subprocesses and reads files, in syscall
terms that might count as a lot of work so perhaps a subpool is
justified.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
is small and read it all into memory.
+
+ if (eof)
+break;
+}
+
+ return NULL;
+}
+
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
and python 2.7.3.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
plausible.
Here is the patch for this part of the issue. Log message:
I have already committed something, but feel free to commit an
additional change if you think it better.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
differences intentional?
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
ignore_enoent is set?
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
: svn_test_main (svn_test_main.c:927)
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
: svn_test_main (svn_test_main.c:949)
==21219==by 0x4011A9: main (prefix-string-test.c:154)
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
)
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Philip Martin philip.mar...@wandisco.com writes:
+ /* Any previously cached children now have a new WCROOT. */
+ for (hi = apr_hash_first(scratch_pool, db-dir_data);
+ hi;
+ hi = apr_hash_next(hi))
+{
+ const char *abspath = apr_hash_this_key(hi
if I should remove the entries
rather than update them. Is it wrong for db-dir_cache to have entries
for paths that are not directories and do not exist?
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
, see r1651759.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
such a test?
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Philip Martin philip.mar...@wandisco.com writes:
$ valgrind -q test/test_all
==11901== Invalid read of size 8
==11901==at 0x403E7CF: serf_config_get_object (config_store.c:278)
==11901==by 0x403F88A: serf__log (logging.c:146)
==11901==by 0x4048E79: serf_log_wrapped_readline
Lieven Govaerts l...@mobsol.be writes:
On Mon, Jan 12, 2015 at 12:22 PM, Philip Martin
I'm not familiar with serf's testsuite. The scenario:
- client sends first request
- server sends response
- server sends extra data, say '\n'
- client handles first request
- client creates
Philip Martin philip.mar...@wandisco.com writes:
Lieven Govaerts l...@mobsol.be writes:
Is there a reason you specifically mention pipelined here? The first
response has been handled so it's just synchronous communication.
The point is that the client reads data from the server before
Philip Martin philip.mar...@wandisco.com writes:
Philip Martin philip.mar...@wandisco.com writes:
While debugging a problem reported on users I accidentally sent an extra
byte to the client: I sent Content-Length of N and then sent N+1 bytes.
The first N bytes made a valid response, so serf
Philip Martin philip.mar...@wandisco.com writes:
While debugging a problem reported on users I accidentally sent an extra
byte to the client: I sent Content-Length of N and then sent N+1 bytes.
The first N bytes made a valid response, so serf was happy at that
stage. When processing the next
;
+}
+res = apr_date_checkmask(ctx-linebuf.line, HTTP/#.# ### *);
if (!res) {
/* Not an HTTP response? Well, at least we won't understand it. */
return SERF_ERROR_BAD_HTTP_RESPONSE;
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Philip Martin philip.mar...@wandisco.com writes:
Index: buckets/response_buckets.c
===
--- buckets/response_buckets.c(revision 2445)
+++ buckets/response_buckets.c(working copy)
@@ -129,7 +129,10 @@ static
Looking at serf_context_run() I see
while (num--) {
serf_connection_t *conn = desc-client_data;
status = serf_event_trigger(ctx, conn, desc);
if (status) {
return status;
}
desc++;
}
When I look at conn[0] gdb it is junk. Looking at
()
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Ivan Zhakov i...@visualsvn.com writes:
On 2 January 2015 at 18:59, Philip Martin philip.mar...@wandisco.com wrote:
Ivan Zhakov i...@visualsvn.com writes:
Do I understand correctly that you have merged branch without
resolving concerns about dirent_t structure naming raised dev@ list
two
.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Philip Martin philip.mar...@wandisco.com writes:
My mistake, I was looking at libsvn_fs_fs/fs.h in my tree and saw the
rename, but it occurred later in r1649062.
Should be looking at libsvn_fs_x/fs.h
--
Philip
completed successfully?
It seems to me we should. What do you think?
Yes, if only to alert people that the apache/MPM they are using may have
bugs.
[...]
or if searching the error_log is the only way to detect these crashes.
I suspect searching the error_log is the only way.
