 12892 gets rolled back. Hmmm... we might need you to add some printfs() to fossil to debug this one, unless there's a switch i'm unaware of to dump the manifest out before committing. What platform are you on? Is there any weird filename encoding in play (Windows/Mac)? -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 8:57 AM, Stephan Beal sgb...@googlemail.com wrote: Hmmm... we might need you to add some printfs() to fossil to debug this one, unless there's a switch i'm unaware of to dump the manifest out before committing. i need to leave for a while, but here's something we can try, and with any luck i'll be back about the time you've got this running (i'm naively assuming you're running a fossil built locally): http://fossil-scm.org/index.html/artifact/468d7f73a4b09fe03d16d8477f9fe6a99481bec9?ln=494 right after that line, please add: cgi_printf(TEST FOR RON:\n%b, pOut); all that will do is dump the manifest (and any other stuff it tries to save) to stdout. Please try with a small commit first, because this will also dump the contents of the file(s) being committed, and send the output to a file: # fossil commit foo.out what i'm mostly interested in seeing is the contents of the manifest file itself. The error is happening during the before-commit verification of the manifest because it cannot load that manifest: http://fossil-scm.org/index.html/artifact/0e2f3aa96c9d2fc79f158f3944378b1e7fe8f3ec?ln=868 -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
Sorry, Ron: 2nd arg is pBlob, not pOut. (mobile phone - please excuse brevity and typos) - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal On Oct 4, 2013 9:17 AM, Stephan Beal sgb...@googlemail.com wrote: On Fri, Oct 4, 2013 at 8:57 AM, Stephan Beal sgb...@googlemail.comwrote: Hmmm... we might need you to add some printfs() to fossil to debug this one, unless there's a switch i'm unaware of to dump the manifest out before committing. i need to leave for a while, but here's something we can try, and with any luck i'll be back about the time you've got this running (i'm naively assuming you're running a fossil built locally): http://fossil-scm.org/index.html/artifact/468d7f73a4b09fe03d16d8477f9fe6a99481bec9?ln=494 right after that line, please add: cgi_printf(TEST FOR RON:\n%b, pOut); all that will do is dump the manifest (and any other stuff it tries to save) to stdout. Please try with a small commit first, because this will also dump the contents of the file(s) being committed, and send the output to a file: # fossil commit foo.out what i'm mostly interested in seeing is the contents of the manifest file itself. The error is happening during the before-commit verification of the manifest because it cannot load that manifest: http://fossil-scm.org/index.html/artifact/0e2f3aa96c9d2fc79f158f3944378b1e7fe8f3ec?ln=868 -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
2013/10/4 Stephan Beal sgb...@googlemail.com: On Fri, Oct 4, 2013 at 8:57 AM, Stephan Beal sgb...@googlemail.com wrote: Hmmm... we might need you to add some printfs() to fossil to debug this one, unless there's a switch i'm unaware of to dump the manifest out before committing. Shouldn't the error-message in this case be more helpful anyway? Something like: http://fossil-scm.org/index.html/info/1eb438d61a Regards, Jan Nijtmans ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 11:30 AM, Jan Nijtmans jan.nijtm...@gmail.comwrote: Shouldn't the error-message in this case be more helpful anyway? Something like: http://fossil-scm.org/index.html/info/1eb438d61a That's essentially how libfossil reports the errors for this case. At this point i don't know if Ron's problem is caused at parse-time (i don't think so - IIRC, manifest parsing errors are immediately fatal) or if it's trying to load a manifest which is not there (which i find hard to believe), or some such. If i can get a copy of the manifest i can run it through the libfossil manifest parser as a sanity check :). -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 12:10 PM, Jan Nijtmans jan.nijtm...