> Hm, darcs's patch viewer isn't much use: is there a way to make `v'
> display unified diffs with any context?
No, unfortunately.
What I do is apply the patch, then do something like
darcs diff -u --patch=Refactor
Thanks for the patch, applied and pushed.
> ... oh, is *that* what internAtom
Dear all,
Since it's after midnight over here, I'm officially allowed another
Polipo snapshot.
You will find the latest snapshot on
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-20071002.tar.gz
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-20071002.tar.gz.asc
Th
> The site I was getting crashes most often was this one:
> http://www.rbc.ru/ Usually only several reloads were enough to pull
> a crash.
I still cannot reproduce the issue, but I have found what is weird
about this site: their DNS server is buggy, it sends ``Refused'' to
requests, which cau
Could some HTML specialist have a look at
http://localhost:8123/polipo/config?
and try to make this page a little more pleasing to the eye? You'll
find the source in local.c, function httpSpecialRequest.
Thanks for your help,
Juliusz
-
Dear all,
I am happy to announce version 1.0.3 of the Polipo HTTP proxy, which
you will find in
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-1.0.3.tar.gz
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-1.0.3.tar.gz.asc
For more information about Polipo, please see
>> It's a very rare bug, and fixing it would require keeping quite a bit
>> of additional housekeeping information; for once, I'll just give up.
> Ok. Thanks for looking into this. We'll work around it by not
> breaking the first connection until the second is complete. The
> overhead should be mi
In util.c, you will find the following code:
#if defined(__linux__)
...
#elif defined(__FreeBSD__)
...
#else
int
physicalMemory()
{
return -1;
}
#endif
This function is supposed to return the physical memory of the machine
we're running on, or -1 if it is infinite (i.e.
> I also included my contrib dir as well. I'm keeping my patches in Tor's
> svn for now at https://tor.eff.org/svn/trunk/contrib/polipo/.
Which reminds me that I forgot to include a config.tor.example, as you
suggested. Sorry for that.
Juliusz
--
> if I am not wrong, polipo (version 1.0.3, linux) should not save the
> data to its disk-cache for requests with 'Cache-Control: no-store' set
> -- but polipo seems to ignore it.
Yes, this is a bug all right. No-store is only implemented for
server replies, it should also be implemented for clie
[CC-ing the mailing list, by the author's permission.]
> But please add logging in acces_log in CLF (Common Log Format) like
> other proxies or web servers.
There's no fundamental reason why it isn't done, I just haven't
implemented it yet. Patches welcome.
> I need use webalizer to see info on
[Replying to the mailing list, by the author's permission]
> polipo is great.
>
> But I found bug.
>
> CONNECT method not working with parent proxy.
Yes, it's a bug. It really should be fixed.
Juliusz
-
[Replying to the list, by the author's permission.]
> One thing I noticed today - SSL (or other tunneled connections)
> are not filtered by forbiddenFile :(
> I could badly use it to fight SSL spam - many sites now use
> https to avoid caches and filtering proxies. Even some "good"
> sites are ne
> Is FTP protocol throught http proxy supported?
Please see the FAQ.
http://www.pps.jussieu.fr/~jch/software/polipo/faq.html
Juliusz
-
This SF.net email is sponsored by: Microsoft
D
>> I think polipo might have some serious problem with caching.
Everything Ziyad said is true. Additionally, I'd like to point out
that a serious bug with cache consistency was fixed in version 0.9.12;
if you are using any earlier version of Polipo, please consider
upgrading (I currently recommen
> I ran oprofile for 30 minutes on my laptop today. After visiting
> roughly 300 urls from my bookmarks file, I ran opreport on polipo. My
> polipo binary is built from a darcs pull this morning.
More or less what I expected. Early versions of Polipo spent 60% of
their time in the HTTP parser.
> this is my bugfix for CONNECT throught parentProxy bug
Could you please send a patch?
Juliusz
-
SF.Net email is sponsored by: The Future of Linux Business White Paper
from Novell. F
Sorry for letting this post through, I should be more careful when
moderating the list.
Michael, please do not send large tarballs to the list. If you wish
to discuss your code, you should send a suitable explanation, and
a patch, not your modified version of Polipo.
Forwarding by permission of the original author.