--
Philip Martin
we allow the
build and XFAIL? At present we do all three.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
NmUo2ZevecgKR4ZH5bQILX5yNVOSfu8EPgyNfYkvBO/vJKJ5Y95P2NDrb4Pb0auH
PIxcDoySPT2hbKhmi20wqjABHSD4W0F5fyFKuq1LjPtqeQQtFNc3bVlPoUtbpR4=
=H+zc
-END PGP SIGNATURE-
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
return NULL,
i.e. a copy with no if header, and skip the walk.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
/jubif0vBJ4ER7X2lDNfAgdbg6BYWy2
CI58JhLhILDjjjy4lQINPIU5X/FpdcoGwfCDhErsTunbmxBX9fx+rnLqtyQxpJ9W
NMKwHb3hShSWaztoj1VH33HjhwkjK6AU/QcQS5nFc9+i+y6mAaJ7/Whwcm38yErk
+2X7fJcaB1TRPUMuc46osBUzw1lh7j2NBMBTu1qYL2kDslK+yBzdPi566OLvJ9U=
=WDzy
-END PGP SIGNATURE-
--
Philip Martin | Subversion Committer
that breaks.
configure and libtool both unset CDPATH as safety measure.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
? Perhaps the
large tree produced by the script in the issue is not large enough to
cause the problem in a normal build?
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
will deactivate the new caching
facilities and cause the server to fall back to 1.6 caching
mechanisms.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Philip Martin philip.mar...@wandisco.com writes:
I'm still trying to determine why this
extra walk is present, perhaps the new version is doing an unnecessary
walk or perhaps the old version has an unfixed bug.
Debian's httpd doesn't have r1497441 which introduced the extra walk for
copies
Philip Martin philip.mar...@wandisco.com writes:
I think this extra walk is pointless most of the time and is just making
our nominally O(1) copy operation slower. Locks on the copy source do
not prevent a copy.
Oops! I meant to write Locks on the copy source held by somebody else
do
Philip Martin philip.mar...@wandisco.com writes:
My pool-debug build uses my own httpd build from 2.2.x@1562432 while my
normal build used the system httpd, Debian's 2.2.22-13+deb7u3. It looks
like some change to mod_dav has added an extra walk over the copy source
in the pool-debug build
mistakenly passed to
svn_ra_do_switch3() by 'switch_loc-url'.
]]]
I fixed the testsuite fail that resulted and committed
r1641564. Backports nominated for 1.8 and 1.7. Thanks!
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Description: compressed dump file
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
(CCFLAGS=['/W4', '/wd4100'])
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
because the
distribution's libtool resolves symbols against the old installed
libraries rather than the new build libraries. If this happens then
moving or removing the installed libraries will allow the build to work.
I've not experienced this error for a while.
--
Philip Martin | Subversion
Mark Phippard markp...@gmail.com writes:
I might be able to try this as I would like to see it get backported.
It merges cleanly into 1.8 if you want to build 1.8 rather than trunk.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
. With current 1.8.x
I saw failures, including a null pointer exception. With r1640730
backported to 1.8.x I can use the keyring successfully.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Mark Phippard markp...@gmail.com writes:
If you nominate this for backport you can include my +1 vote.
I've nominated it for 1.8 and 1.7.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
printf
Patch by: Oleksandr Chumachenko led...@gmail.com
]]]
Yes, approved.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
in an editor.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
diagnosed in the issue.
It's possible that upgrading Subversion, or the APIs used by Subclipse,
might fix the problem.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
the minimum to connect and authenticate. One option would be
to extend that program and produce a testcase that reproduces the
problem. Perhaps it needs several connections? Or multiple
authentication objects? Or multiple threads? Or ...
--
Philip Martin | Subversion Committer
WANdisco
,
callback_done, key_info, NULL);
g_main_loop_run(key_info.loop);
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Philip Martin philip.mar...@wandisco.com writes:
Philip Martin phi...@codematters.co.uk writes:
Philip Martin philip.mar...@wandisco.com writes:
handle_fetch (request=0x73e56038, response=0x73e918b8,
handler_baton=0x73e611e0, pool=0x73e6d028)
at ../src/subversion
but not for
make check BASE_URL=http://...
Perhaps we could move the code from davautocheck.sh to apache.m4. Have
apache.m4 set a make variable containing the version number and add a
make rule to export the value of the make variable in an environment
variable.
--
Philip Martin | Subversion Committer
Philip Martin philip.mar...@wandisco.com writes:
Perhaps we could move the code from davautocheck.sh to apache.m4. Have
apache.m4 set a make variable containing the version number and add a
make rule to export the value of the make variable in an environment
variable.