@gmail.comwrote: Fully agreed! I also don't think that the parser is the problem, maybe it's a bug somewhere else which does not properly fossilize some field which ends up in the manifest later. My hope is that the That's my suspicion. i just got a copy of the repo off-list and am looking into it now using your patched error reporting. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 12:13 PM, Richard Hipp d...@sqlite.org wrote: Doesn't the -n option cause the manifest to be printed on standard output? Quite possibly so - never used it :/. In any case, i'm cloning the repo now (going a bit slowly) and will report back as soon as i know something concrete (and will report whether or not -n spits out the manifest). -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 12:19 PM, Stephan Beal sgb...@googlemail.com wrote: In any case, i'm cloning the repo now (going a bit slowly) and will report back as soon as i know something concrete (and will report whether or not -n spits out the manifest). Follow-up: it's not the debugging which is taking so long, but the reproduce/build process (cloning a remote site) has been running for about 80 minutes now and i have no idea how long it will run. @Ron: this has prio 1 for me today (other than walking my dog ;), and i'll be in touch as soon as possible. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 1:53 PM, Stephan Beal sgb...@googlemail.com wrote: Follow-up: it's not the debugging which is taking so long, but the reproduce/build process (cloning a remote site) has been running for about 80 minutes now and i have no idea how long it will run. @Ron: this has prio 1 for me today (other than walking my dog ;), and i'll be in touch as soon as possible. Bad news: i can't reproduce it :/ [stephan@host:~/cvs/fossil/ron]$ f set | grep autosy autosync (local) 0 [stephan@host:~/cvs/fossil/ron]$ f com -m 'crash test' 21 | tee foo New_Version: 6d5ef0c1b9988dc2400e98024b3ac42d0f14a976 [stephan@host:~/cvs/fossil/ron]$ cat foo New_Version: 6d5ef0c1b9988dc2400e98024b3ac42d0f14a976 [stephan@host:~/cvs/fossil/ron]$ f time -n 1 === 2013-10-04 === 12:13:54 [6d5ef0c1b9] *CURRENT* crash test (user: stephan tags: trunk) [stephan@host:~/cvs/fossil/ron]$ ../libfossil/f-timeline -n 2 ci [6d5ef0c1b998] @ 2013-10-04 14:13:54 by [stephan] in branch [trunk] crash test ci [6c0d3704e381] @ 2013-01-15 11:00:18 by [ron] in branch [trunk] no comment [stephan@host:~/cvs/fossil/ron]$ ../libfossil/f-ls | wc -l 10806 [stephan@host:~/cvs/fossil/ron]$ ../libfossil/f-acat current | head C crash\stest D 2013-10-04T12:13:54.361 F a.js 8051128fc23969a5894bd916de0719fde0a94f29 F about.htm a9c3c9d80131b5021e292e8be0f30957dd836313 F amex.gif 14979ef159d2c3a7caa72215ab7f9824cb7a2854 ... [stephan@host:~/cvs/fossil/ron]$ ../libfossil/f-ls -v current 1 [stephan@host:~/cvs/fossil/ron]$ ../libfossil/f-ls -v prev 2 [stephan@host:~/cvs/fossil/ron]$ diff 1 2 | less 1c1 File list from manifest version 'current' [6d5ef0c1b998] (RID 12814)... --- File list from manifest version 'prev' [6c0d3704e381] (RID 10825)... ... lots of file diffs and the big RID diff between the current and prev commits implies lots of changes were made. i backed it up before the commit, so i can try a few more things out without having to rebuild, but without a reproduce i won't be able to do much :/. This is 64-bit Ubuntu 12.04.2, btw. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 2:22 PM, Stephan Beal sgb...@googlemail.com wrote: ... lots of file diffs and the big RID diff between the current and prev commits implies lots of changes were made. And... [stephan@host:~/cvs/fossil/ron]$ ../libfossil/f-acat tip | ../libfossil/f-mfparse -r Parsing this manifest: - Artifact type=MANIFEST, rid=12814, uuid=6d5ef0c1b9988dc2400e98024b3ac42d0f14a976 10804 files seen in manifest(s). Trying to re-calculate R-card: original=[136592d60d99845173057784350a3afb] Re-calculated R-card: [136592d60d99845173057784350a3afb] R-card match? yes Round-trip re-generated artifact (type=MANIFEST) from input file: Rather large - not dumping to console. Dumping mf to file [mf.out] SHA of [mf.out] = [6d5ef0c1b9988dc2400e98024b3ac42d0f14a976] SHA match? yes i.e. the manifest looks good. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
[fossil-users] Run fossil behind stunnel?
