While encountering problems updating protobuf to 3.1 I foolishly ended up removing the libprotbuf* files in /usr/lib/x86_64-linux-gnu
Fortunately Protoc --version now outputs "libprotoc 3.1.0", but I am now trying to get the libprotobuf.so files back. When trying to use things that link to protobuf like matplotlib.pyplot I get an error like
** (process:15246): WARNING **: Failed to load shared library 'libgdk-3.so.0' referenced by the typelib: libprotobuf-lite.so.9: cannot open shared object file: No such file or directory
I have been able to reinstall some of libprotobuf files but the .so files are still missing.
$ ll /usr/lib/x86_64-linux-gnu/libprotobuf*
-rw-r--r-- 1 root root 2.3M Oct 23 2015 /usr/lib/x86_64-linux-gnu/libprotobuf.a
-rw-r--r-- 1 root root 387K Oct 23 2015 /usr/lib/x86_64-linux-gnu/libprotobuf-lite.a
lrwxrwxrwx 1 root root 25 Oct 23 2015 /usr/lib/x86_64-linux-gnu/libprotobuf-lite.so -> libprotobuf-lite.so.9.0.1
lrwxrwxrwx 1 root root 20 Oct 23 2015 /usr/lib/x86_64-linux-gnu/libprotobuf.so -> libprotobuf.so.9.0.1
Anyone know a way to recover the protobuf.so.9.0.1 and libprotobuf-lite.so.9.0.1 files in ubuntu 16.04?
Build Protobuf 2.6.1 version. After building this version it will create libprotobuf.so.9.0.1 and libprotobuf-lite.so.9.0.1.
download the right protobuf version, build it yourself, and you will be just fine.
the original so and the so you are going to build should be the same, the later one will work just fine.
Related
I'm trying to edit a ruby file in my spacemacs editor.
When I open the ruby file I get this debugging information:
[rbenv] Using the currently activated ruby.
evil-line-move: End of buffer
Suspicious state from syntax checker ruby: Flycheck checker ruby returned 1, but its output contained no errors: ruby: /usr/lib/x86_64-linux-gnu/libcrypt.so.1: version `XCRYPT_2.0' not found (required by /home/sfernandez/.rbenv/versions/2.7.0/lib/libruby.so.2.7)
Try installing a more recent version of ruby, and please open a bug report if the issue persists in the latest release. Thanks!
I've tried the solution from this post:
https://ubuntuforums.org/showthread.php?t=2375927
I tried doing this after changing directory to /usr/lib/x86_64-linux-gnu
sudo mv libcrypt.so.1 libcrypt.so.1.old
After I executed this line I was locked out from doing any commands with sudo. And received an error: sudo: account validation failure, is your account locked. I was able to get sudo access again by just refresh installing my system with Pop!_OS, I guess moving the link for libcrypt.so.1 caused something bad to happen, so won't do that again.
I'm sure this all makes sense, but I feel like I'm missing something and that this shouldn't be hard to fix. Would appreciate any guidance on this.