I did it slightly
Branko Čibej br...@wandisco.com writes:
On 14.11.2014 14:47, Philip Martin wrote:
Philip Martin philip.mar...@wandisco.com writes:
Perhaps we could move the code from davautocheck.sh to apache.m4. Have
apache.m4 set a make variable containing the version number and add a
make rule
.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
-trivial.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
be sufficient to make the output order deterministic.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Justin Erenkrantz jus...@erenkrantz.com writes:
On Wed, Nov 5, 2014 at 9:24 AM, Philip Martin phi...@codematters.co.uk
wrote:
Configure a server with both mod_ssl (HTTPS) and mod_deflate (HTTP
compression). Start a checkout and the kill the server. The client
goes into an infinite loop
Philip Martin philip.mar...@wandisco.com writes:
handle_fetch (request=0x73e56038, response=0x73e918b8,
handler_baton=0x73e611e0, pool=0x73e6d028)
at ../src/subversion/libsvn_ra_serf/update.c:1152
1152if (SERF_BUCKET_READ_ERROR(status))
(gdb) p status
$13
Philip Martin phi...@codematters.co.uk writes:
Philip Martin philip.mar...@wandisco.com writes:
handle_fetch (request=0x73e56038, response=0x73e918b8,
handler_baton=0x73e611e0, pool=0x73e6d028)
at ../src/subversion/libsvn_ra_serf/update.c:1152
1152
Configure a server with both mod_ssl (HTTPS) and mod_deflate (HTTP
compression). Start a checkout and the kill the server. The client
goes into an infinite loop in libsvn_ra_serf/update.c:handle_fetch
(gdb) n
1147 svn_txdelta_window_t delta_window = { 0 };
(gdb) n
1151 status
);
+ dwb-entry-size = 0;
+}
}
}
else if (strcmp(name, resourcetype) == 0)
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Philip Martin philip.mar...@wandisco.com writes:
If the
Solaris port has defined SVN_FILESIZE_T_FMT as %ld, rather than ld,
the resulting format string would be %%ld and would lead to the ouput
shown above. The difficulty here is that breaking the DAV code like
that would also break
Branko Čibej br...@wandisco.com writes:
Good catch. The array should be const, too.
There is a struct declaration that should be moved as well, r1636388.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Philip Martin philip.mar...@wandisco.com writes:
The error below is caused by running the JavaHL regression tests with
non-working SQLite support.
Moving this bit of the discussion to dev. The SQLite rep-cache is
designed to be optional. Force it to fail:
Index: ../src/subversion
if asn1_get_tag were to detect that *p had reached end it would
happen after we had dereferenced **p. This bug occurs in other places:
x509_get_sig, x509_get_alg, etc.
Fix by assigning ASN1_OID explicitly or by moving the assignment after
asn1_get_tag.
--
Philip Martin | Subversion Committer
WANdisco // *Non
Philip Martin philip.mar...@wandisco.com writes:
Sergey Azarkevich azarkev...@gmail.com writes:
I see in pre-commit hook template that changing revisions properties in
transaction are allowed.
But I can't find way for it (I want change svn:author to more readable).
Is here way
Alexey Neyman sti...@att.net writes:
On Wednesday, October 08, 2014 07:56:50 AM Daniel Shahaf wrote:
You said you had two approaches, but you only list one?
Err, was the email truncated in your mailbox somehow?
The mail I received was also truncated.
--
Philip Martin | Subversion
the corresponding
session pool?
Yes, and that a pattern is also used in other places, e.g. svn_fs_t.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
and it fixes the compressed checkout.
Are there any plans for a serf release that includes this fix?
-
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
response 4GB, such as a bulk-mode
REPORT carrying a whole Subversion tree that's larger than 4GB?
I can trigger the decompression error on a 5GB REPORT by setting
http-bulk-updates=yes on the client side.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Philip Martin philip.mar...@wandisco.com writes:
C. Michael Pilato cmpil...@gmail.com writes:
The log message for r2419 mentions files larger than 4Gb, and leads me
to believe that this problem only affects GETs. But here, Philip avoids
the term files and talks about the compressed size
Mark Phippard markp...@gmail.com writes:
On Wed, Oct 1, 2014 at 10:03 AM, Philip Martin philip.mar...@wandisco.com
wrote:
I can trigger the decompression error on a 5GB REPORT by setting
http-bulk-updates=yes on the client side.
This does not really answer the question.
Was your REPORT
for it to do so.
I've committed a modification of the patch that also adds glibtoolize1
and libtool14 in this order:
glibtoolize libtoolize glibtoolize1 libtoolize15 libtoolize14
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Julian Foad julianf...@btopenworld.com writes:
Hi Philip. Sounds good, but it would be good to add your reasoning
(above) into the log message and/or as comments in the code.
Done in r1627325.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
but no change
for the property no-op -r4.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
open_file.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
-in-1.9 svnadmin delrevprop to delete revision
properties, so either svnlook propdel or svnadmin deltxnprop.
Or perhaps the new subcommand to set txnprops should allow a way to
delete as well?
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
propget already combines --revprop
with --transaction and we can error out if --transaction is combined
with --use-pre/post.
(And yes, 'svnlook' should be kept read-only.)