How do I run fossil behind stunnel? Besides getting stunnel to work, which was not dead simple, now there is some problem with the interaction with fossil. I have fossil running via inetd on port 79, and stunnel on port 779. Connecting on 779 with https works fine at first, connecting to https://myhost.com:779 works, but the browser gets redirected to https://www.myhost.com:779 and breaks. Manually entering https://www.myhost.com:779/login?g=/home works but no images are shown, and there is a BASE tag in the html which points to: base href=http://www.myhost.com:779/login; / This does not look right, no https, and Firefox also complains about sending the results of the form a non-encrypted URL. Any ideas about what is going on? Certainly, putting SSL on Fossil is NOT as simple as just drop it behind stunnel as implied on the web page. Of course, I could put it in Apache as a CGI or whatever, but it bothers me to not understand why it's not working. Could not find any config option inside fossil UI either which seemed to make any difference. best regards, Jakob ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
2013/10/4 Stephan Beal sgb...@googlemail.com: On Fri, Oct 4, 2013 at 2:22 PM, Stephan Beal sgb...@googlemail.com wrote: ... lots of file diffs and the big RID diff between the current and prev commits implies lots of changes were made. I did the same, and everything looks fine. But it indeed takes a lot of time. I did it on Cygwin64, without seeing this error message. Noting that both Stephan and I tried it on a 64-bit machine with a lot of memory, could this be an unchecked memory overflow somewhere? Regards, Jan Nijtmans ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 3:00 PM, Jan Nijtmans jan.nijtm...@gmail.com wrote:
Noting that both Stephan and I tried it on a 64-bit machine with a lot
of memory, could this be an unchecked memory overflow somewhere?
i'm also working on a repro in a 32-bit vm in parallel. In theory a
malloc failure is not possible because fossil_malloc() aborts on alloc
error, but we do have a few places which use malloc()/free() instead of
fossil_{malloc,free}().
--
- stephan beal
http://wanderinghorse.net/home/stephan/
http://gplus.to/sgbeal
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 3:01 PM, Stephan Beal sgb...@googlemail.com wrote:
i'm also working on a repro in a 32-bit vm in parallel. In theory a
malloc failure is not possible because fossil_malloc() aborts on alloc
error, but we do have a few places which use malloc()/free() instead of
fossil_{malloc,free}().
Update: we've got the broken manifest now (off-list) and have found the
problem (malformed/empty P-card) but don't yet know why. Jan and i are
playing with it.
@Devs: anyone touched P-card generation lately?
...
F xr.css 35f07555a0420a382004217172b613bd38a3a3f4
P
R e7fa81c6d596b4d5bf77ffdeb62d71ba
U ron
...
the only commit which can legally have a empty P-card is the initial commit
(or one without a parent, but i think that's only possible on the first
commit?).
--
- stephan beal
http://wanderinghorse.net/home/stephan/
http://gplus.to/sgbeal
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 3:01 PM, Stephan Beal sgb...@googlemail.com wrote:
i'm also working on a repro in a 32-bit vm in parallel. In theory a
malloc failure is not possible because fossil_malloc() aborts on alloc
error, but we do have a few places which use malloc()/free() instead of
fossil_{malloc,free}().
Also work on 32-bit for me. Normally one is happy about success, but this
time not.
i'm at a loss.
The problem is definitely the empty P-card (meaning that the manifest looks
like it has no parent version, which is just wrong), but i cannot say why
that happens. As Jan hypothesized, this sounds almost like a memory misuse,
but that seems to have been ruled out. It's too specific of a problem (and
reproducible on Ron's end) to be generic memory corrupt.
Thankfully, the verify-before-commit logic catches this and keeps this from
being immortalized in the repo.
@Ron: which fossil version are you on, and what platform?