Juliusz
--- Begin Message ---
Hi,
2 patches implementing this.
The first does just the filtering, domains compared to domains
and regexes matched against the "domain:port" string.
Obviously regexes startin
> 2) polipo seems to send 2 DNS queries with the same id, when the 1st
> returns an error then the 2nd is not processed.
Correct. The A and queries are sent with the same id.
> IMHO polipo should wait for second answer or use different ids for
> A and query, so it does not return an er
> I remember seeing a post about configuring polipo for a large number of
> users way back in May but am not able to find it now.
Message id <[EMAIL PROTECTED]>,
dated Sun, 03 Sep 2006 22:17:17 +0200.
Message id <[EMAIL PROTECTED]>,
dated Sun, 16 Sep 2007 00:10:48 +0200
> I am trying to
> When I tried to do check what Polipo returns when i issue a connect
> command, it just echos the connect command like:
> CONNECT [Website.com] HTTP/1.1
I cannot reproduce this issue:
$ echo -ne 'CONNECT localhost:22 HTTP/1.1\n\n' | nc localhost 8123
HTTP/1.1 200 Tunnel established
SSH-
It appears that RFC 2616 is currently undergoing revision at the IETF.
The changes planned are mostly minor, but unfortunately include fixing
my favourite typo:
This media type UST NOT be used unless the sender knows that the
recipient can arse it;
(Section 4.4)
> Hallo, I'm using polipo 1.0.2 for mipsel on openwrt, it give me "segmentation
> fault" if diskCacheRoot or localDocumentRoot point to an existing directory,
> if
> I disable them or they point to an unexistent directory polipo starts without
> any problem.
I've just checked that 1.0.3 with a co
Tue Jan 1 23:08:03 CET 2008 Juliusz Chroboczek <[EMAIL PROTECTED]>
* Fix tunnelling with parent proxy.
Thanks to Richard Sputa.
Please test this version. This should fix the following:
- incorrect handling of CONNECT (https) with a parent proxy;
- incorrect handling of pipeline
> i'm using polipo 1.0.3 on openbsd.
> when trying to upload a picture to ebay polipo throws the following error:
>
> Couldn't write to server: Immediate shutdown requested
Could you please provide me with a syscall trace ? (Sorry, I don't
know how to generate one under OpenBSD -- on Linux it's s
Squid finally has IPv6 support.
http://squidproxy.wordpress.com/category/ipv6/
Juliusz
-
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio
> First: In my case, url regexp matching drowns out everything else (90%).
That I know. If you benchmark Polipo, don't use a forbidden file. If
you run a busy Polipo, don't put regexps in your forbidden file.
It might be possible to fix this inefficiency by using a DFA-based
rexexp matcher, but
>> The surprise is that the back-of-queue never triggers.
>> I'm not sure how to use this information, but it may be possible to
>> shave off 20-30% of the calls to timeval_cmp by somehow
>> special-casing this.
>
> I'll do that, then.
Hmm, no, actually I won't. It complicates the code quite a b
> For 2) I did not dig deep into polipo's source, so I'm not sure if my
> observation is right. The
> queries on wire I see
> 09:07:25.875312 IP (..) xxx.xxx.xxx.xxx.62376 > xxx.xxx.xxx.xxx.53: [udp sum
> ok] 26981+ A? www.glb.intel.com. (35)
> 09:07:25.875602 IP (..) xxx.xxx.xxx.xxx.62376 > xxx
> what exactly is "this version"? i checked the download areas and the
> newest things I see are 1.0.3 and 2007-10-02.
The Darcs tree.
darcs get http://www.pps.jussieu.fr/~jch/software/repos/polipo
Please see
http://www.pps.jussieu.fr/~jch/software/polipo/faq.html#current
> Is there a pre-compiled package for operwrt? I use kamikaze 7.09. I do not
> have the chance to compire polipo for openwrt by myself.
I've just put a current build (1.0.3 with the tunnelling fixes) on
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-20080103-mipsel.gz
http://www
Ya Ziad,
> And yes, there's a way: Compile Polipo with debug symbols (by passing
> "-g" to the compiler, note that by default the compile process will use
> the debug symbols and you should not do anything more) and don't run
> "stip" on the resulting binary, then run Polipo like this:
> gdb
> here is the trace,
Thanks, that should be enough to understand the problem.