If it helps here's my system info with neofetch
///////////// sfernandez#pop-os
///////////////////// -----------------
///////*767//////////////// OS: Pop!_OS 20.04 LTS x86_64
//////7676767676*////////////// Host: Inspiron 7559 1.3.1
/////76767//7676767////////////// Kernel: 5.4.0-7634-generic
/////767676///*76767/////////////// Uptime: 9 mins
///////767676///76767.///7676*/////// Packages: 1763 (dpkg), 20 (flatpak)
/////////767676//76767///767676//////// Shell: bash 5.0.16
//////////76767676767////76767///////// Resolution: 1920x1080, 1920x1080
///////////76767676//////7676////////// DE: GNOME
////////////,7676,///////767/////////// WM: Mutter
/////////////*7676///////76//////////// WM Theme: Pop
///////////////7676//////////////////// Theme: Pop [GTK2/3]
///////////////7676///767//////////// Icons: Pop [GTK2/3]
//////////////////////'//////////// Terminal: gnome-terminal
//////.7676767676767676767,////// CPU: Intel i5-6300HQ (4) # 3.200GHz
/////767676767676767676767///// GPU: Intel HD Graphics 530
/////////////////////////// GPU: NVIDIA GeForce GTX 960M
///////////////////// Memory: 2328MiB / 7828MiB
/////////////
After doing some more work on this I've learned a little about symbolic links. I know that libcrypt.so.1 is a symbolic link to libcrypt.so.1.1.0 after running ls -l | grep "libcrypt.so.1" I get the following:
lrwxrwxrwx 1 root root 39 Mar 10 12:24 libcrypt.so -> /lib/x86_64-linux-gnu/libcrypt.so.1.1.0
lrwxrwxrwx 1 root root 17 Mar 10 12:24 libcrypt.so.1 -> libcrypt.so.1.1.0
-rw-r--r-- 1 root root 202760 Mar 10 12:24 libcrypt.so.1.1.0
Running ls -l after changing directory to $HOME/.rbenv/versions/2.7.0/lib I get the following output:
lrwxrwxrwx 1 sfernandez sfernandez 16 Jun 27 20:25 libruby.so -> libruby.so.2.7.0
lrwxrwxrwx 1 sfernandez sfernandez 16 Jun 27 20:25 libruby.so.2.7 -> libruby.so.2.7.0
Now, I think the problem is that libruby.so.2.7 can reference the ruby code just fine, but it also needs access to whatever is referenced by libcrypt.so.1. Anyone got any ideas?
This question is still relevant please help
I have the same issue (pop_os 20.04 LTS, spacemacs 0.300.0#27.1).
The only way I managed to make it work was downgrading emacs to version 26.3.
[UPDATE]
So after about half a year I had to set up a new laptop and was faced with the same problem. After hours of debugging I've finally found my old answer and managed to fix the problem.
Again, the solution for pop_os is to install the app from the pop!_shop that is named Emacs (Terminal).
I am developing a OSX application in Objective-C and OpenCV-2.4.12 using Xcode.
I followed this installation video guide and this setup video guide to integrate the OpenCV. Basically in the second video, the author uses the OpenCV dynamic libraries (.dylib files).
This work perfectly in my mac, but in another macs without OpenCV installed the application gives an error and I can´t open and run it.
I have searched and I think the solution is to use static libraries (.a files). But I can´t figure out how to replace the dynamic libraries with static libraries.
Which static libraries I need to use? Where I find them? How I link them in Xcode?
My settings right now:
I added to the project the libopencv_core.2.4.12.dylib, libopencv_highgui.2.4.12.dylib, libopencv_imgproc.2.4.12.dylib.
In the Search Paths I have (like explained in the second video):
Framework Search Paths: /usr/local/lib
Header Search Paths: /usr/local/include
Library Search Paths: "OpenCVFolderPath"/build/lib
Thank you.
I tend to install OpenCV using homebrew and if I look at the options I can specify for OpenCV Version 3 like this:
brew options opencv3
I get this output:
--32-bit
Build 32-bit only
--c++11
Build using C++11 mode
--with-contrib
Build "extra" contributed modules
--with-cuda
Build with CUDA v7.0+ support
--with-examples
Install C and python examples (sources)
--with-ffmpeg
Build with ffmpeg support
...
...
