<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Recent changes to bugs</title><link>https://sourceforge.net/p/middle-man/bugs/</link><description>Recent changes to bugs</description><atom:link href="https://sourceforge.net/p/middle-man/bugs/feed.rss" rel="self"/><language>en</language><lastBuildDate>Wed, 03 Oct 2007 17:13:46 -0000</lastBuildDate><atom:link href="https://sourceforge.net/p/middle-man/bugs/feed.rss" rel="self" type="application/rss+xml"/><item><title>Can't compile under GCC 3+, patch included</title><link>https://sourceforge.net/p/middle-man/bugs/50/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;lvalue casts are illegal, &lt;/p&gt;
&lt;h1 id="index-mainc"&gt;Index: main.c&lt;/h1&gt;
&lt;p&gt;RCS file: /cvsroot/middle-man/middleman/src/main.c,v&lt;br /&gt;
retrieving revision 1.217&lt;br /&gt;
diff -u -r1.217 main.c&lt;br /&gt;
--- main.c      5 Aug 2004 02:01:16 -0000       1.217&lt;br /&gt;
+++ main.c      3 Oct 2007 17:13:12 -0000&lt;br /&gt;
@@ -482,7 +482,7 @@&lt;br /&gt;
#else&lt;br /&gt;
pcre_free = xfree;&lt;br /&gt;
// grr... argument to malloc is int on some systems, and long on others&lt;br /&gt;
-       (void *) pcre_malloc = (void *) xmalloc;&lt;br /&gt;
+       pcre_malloc = (void* (*)(size_t)) xmalloc;&lt;br /&gt;
#endif&lt;/p&gt;
&lt;p&gt;global-&amp;gt;dns_cache = hash_create(DNS_HASH_SIZE);&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Charlie Hedlin</dc:creator><pubDate>Wed, 03 Oct 2007 17:13:46 -0000</pubDate><guid isPermaLink="false">https://sourceforge.nete476af1dc906286e71dc00e21fef94fac05b2ca3</guid></item><item><title>Stack smashing with long urls from log code</title><link>https://sourceforge.net/p/middle-man/bugs/49/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;This is probably the cause of several bug reports about specific URLs that crash middleman, ie 888197.&lt;/p&gt;
&lt;p&gt;Here is the output of cvs diff -uN log.c:&lt;br /&gt;
Index: log.c&lt;br /&gt;
===================================================================&lt;br /&gt;
RCS file: /cvsroot/middle-man/middleman/src/log.c,v&lt;br /&gt;
retrieving revision 1.31&lt;br /&gt;
diff -u -r1.31 log.c&lt;br /&gt;
--- log.c       2 Apr 2004 14:31:37 -0000       1.31&lt;br /&gt;
+++ log.c       3 Oct 2007 14:38:03 -0000&lt;br /&gt;
@@ -145,6 +145,7 @@&lt;br /&gt;
logbuffer_add(global-&amp;gt;logbuffer, buf);&lt;/p&gt;
&lt;p&gt;if (*(ptr - 1) != '\n') {&lt;br /&gt;
+               ptr--;&lt;br /&gt;
*(ptr++) = '\n';&lt;br /&gt;
*ptr = '\0';&lt;br /&gt;
}&lt;br /&gt;
===================================================================&lt;/p&gt;
&lt;p&gt;The code before this can fill the buffer completely, so&lt;br /&gt;
adding the newline and null may overflow the buffer.&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Charlie Hedlin</dc:creator><pubDate>Wed, 03 Oct 2007 14:39:53 -0000</pubDate><guid isPermaLink="false">https://sourceforge.nete701310ed0c3d2f31d386422b358937f451d16ce</guid></item><item><title>Large memory leak</title><link>https://sourceforge.net/p/middle-man/bugs/48/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;Seems that middleman-2.0.1 have some large memory leak.&lt;br /&gt;
Right after startup:&lt;/p&gt;
&lt;p&gt;&lt;a href="http://mman/stats" rel="nofollow"&gt;http://mman/stats&lt;/a&gt;&lt;br /&gt;
memory resident 3.2 MB&lt;br /&gt;
memory size 17.6 MB&lt;/p&gt;
&lt;p&gt;I have done some browsing (well, half of day of ocassional &lt;br /&gt;
browsing) and numbers went to&lt;br /&gt;
memory resident 234.9 MB  &lt;br /&gt;
memory size 282.8 MB&lt;br /&gt;
bytes in    63.1 MB &lt;br /&gt;
bytes out   73.6 MB&lt;/p&gt;
&lt;p&gt;I have done wget -r [my site with about 5-10mb of files] -Y (proxy via &lt;br /&gt;
middleman)&lt;br /&gt;
about 6 times and the numbers went to&lt;/p&gt;
&lt;p&gt;memory resident 260.6 MB  &lt;br /&gt;
memory size 308.5 MB  &lt;br /&gt;
bytes in    89.8 MB &lt;br /&gt;
bytes out   107.8 MB&lt;/p&gt;
&lt;p&gt;Memory cache size is set to 5 MB, disk cache is off on my &lt;br /&gt;
configuration...&lt;/p&gt;
&lt;p&gt;Rest of configuration (filters, etc ... ) is quite standard.&lt;br /&gt;