Just to confirm that I'm reading the issue right -- what happens if
you run Polipo with
expectContinue=false
?
Juliusz
--
Thomas,
Does the following patch work around the issue ? (It's not correct,
but if it works around the issue, I'll think up a proper solution.)
(Please reply to the list, except of course when sending attachments.
It's more convenient for me.)
Juliusz
dif
I've just committed a proper fix into the Darcs repository. Could you
please test it, I'd like to release 1.0.4?
I'm also attaching a patch, but I'd much rather people tested the code
in the repo (there are some other tweaks).
Thanks,
Juliusz
diff -rN -u
Dear all,
The first release candidate for Polipo 1.0.4 is available on
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-20080106.tar.gz
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-20080106.tar.gz.asc
This release fixes two bugs, there are no user-visible changes.
> I sent you a back trace together with a script to reproduce it
> on 2007-12-02, but so far I didn't hear back from you.
I do indeed have your mail. Sorry for forgetting about it.
Juliusz
--
> A single client request is still enough to trigger:
>
> [EMAIL PROTECTED] ~/test/polipo/polipo-20080106 $./polipo
> proxyAddress=127.0.0.1
> Disabling disk cache: No such file or directory
> Disabling local tree: No such file or directory
> Established listening socket on port 8123.
> Assertion
Dear all,
The second release candidate for Polipo 1.0.4 is available from
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-20080106-2.tar.gz
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-20080106-2.tar.gz.asc
This is a bug fix only release, there are no user-visible
Dear all,
Version 1.0.4 of Polipo, the caching web proxy, is available from
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-1.0.4.tar.gz
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-1.0.4.tar.gz.asc
For more information about Polipo, please see
http://www.pps.ju
I've just put an OpenWRT (mipsel) binary of Polipo 1.0.4 on
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-1.0.4-mipsel.gz
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-1.0.4-mipsel.gz.asc
For information, Polipo is included in OpenWRT 7 (Kamikaze), but it is
not b
An experimental native Windows binary for Polipo 1.0.4 is available from
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-1.0.4-win32.zip
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-1.0.4-win32.zip.asc
For more information about Polipo, please see
http://www.pps
> http://interloper.org/tmp/polipo/
Excellent. I've added a mention of the Windows installer on the
Polipo page.
Juliusz
-
Check out the new SourceForge.net Marketplace.
It's the best
> I've just installed polipo 104 on windows. I understood the below
> email to fix the case where you have a parent proxy and connect to a
> secure site. Is that right?
That's correct.
> Well, it still doesn't seem to work for me. When I request a secure
> site from polipo, the browser just sit
> I think it would be good to bundle some sample shrink-the-cache script
> along with polipo.
I'm not quite sure what you mean -- in what way does ``polipo -x'' not
suit your needs ?
http://www.pps.jussieu.fr/~jch/software/polipo/manual/Purging.html
Juli
> Am I doing something wrong that I don't see any logs for the secure
> request with level 0xFF?
I'm afraid that logging of tunnelling is not implemented.
Juliusz
-
Check out the new S
> which should be fine. If I start Tor and than Polipo everything is
> working. But if I start Polipo and than Tor it's not working.
Something else may have changed in the meantime, and Polipo may have
configured itself differently. Please go to
http://localhost:8123/polipo/config?
and check
> I also think the actual bug is on the server side.
Yes, it is. RFC 2616 Section 14.16.
But then, what do you expect from a server written in Python.
> +if(token_compare(buf, i, i + 5, "bytes")) {
> +i += 5;
> +i = skipWhitespace(buf, i);
> +} else {
> +do_log(L
> || Couldn't write to client : Unknown error
This means that error-handling is broken somewhere. If you're using
Windows, that's a known issue. (I need help with this.)
> || Couldn't write to client : Immediate shutdown requested
That's perfectly normal. It happens when you press ``stop''
>> > I also think the actual bug is on the server side.
>> Yes, it is. RFC 2616 Section 14.16.
>> But then, what do you expect from a server written in Python.
> You never programmed in Python, did you?
Yes, I have. It's a language that's great for writing quick hacks
(much better than Perl).
