similar to: Too many open files, etc.

Hello, I just found this message in our dovecot logfiles: "Disconnected: BUG: Unknown internal error" Should I worry about this ? There are no complaints yet (although it is not limited to one user). It seems to occur with different clients (Apple/MacOS mail, Windows/Thunderbird, Pine). We're running dovecot 1.0.10 on Solaris 10 with mboxes (NFS) and local index storage. There are

Hello, I have a question about the way Dovecot limits file system access. Currently we're using Dovecot 1.0.5 (Solaris 10). In some cases users have a symbolic link like "z: -> /" in their mail directory. As a result there are log entries like Jan 25 13:30:31 imap1 dovecot: [ID 107833 mail.error] IMAP(xyz):

Hi, thank You for a reply, compiling with Solaris Studio 12.5 solved this case. Martin Preen <preen at informatik.uni-freiburg.de> ra??: > Hello, > I don't have problems building 2.2.27 on Solaris 10 > (using Sun Workshop compiler 5.11). > > The configuration is the same as your. > Maybe a compiler/version problem on your system ? > > Regards, > Martin >

Hello everybody, I get the following assertion error with dovecot 1.0-stable when accessing a large folder: file istream-raw-mbox.c: line 383 (istream_raw_mbox_get_body_size): assertion failed: (rstream->body_offset != (uoff_t)-1) There is no error when using dovecot 0.99.14. Any ideas ? Martin ----------------------------------------------------------------- Martin Preen, Universitaet

Hello, with some mbox folders I got this error: Mar 18 14:48:12 imap2 dovecot: [ID 107833 mail.crit] Panic: IMAP(xyz): file charset-iconv.c: line 122: unreached I assume that this is related to the folder contents (and not the configration). What does that mean to me ? Martin --------------------------------------------------------------- Martin Preen, Universit?t Freiburg, Institut f?r

Hi, My setup: Solaris 9, /var/mail is NFS mounted from another Solaris 9 box, home directories and dovecot index stuff is on local ufs filesystems. I upgraded from alpha4 to a5 about 36 hours ago. Dovecot is configured like so during the build: VERSION=1.0.alpha5 CC=gcc CFLAGS="-g -O" CPPFLAGS=-I/opt/openssl/include LDFLAGS=-L/opt/openssl/lib \ ./configure

Hello, I am tying to compile Dovecot 2.2.27 on Solaris 10, and I get this error: test-ioloop.c: In function `test_ioloop_pending_io': test-ioloop.c:188: error: size of array `type name' is negative My configuration is like this: Install prefix . : /usr/local File offsets ... : 64bit I/O polling .... : poll I/O notifys .... : none SSL ............ : yes (OpenSSL) GSSAPI ......... : no

Hello, I'm currently testing dovecot/imap (Version 0.99.13). I found this in the log: imap(xxxxx): [ID 265587 mail.crit] file imap-bodystructure.c: line 278 (part_parse_headers): assertion failed: (part->physical_pos >= input->v_offset - start_offset) dovecot: [ID 684838 mail.error] child 7376 (imap) killed with signal 6 The user tried to setup a new folder (with Thunderbird)

Hello, I'm testing version 1.1.12 and occasionally there is Mar 18 14:40:44 imap2 dovecot: [ID 107833 mail.error] IMAP(xyz): nfs_flush_file_handle_cache_dir: rmdir(/home/xyz) failed: Device busy It seems that it is related to a crash of the previous imap process of that user. And it doesn't seem to cause any problems. So can it safely be ignored ? It is on Solaris 10. Index storage

Hello, currently we're using version 1.0.13 with 32bit file offsets. Is it safe to switch to a new version with largefile support enabled ? We want to reuse existing index/cache or do we have to expect errors with that ? Regards. Martin --------------------------------------------------------------- Martin Preen, Universit?t Freiburg, Institut f?r Informatik Georges-Koehler-Allee 52, Raum

My mailhosting provider has been testing dovecot 1.0 beta 3 as a proxy server and they asked the customers if they would like to test it out and report any issues. I am using Thunderbird 1.5 (IMAP) and have been with this mailhosting company for two months and have no issues at all with their current setup which uses Perdition as their proxy servers. The problem(s) I am encountered have been

Hello, I tried to find out how about to use the hardware crypto engines under Solaris (Sun Fire T2000). It seems, that its not just a compilation issue: > For operations that are to be offloaded, it is necessary to restrict use to subset > of OpenSSL functions (the EVP_ functions) and explicitly indicate the use of the PKCS11 > engine; something like the following works for bulk

Hello, > I upgraded from alpha3 to alpha4 yesterday, and I am seeing > a few of the following asserts: > > imap(user): file mbox-lock.c: line 379 (mbox_lock_fcntl): > assertion failed: (lock_type == F_UNLCK) I'm getting this error too after upgrading to alpha4. > Gdb info on a couple of the cores is attached. It looks > like the arg lock_type for mbox_lock_fcntl()

Hello, I don't have problems building 2.2.27 on Solaris 10 (using Sun Workshop compiler 5.11). The configuration is the same as your. Maybe a compiler/version problem on your system ? Regards, Martin Mantas Gegu?is wrote: > Hello, > > I am tying to compile Dovecot 2.2.27 on Solaris 10, and I get this error: > test-ioloop.c: In function `test_ioloop_pending_io': >

Hello, is someone using Dovecot with Solaris on a T1 cpu ? I would like to hear about pros and cons. Martin --------------------------------------------------------------- Martin Preen, Universitaet Freiburg, Institut fuer Informatik Georges-Koehler-Allee 52, Raum 00-006, 79110 Freiburg, Germany phone: ++49 761 203-8250 preen at informatik.uni-freiburg.de fax: ++49 761 203-8242

http://dovecot.org/test/ Two potentially dangerous changes: - mbox parser should now be able to handle CR+LF linefeeds. Hopefully doesn't break anything. - Some mail header parser changes. Hopefully doesn't break anything. And other stuff: - COPY shouldn't crash anymore with 64bit systems - If PAM modifies username, keep it - When auth workers died they leaked file

Hallo, I have a question concerning SVM regression in R. I intend to use SVMs for feature selection (and knowledge discovery). For this purpose I will need to extract the weights that are associated with my features. I understand from a previous thread on SVM classification, that predictive models can be derived from SVs, coefficiants and rhos, but it is unclear for me how to transfer this

On 2017.02.03. 16:10, Mantas Gegu?is wrote: > Hi, > > thank You for a reply, compiling with Solaris Studio 12.5 solved this case. > > > Martin Preen <preen at informatik.uni-freiburg.de> ra??: > >> Hello, >> I don't have problems building 2.2.27 on Solaris 10 >> (using Sun Workshop compiler 5.11). >> >> The configuration is the same as

Hello, We are running units in the field that are headless. Sometimes we get units returned that we when we boot them up have some type of filesystem inconsistency that the default preen doesn't fix but running fsck -y does. I want to eliminate the -p (preen option) and always do the -y option anyone know where to make this change? Thanks, -- Stephen Clark *NetWolves* Director of

Dear all, I just stumbled upon the fact, that when I perform a regression on multivariate responses, that are not centred, I get a ricilulously high R-squared value. After reading the code of summary.lm, I found a bug in the function summary.lm: mss is calculated by: mss <-sum((f - mean(f))^2) - where f are the fitted values. This works only for a single response variable, because