--
- stephan beal
http://wanderinghorse.net/home/stephan/
http://gplus.to/sgbeal
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On 10/04/2013 04:24 PM, Stephan Beal wrote: @Ron: which fossil version are you on, and what platform? I'm using fossil 816e893d3b on Linux Mint 15 64bit (kernel 3.8.0-30-generic) -- For confidential messages, please use my GnuPG key http://ronware.org/gpg_key.html ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 3:28 PM, Ron Aaron r...@ronware.org wrote: I'm using fossil 816e893d3b on Linux Mint 15 64bit (kernel 3.8.0-30-generic) Bummer: still works for me :( [stephan@host:~/cvs/fossil/ron]$ f version This is fossil version 1.27 [816e893d3b] 2013-10-04 02:50:26 UTC [stephan@host:~/cvs/fossil/ron]$ fst | head repository: /home/stephan/cvs/fossil/ron/../mtr.fsl local-root: /home/stephan/cvs/fossil/ron/ config-db:/home/stephan/.fossil checkout: 6c0d3704e381430edaef8424b1c1d8d059325a65 2013-01-15 10:00:18 UTC parent: 221b639542e0ef25f06e59a3fc24f6bf5ec9eda3 2012-12-07 09:01:38 UTC tags: trunk comment: no comment (user: ron) EDITED b/r/r1503.htm EDITED b/r/r1903.htm EDITED b/r/r2111.htm [stephan@host:~/cvs/fossil/ron]$ f set autosync 0 [stephan@host:~/cvs/fossil/ron]$ f com -m blah 12 | tee foo New_Version: cfc9450f5ff5a2c3d17d03ceccb0f4617d0d2520 [stephan@host:~/cvs/fossil/ron]$ f artifact cfc9450f5ff5a2c3d17d03ceccb0f4617d0d2520 | grep '^P ' P 6c0d3704e381430edaef8424b1c1d8d059325a65 that's the expected value. @Ron: please try: fossil test-integrity The output on my machine (and Jan's - thanks to him for this idea): [stephan@host:~/cvs/fossil/ron]$ f test-integrity 10826 non-phantom blobs (out of 10826 total) checked: 0 errors [stephan@host:~/cvs/fossil/ron]$ f test-integrity --parse 10826 non-phantom blobs (out of 10826 total) checked: 0 errors 22 total control artifacts 6 manifests 16 clusters -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
ron@monster ~/proj/mtrweb $ fossil test-integrity 10826 non-phantom blobs (out of 10826 total) checked: 0 errors ron@monster ~/proj/mtrweb $ fossil test-integrity --parse 10826 non-phantom blobs (out of 10826 total) checked: 0 errors 22 total control artifacts 6 manifests 16 clusters sigh On 10/04/2013 04:37 PM, Stephan Beal wrote: On Fri, Oct 4, 2013 at 3:28 PM, Ron Aaron r...@ronware.org mailto:r...@ronware.org wrote: I'm using fossil 816e893d3b on Linux Mint 15 64bit (kernel 3.8.0-30-generic) Bummer: still works for me :( [stephan@host:~/cvs/fossil/ron]$ f version This is fossil version 1.27 [816e893d3b] 2013-10-04 02:50:26 UTC [stephan@host:~/cvs/fossil/ron]$ fst | head repository: /home/stephan/cvs/fossil/ron/../mtr.fsl local-root: /home/stephan/cvs/fossil/ron/ config-db:/home/stephan/.fossil checkout: 6c0d3704e381430edaef8424b1c1d8d059325a65 2013-01-15 10:00:18 UTC parent: 221b639542e0ef25f06e59a3fc24f6bf5ec9eda3 2012-12-07 09:01:38 UTC tags: trunk comment: no comment (user: ron) EDITED b/r/r1503.htm EDITED b/r/r1903.htm EDITED b/r/r2111.htm [stephan@host:~/cvs/fossil/ron]$ f set autosync 0 [stephan@host:~/cvs/fossil/ron]$ f com -m blah 12 | tee foo New_Version: cfc9450f5ff5a2c3d17d03ceccb0f4617d0d2520 [stephan@host:~/cvs/fossil/ron]$ f artifact cfc9450f5ff5a2c3d17d03ceccb0f4617d0d2520 | grep '^P ' P 6c0d3704e381430edaef8424b1c1d8d059325a65 that's the expected value. @Ron: please try: fossil test-integrity The output on my machine (and Jan's - thanks to him for this idea): [stephan@host:~/cvs/fossil/ron]$ f test-integrity 10826 non-phantom blobs (out of 10826 total) checked: 0 errors [stephan@host:~/cvs/fossil/ron]$ f test-integrity --parse 10826 non-phantom blobs (out of 10826 total) checked: 0 errors 22 total control artifacts 6 manifests 16 clusters -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users -- For confidential messages, please use my GnuPG key http://ronware.org/gpg_key.html ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
2013/10/4 Stephan Beal sgb...@googlemail.com: the only commit which can legally have a empty P-card is the initial commit (or one without a parent, but i think that's only possible on the first commit?). An empty P-card (or a manifest without a P-card) is allowed for any manifest, but then it shouldn't have a space after the P (it would look strange in the timeline, here is one: http://fossil-scm.org/index.html/timeline?c=2013-08-21T14:49:41.624 ). Of course we could remove the space when there is no parent manifest, but that would only mask the problem: in this case the parent blob should be there. Regards, Jan Nijtmans ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] Run fossil behind stunnel?