> This is probably causing the forbidden functionality to not exist in my
> win32 binaries.
Forbidden functionality is disabled by default on Windows. In order
to compile with the forbidden functionality, you need to make sure you
have a POSIX regex library, and compile with -DHAVE_REGEX:
ma
> Zenkuje. Pray tell, what is an admissible logLevel for debugging
> purposes ? I tried 0x7F but it did not seem like it showed anything
> more than the default.
Debugging code is compiled-out by default. You'll need to modify
log.h,
-#define LOGGING_MAX 0xFF
+#define LOGGING_MAX (~0)
and
> I hope Juliusz is listening and will provide the selective
> forwarding functionality we need for Polipo to compete as a serious
> replacement for Privoxy !
I am listening, but I've got very little time to work on Polipo this
year. So don't expect any new features before the summer, unless you
Forwarded by permission of the author.
Juliusz
--- Begin Message ---
Hello,
I'm unable to send an email to sourceforge (it bounces because my domain
doesn't have a "postmaster" address).
Here is a copy of the bug report:
Orig
> In supersede case why you called "delayedHttpClientNoticeRequest" in
> "httpClientGetHandler".
> In this supersede case you have create new object and that object will be
> replace to old object in httpClientGetHandler then why you call this
> delayedHttpClientNoticeRequest, why don't you g
> Thx for the reply. Is there a chance that the request would get resent
> again when httpClientGetHandler gets called with the new_object or do u do
> this just for validation with request parameters ?
Just the latter.
What are you trying to do?
Juliu
Sorry for the forward, apparently one of my co-moderators rejected
this posting by mistake.
Juliusz
-
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visua
Bis repetita. It will work this time.
Juliusz
--- Begin Message ---
Hi all!
Has anyone had trouble using Polipo on OS X?
It seems as if the allowedClients directive is being totally ignored by
the server at runtime.
I try connecting to it from a non-loca
> Is there a way to get polipo to recognize my hosts file ?
No.
> I kinda like how I am able to block ads and junk.
You may either use Polipo's built-in ad blocker, or use an external
redirector, or use a local DNS proxy. Sorry, I'm not planning to
parse /etc/hosts, although I might accept a pa
> 1) to change the page polipo gives when an URL matches the forbidden list (I'm
> in Italy and should be better have a page in Italian), how can I do this?
Check the configuration variable forbiddenUrl. Note that setting that
will cause forbidden pages to result in a 302 redirect, rather than
a
> Even better, use a local dns proxy. That way polipo won't block on name
> lookups.
Polipo will not block on name lookups, even if there's no local DNS
proxy. That's the very reason why Polipo has its own DNS resolver.
If you run Polipo with dnsUseGethostbyname set to true, Polipo will
block on
> Attached patch fixes the description.
Thanks, applied.
Juliusz
-
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.c
> I guess you should avoid the double-quotes entirely
> or quote each of them ("127.0.0.1", "134.157.168.57").
Fixed. Thanks.
Juliusz
-
This SF.net email is sponsored by: Microsoft
De
> I'm japanese developer for mac osx. I found polipo a week before
> last week. and I had very impressive from it.
Thanks for the kind words.
> but it is a bit difficult to setup and use for common users.
> I made polipo's GUI wrapper for osx.
> I published it here.
> http://drikin.com/dolipo/
> It would be helpful to post screenshots and source code of your GUI
> wrapper. I don't see any source code in the .zip file.
I didn't realise the source is not available. Kohichi-san, could you
please clarify whether your GUI is Free Software ?
Juliusz
Dear Nathan,
Sorry for the delay.
> I'm using Polipo to allow a network of users to search previously
> browsed traffic. Currently, traffic from multiple users is lumped
> together (users can't search for objects browsed by their user name).
> What would be the best/easiest way to implement a
> Can an instance of Polipo be set to listen on several IP adresses at the same
> time like, say, on both 127.0.0.1 and 10.0.0.1 (but not 0.0.0.0) ?
Right now, it cannot. I agree that this is a serious issue, one that
I should fix.
Juliusz
---
> I use Polipo with Tor, as such I want to disable diskCacheRoot (which
> seems to be the case anyway).
Congratulations, you have successfully disabled the on-disk cache.