I tried turning off my external filter (pipe, simple program in perl to &lt;br /&gt;
filter via perl regexps), but the memory consumption is growing too, &lt;br /&gt;
so the cause probably isn't here..&lt;/p&gt;
&lt;p&gt;I'm attaching my config.xml (zipped)&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Martin Petříček</dc:creator><pubDate>Fri, 03 Jun 2005 01:51:10 -0000</pubDate><guid isPermaLink="false">https://sourceforge.net41212f7d4a92e5236decad97cd5bd661c48c3f78</guid></item><item><title>url filtering problem</title><link>https://sourceforge.net/p/middle-man/bugs/47/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;When I add a new allow url filtering, middleman (v&lt;br /&gt;
2.0.1) ignore ip range and only match if "ip range" is&lt;br /&gt;
empy. I can't assign a filter to an ip.&lt;/p&gt;
&lt;p&gt;Default policy for url filtering is deny.&lt;/p&gt;
&lt;p&gt;Regards.&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Anonymous</dc:creator><pubDate>Wed, 26 Jan 2005 13:33:31 -0000</pubDate><guid isPermaLink="false">https://sourceforge.net4dfd70c89ee3b019a1be0aaf2e699d96964e4c88</guid></item><item><title>File gets decompressed</title><link>https://sourceforge.net/p/middle-man/bugs/46/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;If I download the following file using middleman:&lt;br /&gt;
&lt;a href="http://ftp.debian.org/debian/pool/main/n/ncompress/ncompress_4.2.4-15.diff.gz" rel="nofollow"&gt;http://ftp.debian.org/debian/pool/main/n/ncompress/ncompress_4.2.4-15.diff.gz&lt;/a&gt;&lt;br /&gt;
the file is decompressed.&lt;br /&gt;
I haven't seen this behavior on any other (similar)&lt;br /&gt;
file I downloaded with middleman.&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Stephan Rutten</dc:creator><pubDate>Mon, 24 Jan 2005 11:01:33 -0000</pubDate><guid isPermaLink="false">https://sourceforge.net334c8670a5053fc2b4f49d3c4966394e4e106a06</guid></item><item><title>cache problem with ftp (cvs-version of mman)</title><link>https://sourceforge.net/p/middle-man/bugs/45/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;client: mozilla-1.7.5&lt;br /&gt;
ftp-server: first noticed with a public ftpd --&lt;br /&gt;
reproduced with oftpd-0.3.7&lt;br /&gt;
running on localhost&lt;br /&gt;
mman: cvs-snapshot from sf.net as of yesterday&lt;/p&gt;
&lt;p&gt;fetching a file via ftp for the first time is ok, all&lt;br /&gt;
successive fetches with&lt;br /&gt;
mozilla of a file (that should be cached) fail.&lt;br /&gt;
if the "fresh.." url-command is added it works fine.&lt;/p&gt;
&lt;p&gt;i'll happily provide log-snipplets of both mman and&lt;br /&gt;
oftpd if needed.&lt;br /&gt;
contact: eckert.thomas@gmx.net&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Anonymous</dc:creator><pubDate>Sat, 08 Jan 2005 13:51:32 -0000</pubDate><guid isPermaLink="false">https://sourceforge.net527db06305d50e84676cfae1e606c930a095549a</guid></item><item><title>apt-get hangs after 1 file</title><link>https://sourceforge.net/p/middle-man/bugs/44/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;To reproduce this, make sure all the debs are in your&lt;br /&gt;
middleman cache but not in the apt cache.  The example&lt;br /&gt;
below uses iftop because it is small and consists of&lt;br /&gt;
two separate packages.&lt;/p&gt;
&lt;p&gt;apt-get install iftop   # load into mman cache, grep&lt;br /&gt;
the journal to verify&lt;br /&gt;
apt-get remove libpcap0.7   # prepare to re-install&lt;br /&gt;
cd /var/cache/apt/archives; rm libpcap* iftop*   #&lt;br /&gt;
force apt to fetch the debs from mman&lt;br /&gt;
tail -f /var/log/middleman/middleman.log   # watch what&lt;br /&gt;
happens&lt;br /&gt;
apt-get install iftop   # this will hang on the second&lt;br /&gt;
fetch (iftop)&lt;/p&gt;
&lt;p&gt;apt-get will fail to get iftop, close the connection,&lt;br /&gt;
retry, and succeed.  It seems like cache code is&lt;br /&gt;
confusing mman enough to prevent it from serving more&lt;br /&gt;
files via a keepalive connection.&lt;/p&gt;