On Fri, Oct 04, 2013 at 02:58:09PM +0200, Jakob Eriksson wrote: Hi there, How do I run fossil behind stunnel? fossil http --https and there is a BASE tag in the html which points to: base href=http://www.myhost.com:779/login; / This does not look right, no https, Correct. stunnel and inetd take care of the network side of things, and fossil just speaks http on stdin and stdout. So it must be told if anything unusual is in use; and in this context, https is unusual. Certainly, putting SSL on Fossil is NOT as simple as just drop it behind stunnel as implied on the web page. Could you provide a link to the web page? Perhaps someone on the list will have access to edit it to be more clear or correct, for the next person. Could not find any config option inside fossil UI either which seemed to make any difference. fossil http --help, or http://www.fossil-scm.org/xfer/help?cmd=http, is the documentation, although it may not be clear enough if you don't already know the answer. Cheers, f -- Francis Dalyfran...@daoine.org ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
Ron, can you run: sqlite3 .fslckout '.dump v%' And post make that output available to us? The .fslckout file should be at the root of your check-out. It might be called _FOSSIL_ if your checkout was originally opened with a much older version of Fossil. -- D. Richard Hipp d...@sqlite.org ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 3:42 PM, Jan Nijtmans jan.nijtm...@gmail.com wrote: An empty P-card (or a manifest without a P-card) is allowed for any manifest, but then it shouldn't have a space after the P (it would look strange in the timeline, here is one: http://fossil-scm.org/index.html/timeline?c=2013-08-21T14:49:41.624 ). Of course we could remove the space when there is no parent manifest, but that would only mask the problem: in this case the parent blob should be there. You are certainly correct. i had the P (with space) answer backwards in my memory, but found a comment to that effect: /* Evil ugly hack, primarily for round-trip compatibility with manifest #1, which has an empty P card. fossil(1) ignores empty P-cards in all cases, and must continue to do so for backwards compatibility with rid #1 in all repos. Pedantic note: there must be no space between the 'P' and the newline. */ fsl_deck_append(os, P\n); -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
Done, you can pick it up here: http://ronware.org/dump.sql.gz Thanks! On 10/04/2013 04:44 PM, Richard Hipp wrote: sqlite3 .fslckout '.dump v%' -- For confidential messages, please use my GnuPG key http://ronware.org/gpg_key.html ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 9:50 AM, Ron Aaron r...@ronware.org wrote: Done, you can pick it up here: http://ronware.org/dump.sql.gz That looks ok. What does the following query against the repository give: SELECT uuid FROM blob WHERE rid=15513; It should yield a SHA1 hash, which will become the argument to the P card. If in your repository it yields an empty string, that would explain the malformed manifest. (Still unexplained is how the above query could possible yield an empty string.) -- D. Richard Hipp d...@sqlite.org ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 3:50 PM, Ron Aaron r...@ronware.org wrote:
Done, you can pick it up here: http://ronware.org/dump.sql.gz
i think what Richard wanted there was:
INSERT INTO vvar VALUES('checkout','15513');
So that we can check:
[stephan@host:~/cvs/fossil/ron]$ echo select uuid from blob where
rid=15513; | fossil sqlite
(no results)
(please try that on your machine)
which is somewhat surprising. i was expecting to see:
[stephan@host:~/cvs/fossil/ron]$ echo select rid from blob where
uuid='6c0d3704e381430edaef8424b1c1d8d059325a65'; | fossil sqlite
10825
That said, the RIDs can theoretically be different across sites, so this
isn't definitive.
--
- stephan beal
http://wanderinghorse.net/home/stephan/
http://gplus.to/sgbeal
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
2013/10/4 Ron Aaron r...@ronware.org:
Done, you can pick it up here: http://ronware.org/dump.sql.gz
Fossil code:
vid = db_lget_int(checkout, 0);
zParentUuid = db_text(0, SELECT uuid FROM blob WHERE rid=%d, vid);
From the dump:
INSERT INTO vvar VALUES('checkout','15513');
So the parent blob has rid 15513. that looks OK.
In my cloned repository the rid of this blob is 10825.