There's nothing more to do. The four warnings that you get are
completely unrelated to the on-disk cache.
> -"DNS: couldn't o
> As you can see there, the Content-Type header is missing, when I use
> the proxy (2nd invocation of wget).
Weird -- I cannot reproduce your issue. I get:
Connecting to localhost|127.0.0.1|:8123... connected.
Proxy request sent, awaiting response...
HTTP/1.1 200 OK
Content-Length: 20293
> I noticed today that mingw_setnonblocking() doesn't use its second
> argument.
[...]
> I guess something like the below would make it behave as documented.
Could some Windows person confirm that?
Juliusz
--
> I'm trying to use Polipo to cache connections to reduce overhead in an
> application which is doing RPC via HTTP POSTs.
>
> I've got it to use the proxy, but I still only see one request per
> connection (despite the fact that it's keeping persistent connections
> open).
Polipo will always use a
>> Polipo will always use a fresh connection for a POST or PUT request;
>> this is to avoid a race condition that is intrinsic in the HTTP protocol.
>> An HTTP server is allowed to shut down a persistent connection at any
>> time. This means in particular that it may shut down the connection
>> b
>> The solution is to use a reliable protocol, not HTTP, for RPC.
>
> I'm trying to make a bad system (which already uses HTTP for RPC)
> better here. I have complete control over the HTTP server and know
> what its behaviour is likely to be, moreover I know that resubmitting
> the POST request wo
> I'm new to polipo.
Welcome.
> I was wondering where (if it exists) or how I can find
> stats on Cache hits/misses
Polipo does not keep this kind of statistics.
Actually, this kind of statistics don't make much sense for Polipo,
since there isn't a one-to-one correspondence between client and
> I guess you can have an entry in the "servers" list but it may not have any
> active connections to those servers, am I correct?
Correct. Look at the ``connections'' column.
> I am looking for a few different things:
> 1) Cache hit/misses. I guess it does not exist but perhaps you can point
Bhavesh, could you please quote properly.
>> "httpClientNoticeRequest. Note however that what makes this tricky is
>> that on a miss we will go through this function twice -- the first
>> time we'll get a miss or a revalidation, the second time a hit.
>> Note also that a hit does not mean that w
Dear Roy,
Would you be so kind as to specify precisely what features you're looking for?
Juliusz
-
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
Don't miss this ye
Gabriel Kerneis and myself have just finished a set of latency tests
for a set of web servers, including Polipo. You can see the results on
http://www.pps.jussieu.fr/~jch/software/cpc/bench/
Juliusz
--
[CC-ing the list, by permission from the author]
> Now the problem is that I do want to have a long term cache, but,
> I wouldn't like it to be on a file system (for various reasons, one
> being also efficiency).
> So my idea was to download the source code, and see if I can
> somehow make po
> what does Juliusz think of merging into polipo privoxy's best [privacy
> enhancing] features, possibly with the assistance of the privoxy development
> team?
Feel free to send patches, but be careful with licensing issues. No
promises, though, except that I will read your patches carefully.
> I am on dial-up. I would like to change the 'proxyName' within polipo's
> configuration file at intervals so to become less easily identifiable in
> general.
By default, proxyName is not sent to the server, so I think it's
pretty pointless.
proxyName is used for two purposes:
- it is display
On recent Linux systems, Polipo will no longer set the access time of
files in the on-disk cache. If you're relying on the atime for anything,
you'll want to set
diskCacheNoAtime = false
in your config file.
Juliusz
> I was wondering if anyone had done a polipo-patch to have an access
> log that would show which urls are being loaded by polipo, and for
> what reasons.
You can compile polipo with debugging, and enable D_CLIENT_REQ or
D_SERVER_REQ. This might or might not suit your needs.
> The current log on
--- Begin Message ---
Package: polipo
Version: 1.0.4-1
Denis,
You apply the following patch to Debian's polipo:
polipo-1.0.4/log.c
+@@ -80,7 +80,9 @@
+
+ if(logFile != NULL && logFile->length > 0) {
+ FILE *f;
++ mode_t mask = umask(026);
+ f = fopen(logFile->strin
> O_NOATIME only works if the kernel is 2.6.8 or above. Wouldn't it make
> sense to try again opening the file without that flag if the first
> open() fails?