I prefer an svnadmin solution as well.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Ivan Zhakov i...@visualsvn.com writes:
On 12 August 2014 12:46, Philip Martin philip.mar...@wandisco.com wrote:
Philip Martin philip.mar...@wandisco.com writes:
The txn_dir_cache gets cleared when the pool associated with the
txn_root is cleared
The net result is that the txn_dir_cache
missing
from the compatibility function?
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Bert Huijben b...@qqmail.nl writes:
Philip Martin wrote:
../src/subversion/libsvn_wc/deprecated.c: In function
'svn_wc_get_diff_editor6':
../src/subversion/libsvn_wc/deprecated.c:1991:39: warning: parameter
'ignore_ancestry' set but not used [-Wunused-but-set-parameter]
It was added
it
to work? fs-test 44 is the new test I wrote: it FAILS for FSFS but is a
PASS for BDB and FSX. It is also a PASS for 1.6 FSFS.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Philip Martin philip.mar...@wandisco.com writes:
The txn_dir_cache gets cleared when the pool associated with the
txn_root is cleared
The net result is that the txn_dir_cache doesn't really work with
apache/mod_dav_svn. Import a directory containing a large number of
small files and the rate
/mod_dav_svn?
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Philip Martin philip.mar...@wandisco.com writes:
1. client sends MKCOL
2. apache process MKCOL and populates txn_dir_cache
3. apache sends MKCOL response
4. apache clears MKCOL request pool
5. client sends MERGE
6. apache process MERGE
5 can happen before 4 since two processes
on disk
makes it into the revision.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Philip Martin philip.mar...@wandisco.com writes:
svn_fs_txn_root() calls svn_fs_fs__initialize_txn_caches() and if that
detects concurrent transactions it sets ffd-txn_dir_cache to NULL so
the out-of-date cache doesn't get used and the property change on disk
makes it into the revision
+//xHeD7JCzXx43ChodFoqk4vQXJ/ijaCfsn05p
G83CR7J2DaaRZMEmSW0ShQWFcUN7W2c7AHktLirLtXU2L788Q8axg6kbLbjmaKMO
IVyuOzYDzl37WxeVCcV70YAxl7LCGb6sIG/GfA9Gnb6uQfZ8ErFI9WXhCa0up4sH
Kf97uQZizvog7ayVuUlmdPC93AORNGd0uXv4ilyq/AULH+9IOzakuF+ONOv2mIk=
=Uh2C
-END PGP SIGNATURE-
--
Philip Martin | Subversion Committer
/Ib0Cf5JhxZ1fJsIQUsceurDN5PSBeXGaY
nRQTzX9lGvpmeXoaEm/aygMjtZ/BMyyf+GO2VK/l5wRHj2nxotSOHHikw4IPfk6T
f9or9GhgANiCZS5q9hmsirxEZNOVQN+xvv+kntSTABWn8S3RApEaPt7BaqLDqyO3
s238QHvObk+h6VT3eX33tfu5T+7EZej0vwbsy0GD5U6ZAsBT293c4dn8khbB/Xo=
=HQwP
-END PGP SIGNATURE-
--
Philip Martin | Subversion Committer
the existing slave repository online but this requires doing
somthing to ensure that revprop changes made on the master during the
svnsync get to the copy.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
operations are quick on the same disk, so the repository is
only offline for a short period.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
sophisticated behaviour in future releases?
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
debugging enabled and 0x41 is the poison byte
written to memory before free.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Branko Čibej br...@wandisco.com writes:
On 25.06.2014 14:43, Philip Martin wrote:
Philip Martin philip.mar...@wandisco.com writes:
I'm getting a SEGV in the JVM and it varies from run to run, two
examples at the end. I tried r1603298 and r1603297 and those also SEGV.
Looks like I made
= false, m_cred = 0x7fffd8060c60}
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Branko Čibej br...@wandisco.com writes:
On 26.06.2014 13:04, Philip Martin wrote:
Yup, I've been looking at exactly this spot and I suspect that I need to
make a copy of the hash table into the correct pool.
(gdb) p ht
$1 = (apr_hash_t *) 0x7fffd8060c60
(gdb) p ht[0]
$2 = {pool
Branko Čibej br...@wandisco.com writes:
Should be fixed in r1605739.
Yes, the tests now pass on my box.
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
in start_thread (arg=optimized out)
at pthread_create.c:304
#22 0x775000ed in clone ()
at ../sysdeps/unix/sysv/linux/x86_64/clone.S:112
#23 0x in ?? ()
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Philip Martin philip.mar...@wandisco.com writes:
I'm getting a SEGV in the JVM and it varies from run to run, two
examples at the end. I tried r1603298 and r1603297 and those also SEGV.
Looks like I made a mistake when building the older versions. I am now
seeing the SEGV with r1603306
401 - 500 of 2536 matches
Mail list logo