$ fossil sqlite
SQLite version 3.8.1 2013-09-16 12:57:19
Enter .help for instructions
Enter SQL statements terminated with a ;
sqlite SELECT * FROM blob where rid=10825;
10825|1|627733|6c0d3704e381430edaef8424b1c1d8d059325a65|
sqlite .exit
So, Ron, can you try:
$fossil sqlite
sqlite SELECT * FROM blob where rid=15513
Regards,
Jan Nijtmans
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 4:03 PM, Jan Nijtmans jan.nijtm...@gmail.com wrote: sqlite SELECT * FROM blob where rid=15513 lol... all of us racing at once to say the same thing. At least the answers are consistent, though ;). -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
Indeed, it gives an empty string. On 10/04/2013 05:02 PM, Richard Hipp wrote: SELECT uuid FROM blob WHERE rid=15513; -- For confidential messages, please use my GnuPG key http://ronware.org/gpg_key.html ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On 10/04/2013 05:02 PM, Stephan Beal wrote: echo select rid from blob where uuid='6c0d3704e381430edaef8424b1c1d8d059325a65'; | fossil sqlite For me it's 10798 -- For confidential messages, please use my GnuPG key http://ronware.org/gpg_key.html ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On 10/04/2013 05:03 PM, Jan Nijtmans wrote:
2013/10/4 Ron Aaron r...@ronware.org:
Done, you can pick it up here: http://ronware.org/dump.sql.gz
Fossil code:
vid = db_lget_int(checkout, 0);
zParentUuid = db_text(0, SELECT uuid FROM blob WHERE rid=%d, vid);
From the dump:
INSERT INTO vvar VALUES('checkout','15513');
So the parent blob has rid 15513. that looks OK.
In my cloned repository the rid of this blob is 10825.
$ fossil sqlite
SQLite version 3.8.1 2013-09-16 12:57:19
Enter .help for instructions
Enter SQL statements terminated with a ;
sqlite SELECT * FROM blob where rid=10825;
10825|1|627733|6c0d3704e381430edaef8424b1c1d8d059325a65|
sqlite .exit
So, Ron, can you try:
$fossil sqlite
sqlite SELECT * FROM blob where rid=15513
it's blank (empty) for me.
Regards,
Jan Nijtmans
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
--
For confidential messages, please use my GnuPG key
http://ronware.org/gpg_key.html
___
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
Interestingly, select max(rid) from blob gives me 10826... On 10/04/2013 05:02 PM, Richard Hipp wrote: On Fri, Oct 4, 2013 at 9:50 AM, Ron Aaron r...@ronware.org mailto:r...@ronware.org wrote: Done, you can pick it up here: http://ronware.org/dump.sql.gz That looks ok. What does the following query against the repository give: SELECT uuid FROM blob WHERE rid=15513; It should yield a SHA1 hash, which will become the argument to the P card. If in your repository it yields an empty string, that would explain the malformed manifest. (Still unexplained is how the above query could possible yield an empty string.) -- D. Richard Hipp d...@sqlite.org mailto:d...@sqlite.org ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users -- For confidential messages, please use my GnuPG key http://ronware.org/gpg_key.html ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 4:07 PM, Ron Aaron r...@ronware.org wrote: On 10/04/2013 05:02 PM, Stephan Beal wrote: echo select rid from blob where uuid='6c0d3704e381430edaef8424b1c1d8d059325a65'; | fossil sqlite For me it's 10798 Somehow the record of your checkout version and the actual checkout seem to have gotten out of sync. i can't currently speculate as to how that might happen. Rather than fiddle without the checkout db, i'd suggest trying the following: fossil close --force cd .. tar cf foo.tar $OLDPWD cd - fossil open /path/to/repo.fsl --keep cd - tar xf foo.tar # probably not necessary with --keep, but can't hurt cd - fossil status should show you your list of changes (several hundred, IIRC). And then do your checkin again. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
2013/10/4 Ron Aaron r...@ronware.org: Indeed, it gives an empty string. Then I have a theory what happened. My guess is that you cloned the repository and opened it. Later you cloned the repository again (which got different row-id's), storing it in the exact original location. Then the row-id's of the .fslckout database and the repository don't match any more. What you can do to fix this: - fossil update trunk (find the new rowid of trunk, and bring the checkout there) - fossil revert (clean up any garbage) then do a make and everything else, and try the commit. Am I close? Regards, Jan Nijtmans ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 4:10 PM, Ron Aaron r...