It is my understanding that the flag is silently ignored on earlier
kernels.
However, according to the manual page,
EPERM The O_N
> I'm sorry Juliusz, but I have send you patch with ask to test it.
Yes, and I acknowledge that I didn't reply speedily enough. I was
generally feeling uneasy about your patch, but didn't have time to
think about why.
The recent OpenSSL stuff was what I needed to take the time to think
about it.
>> 1. You did apply a patch relating to security without an explicit ack
>> From upstream.
>> I hope it is clear from the recent OpenSSL debacle why this must not
>> be done.
> Sorry, but no, it isn't.
Julien, you're trolling. Among the DDs I know, you're one of the most
active in making sure y
> Maybe it's better to revert when you are do that in the upstream code?
> My patch is not dangerous so it's not necessary to do it as soon as
> possible.
Then please be so kind to include a large, friendly disclaimer in the
Debian-specific README.
Juliusz
> As an aside, in the specific instance of the OpenSSL patch, the
> maintainer *did* get an ack from upstream.
As far as I'm aware, he got a reply from a person he didn't know on
a mailing list he wasn't intimately familiar with. ``Yeah, I met
this friendly guy in the lift, he told me it's probab
> For log messages like "Unsupported Cache-Control directive x-gzip-ok --
> ignored." would it be possible to include the url from which this header was
> produced?
Will do.
> A global config option to include the url would be appreciated.
Yep.
> Pointers on where to do a possible patch for thi
> Juliusz, let's do it more logical? I think `proxyOffline = false', as
> kazaam suggesting, is better.
> What do you think?
I agree.
proxyOffline is designed for people who want to read files in the
cache when the origin server is not available. It can be useful at
times, but it can be confusin
test this
code if you're running on an exotic platform.
Regards,
Juliusz
Wed May 21 00:02:23 CEST 2008 Juliusz Chroboczek <[EMAIL PROTECTED]>
* Implement logFilePermissions.
diff -rN -u old-polipo/log.c new-polipo/log.c
--- old-polipo/log.c
> For log messages like "Unsupported Cache-Control directive x-gzip-ok --
> ignored." would it be possible to include the url from which this header was
> produced? A global config option to include the url would be appreciated.
> Pointers on where to do a possible patch for this would also be wel
> rc = writeHeaders(entry->fd, &entry->body_offset, object, NULL, 0);
> if(rc == -2) {
> rc = rewriteEntry(object);
> if(rc < 0) return 0;
> return 1;
> }
> As far as I can see, the writeHeaders() would not return a -2, it can only
> return -1 or something posi
issing data). Unless an emergency memory purge happens in
between, no data will be lost.
But since I'm the friendly type, here's a change especially for you:
Fri May 23 23:55:30 CEST 2008 Juliusz Chroboczek <[EMAIL PROTECTED]>
* Protect against EINTR in rewriteEntry.
Protects ag
Sat May 24 00:05:48 CEST 2008 Juliusz Chroboczek <[EMAIL PROTECTED]>
* Fix return value of writeHeaders.
This bug was introduced with the variable header size, and prevented
rewriteHeaders from ever happening. Thanks to Ming Fu.
J
At some point in history, I was critical of using generic database
implementations unthinkingly:
> I'd think about it twice. Given a good underlying filesystem
> (Reiserfs, XFS or a recent release of Ext3, if you happen to be
> running Linux), it should beat any generic DB implementation hands on
Dear all,
A first release candidate for Polipo 1.0.5 is available from
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-20080524.tar.gz
http://www.pps.jussieu.fr/~jch/software/files/polipo/polipo-20080524.tar.gz.asc
For more information about Polipo, the all-dancing, all-singing,
> The function do_daemonise() did not redirect stdin/stdout/stderr to proper
> files such as /dev/null when they are not needed. This cause assert to send
> error message to what ever socket or file descriptor that reuses the stderr
> (fd #2).
>
> This problem only happen when polipo is under stres
> Is there any document explains the difference about private and public
> objects in polipo?
A public object sits in the objectHashTable.
A private object is not in the global hash table, so no new references
to it can be created. Its reference count can only decrease, so it
will eventually be
1 - 100 of 521 matches
Mail list logo