&lt;p&gt;Hopefully you can reproduce the problem easily now.&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Joshua Pritikin</dc:creator><pubDate>Mon, 26 Jul 2004 15:34:43 -0000</pubDate><guid isPermaLink="false">https://sourceforge.netf8ab0fed2200d569ba9ad979d23d729c379608b2</guid></item><item><title>Very high load avg when opening more pages</title><link>https://sourceforge.net/p/middle-man/bugs/43/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;I noticed, that when I open a couple of pages at a&lt;br /&gt;
time, Middleman works *very* slowly (e.g., have to wait&lt;br /&gt;
5 minutes or so for a page to load), and system's load&lt;br /&gt;
avg is very high.&lt;/p&gt;
&lt;p&gt;Below - top showing mman's CPU consumption (cloes to&lt;br /&gt;
50%...).&lt;/p&gt;
&lt;p&gt;I'm using recent CVS.&lt;/p&gt;
&lt;p&gt;top - 11:44:13 up 9 days,  3:29,  1 user,  load&lt;br /&gt;
average: 4.77, 2.63, 1.11&lt;br /&gt;
Tasks: 103 total,   5 running,  98 sleeping,   0&lt;br /&gt;
stopped,   0 zombie&lt;br /&gt;
Cpu(s):  78.0% user,  22.0% system,   0.0% nice,   0.0%&lt;br /&gt;
idle&lt;br /&gt;
Mem:    514632k total,   508528k used,     6104k free,&lt;br /&gt;
8004k buffers&lt;br /&gt;
Swap:        0k total,        0k used,        0k free,&lt;br /&gt;
271824k cached&lt;/p&gt;
&lt;p&gt;PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM &lt;br /&gt;
TIME+  COMMAND&lt;br /&gt;
6438 root      20   0 29036  28m  16m R 46.2  5.6&lt;br /&gt;
0:43.11 mman&lt;br /&gt;
6440 root      15   0 29036  28m  16m R 46.2  5.6&lt;br /&gt;
0:43.24 mman&lt;br /&gt;
6526 root      15   0 29036  28m  16m R 46.2  5.6&lt;br /&gt;
0:42.04 mman&lt;br /&gt;
6662 root      15   0  1016 1016  788 R 46.2  0.2&lt;br /&gt;
0:00.63 top&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">mangoo</dc:creator><pubDate>Fri, 16 Jul 2004 09:55:05 -0000</pubDate><guid isPermaLink="false">https://sourceforge.net2e0c6c850dee1f0faf37315e00f0947665d928b8</guid></item><item><title>redirect -&gt; "bypass sourceforge mirror selection" broken</title><link>https://sourceforge.net/p/middle-man/bugs/42/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;Config -&amp;gt; redirect -&amp;gt; "bypass sourceforge mirror&lt;br /&gt;
selection" is broken.&lt;/p&gt;
&lt;p&gt;When it's enabled, I'm unable to access "Download -&amp;gt;&lt;br /&gt;
Stable" section on www.clamav.net (alert window pops&lt;br /&gt;
up, and says "Redirect limit for this URL exceedeed.&lt;br /&gt;
Unable to load the requested page. This may be caused&lt;br /&gt;
by cookies that are blocked.").&lt;/p&gt;
&lt;p&gt;When I disable Config -&amp;gt; redirect -&amp;gt; "bypass&lt;br /&gt;
sourceforge mirror selection", I can access the page&lt;br /&gt;
mentioned.&lt;/p&gt;
&lt;p&gt;I have cookies in my browser enabled, as well as in&lt;br /&gt;
Middleman.&lt;br /&gt;
I'm using CVS version (14.07.04).&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">mangoo</dc:creator><pubDate>Thu, 15 Jul 2004 09:37:02 -0000</pubDate><guid isPermaLink="false">https://sourceforge.net46e0652913f9f50cc2a7bf43a063d924c2a2f334</guid></item><item><title>Chunked Transfer in HTTP/1.0 with 1.9.1</title><link>https://sourceforge.net/p/middle-man/bugs/41/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;I just tracked down the funny hex numbers i see in dillo, and some other problems when using HTTP connections via MiddleMan 1.9.1 with some clients.&lt;/p&gt;
&lt;p&gt;Dillo uses HTTP/1.0, sends out the request in HTTP/1.0.&lt;br /&gt;
The answer from MiddleMan comes in HTTP/1.1 with chunked transfer encoding. Dillo can't handle this and display the hex numbers (chunk lengths) in the result.&lt;/p&gt;
&lt;p&gt;How to reproduce:&lt;br /&gt;
Use dillo and middleman 1.9.1, go to &lt;a href="http://www.google.com/" rel="nofollow"&gt;http://www.google.com/&lt;/a&gt; and search for something. You should see the first hex number in the upper left corner of the page.&lt;/p&gt;
&lt;p&gt;Didn't find a good place while scanning the sources for a fix for this problem, maybe someone knows the source better and knows where to drop in the necessary line of code :)&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Michael Kukat</dc:creator><pubDate>Fri, 07 May 2004 21:29:40 -0000</pubDate><guid isPermaLink="false">https://sourceforge.net5edf2f569cf81b1414d1bf51b526f2cebe8cb5fd</guid></item></channel></rss>