@ronware.org wrote: Interestingly, select max(rid) from blob gives me 10826... The RIDs aren't 100% dependable as a point of reference, though. They are sequential but not necessarily in increments of one. In theory the highest-number RID must refer to an Fossil Artifact (as opposed a user file), but i'm largely speculating based on my understanding of the model. In any case, it's not guaranteed that RIDs across different systems (or even clones) are identical (though in practice they normally would be identical across clones). -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 10:09 AM, Ron Aaron r...@ronware.org wrote: So, Ron, can you try: $fossil sqlite sqlite SELECT * FROM blob where rid=15513 it's blank (empty) for me. Blank as in the row does not exist, or blank in the sense that the row exists but contains no content? -- D. Richard Hipp d...@sqlite.org ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
Don't know, I'll have to try and see. I can't recall doing that, but it's certainly possible. On 10/04/2013 05:12 PM, Jan Nijtmans wrote: 2013/10/4 Ron Aaron r...@ronware.org: Indeed, it gives an empty string. Then I have a theory what happened. My guess is that you cloned the repository and opened it. Later you cloned the repository again (which got different row-id's), storing it in the exact original location. Then the row-id's of the .fslckout database and the repository don't match any more. What you can do to fix this: - fossil update trunk (find the new rowid of trunk, and bring the checkout there) - fossil revert (clean up any garbage) then do a make and everything else, and try the commit. Am I close? Regards, Jan Nijtmans ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users -- For confidential messages, please use my GnuPG key http://ronware.org/gpg_key.html ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 4:21 PM, Ron Aaron r...@ronware.org wrote: Well, boy howdy! That seems to have cured it. :-D Somehow I must have done Something Bad, but not clear just what it may have been... Jan's hypothesis sounds better than anything i can come up with. It was, in any case, a mismatch between expected and available RID values. The RID values are transient, local to a specific copy of a repo. Jan's idea of a re-clone would explain the difference there (and i can't think of any other cause for it, other than restoring from backup or similar, but the end effect would be the same as a re-clone). -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
Well, boy howdy! That seems to have cured it. Somehow I must have done Something Bad, but not clear just what it may have been... Thanks to all of you for your help Ron On 10/04/2013 05:12 PM, Stephan Beal wrote: On Fri, Oct 4, 2013 at 4:07 PM, Ron Aaron r...@ronware.org mailto:r...@ronware.org wrote: On 10/04/2013 05:02 PM, Stephan Beal wrote: echo select rid from blob where uuid='6c0d3704e381430edaef8424b1c1d8d059325a65'; | fossil sqlite For me it's 10798 Somehow the record of your checkout version and the actual checkout seem to have gotten out of sync. i can't currently speculate as to how that might happen. Rather than fiddle without the checkout db, i'd suggest trying the following: fossil close --force cd .. tar cf foo.tar $OLDPWD cd - fossil open /path/to/repo.fsl --keep cd - tar xf foo.tar # probably not necessary with --keep, but can't hurt cd - fossil status should show you your list of changes (several hundred, IIRC). And then do your checkin again. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users -- For confidential messages, please use my GnuPG key http://ronware.org/gpg_key.html ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 4:25 PM, Stephan Beal sgb...@googlemail.com wrote: On Fri, Oct 4, 2013 at 4:21 PM, Ron Aaron r...@ronware.org wrote: Well, boy howdy! That seems to have cured it. :-D Somehow I must have done Something Bad, but not clear just what it may have been... Jan's hypothesis sounds better than anything i can come up with. It was, in any case, a mismatch between expected and available RID values. The RID values are transient, local to a specific copy of a repo. You did a rebuild in the middle, too, which might have further changed the RIDs (they get regenerated during a rebuild, and not necessarily in the same order as before). The rebuild didn't _cause_ this, but might have exacerbated it (why is it that i can spell exacerbated on the first try, but can't remember how many pairs are in occasionally without the help of a spell checker?). -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 4:29 PM, Stephan Beal sgb...@googlemail.com wrote: You did a rebuild in the middle, too, which might have further changed the RIDs (they get regenerated during a rebuild, and not necessarily in the same order as before). The rebuild didn't _cause_ this, but might have exacerbated it (why is it that i can spell exacerbated on the first try, but can't remember how many pairs are in occasionally without the help of a spell checker?). New theory! @Jan: please help me out here: fossil rebuild -R repo from outside the checkout could change the RIDs out from under any checkouts. OTOH, i would have expected to see this problem before now as a side effect of rebuild all. :-? (that's a question for Jan and Richard) -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 10:32 AM, Stephan Beal sgb...@googlemail.com wrote: New theory! @Jan: please help me out here: fossil rebuild -R repo from outside the checkout could change the RIDs out from under any checkouts. I don't think rebuild changes RIDs. The RIDs are set by the BLOB table and BLOB is not changed by a rebuild. -- D. Richard Hipp d...@sqlite.org ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
2013/10/4 Richard Hipp d...@sqlite.org: On Fri, Oct 4, 2013 at 10:32 AM, Stephan Beal sgb...@googlemail.com wrote: I don't think rebuild changes RIDs. The RIDs are set by the BLOB table and BLOB is not changed by a rebuild. I was about to answer the same: A clone is the only thing I know that changes RID's of Blobs. A rebuild doesn't do that. Regards, Jan Nijtmans ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 4:35 PM, Richard Hipp d...@sqlite.org wrote: I don't think rebuild changes RIDs. The RIDs are set by the BLOB table and BLOB is not changed by a rebuild. Good to know. i thought (haven't ported that far yet) that a rebuild reconstructed the whole blob table one artifact at a time. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] manifest file (12892) is malformed
On Fri, Oct 4, 2013 at 4:21 PM, Ron Aaron r...@ronware.org wrote: Well, boy howdy! That seems to have cured it. Next time it won't take so long: http://www.fossil-scm.org/index.html/info/7bf9fdb068 that adds a check for this specific case and fails earlier in the process. -- - stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] Run fossil behind stunnel?
On Fri, Oct 4, 2013 at 9:43 AM, Francis Daly fran...@daoine.org wrote: On Fri, Oct 04, 2013 at 02:58:09PM +0200, Jakob Eriksson wrote: How do I run fossil behind stunnel? [...] Certainly, putting SSL on Fossil is NOT as simple as just drop it behind stunnel as implied on the web page. Could you provide a link to the web page? Perhaps someone on the list will have access to edit it to be more clear or correct, for the next person. Just looked myself - The Fossil as an inetd/xinetd or stunnel service section on http://fossil-scm.org/index.html/doc/tip/www/server.wiki There is nothing in that section about the --https option to fossil. ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] Run fossil behind stunnel?
I remember having this exact problem in the past. The documentation should be improved. On Fri, Oct 4, 2013 at 6:06 PM, Ron Wilson ronw.m...@gmail.com wrote: On Fri, Oct 4, 2013 at 9:43 AM, Francis Daly fran...@daoine.org wrote: On Fri, Oct 04, 2013 at 02:58:09PM +0200, Jakob Eriksson wrote: How do I run fossil behind stunnel? [...] Certainly, putting SSL on Fossil is NOT as simple as just drop it behind stunnel as implied on the web page. Could you provide a link to the web page? Perhaps someone on the list will have access to edit it to be more clear or correct, for the next person. Just looked myself - The Fossil as an inetd/xinetd or stunnel service section on http://fossil-scm.org/index.html/doc/tip/www/server.wiki There is nothing in that section about the --https option to fossil. ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] Run fossil behind stunnel?
On Fri, Oct 4, 2013 at 12:06 PM, Ron Wilson ronw.m...@gmail.com wrote: Just looked myself - The Fossil as an inetd/xinetd or stunnel service section on http://fossil-scm.org/index.html/doc/tip/www/server.wiki There is nothing in that section about the --https option to fossil. I just checked in a fix for this. -- D. Richard Hipp d...@sqlite.org ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
Re: [fossil-users] Run fossil behind stunnel?
Thanks, what a turnaround time! Richard Hipp d...@sqlite.org skrev: On Fri, Oct 4, 2013 at 12:06 PM, Ron Wilson ronw.m...@gmail.com wrote: Just looked myself - The Fossil as an inetd/xinetd or stunnel service section on http://fossil-scm.org/index.html/doc/tip/www/server.wiki There is nothing in that section about the --https option to fossil. I just checked in a fix for this. -- D. Richard Hipp d...@sqlite.org ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users ___ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