--with-python3
Build with python3 support
--with-qt
Build the Qt4 backend to HighGUI
--with-qt5
Build the Qt5 backend to HighGUI
--with-quicktime
Use QuickTime for Video I/O instead of QTKit
--with-static <--- *** HERE ***
Build static libraries
--with-tbb
Enable parallel code in OpenCV using Intel TBB
--with-vtk
Build with vtk support
--without-eigen
Build without eigen support
--without-numpy
Use a numpy you've installed yourself instead of a Homebrew-packaged numpy
--without-opencl
Disable GPU code in OpenCV using OpenCL
--without-openexr
Build without openexr support
--without-python
Build without Python support
--without-test
Build without accuracy & performance tests
--HEAD
Install HEAD version
So, if I then go ahead and install like this:
brew install opencv3 --with-static
I end up with these static libraries which look like what you want:
ls -l /usr/local/Cellar/opencv3/3.1.0_3/lib/
-r--r--r-- 1 mark admin 1922864 21 Jul 14:23 libopencv_calib3d.a
-r--r--r-- 1 mark admin 3215896 21 Jul 14:23 libopencv_core.a
-r--r--r-- 1 mark admin 965336 21 Jul 14:23 libopencv_features2d.a
-r--r--r-- 1 mark admin 650552 21 Jul 14:23 libopencv_flann.a
-r--r--r-- 1 mark admin 60016 21 Jul 14:23 libopencv_highgui.a
-r--r--r-- 1 mark admin 378128 21 Jul 14:23 libopencv_imgcodecs.a
-r--r--r-- 1 mark admin 3946248 21 Jul 14:23 libopencv_imgproc.a
-r--r--r-- 1 mark admin 954296 21 Jul 14:23 libopencv_ml.a
-r--r--r-- 1 mark admin 542808 21 Jul 14:23 libopencv_objdetect.a
-r--r--r-- 1 mark admin 1078504 21 Jul 14:23 libopencv_photo.a
-r--r--r-- 1 mark admin 269248 21 Jul 14:23 libopencv_shape.a
-r--r--r-- 1 mark admin 694056 21 Jul 14:23 libopencv_stitching.a
-r--r--r-- 1 mark admin 193416 21 Jul 14:23 libopencv_superres.a
-r--r--r-- 1 mark admin 478776 21 Jul 14:23 libopencv_video.a
-r--r--r-- 1 mark admin 149040 21 Jul 14:23 libopencv_videoio.a
-r--r--r-- 1 mark admin 390288 21 Jul 14:23 libopencv_videostab.a
drwxr-xr-x 3 mark admin 102 21 Jul 14:23 pkgconfig
drwxr-xr-x 3 mark admin 102 21 Jul 14:23 python2.7
There doesn't appear to be an equivalent option with OpenCV2, though I am not sure why you would want to use that version.
I've tried to configure my IntelliJ 13.1.5 + Ruby 6.5.2 plugin to run RSpec tests.
When I go to Run -> Edit Configurations and check the checkbox "Run the script in context of the bundle (bundle exec)" in Bundler tab, I get error "Run configuration error: No Gemfile found". See attached screenshot.
My Gemfile is definitely present in the root of working directory. Look for ls -la output below:
drwxr-xr-x 27 lukasm staff 918 27 Oct 09:23 .
drwxr-xr-x 38 lukasm staff 1292 13 Oct 15:53 ..
-rw-r--r--# 1 lukasm staff 6148 20 Oct 11:12 .DS_Store
drwxr-xr-x 16 lukasm staff 544 27 Oct 15:50 .git
-rw-r--r-- 1 lukasm staff 334 7 Jul 12:14 .gitignore
drwxr-xr-x 13 lukasm staff 442 27 Oct 15:51 .idea
-rw-r--r-- 1 lukasm staff 203 29 Aug 11:02 .rubocop.yml
-rw-r--r-- 1 lukasm staff 4 7 Jul 12:14 .ruby-gemset
-rw-r--r-- 1 lukasm staff 6 8 Jul 10:51 .ruby-version
-rw-r--r-- 1 lukasm staff 704 20 Oct 18:30 CPS.iml
-rw-r--r-- 1 lukasm staff 396 3 Sep 15:13 Gemfile
-rw-r--r-- 1 lukasm staff 2919 16 Oct 15:04 Gemfile.lock
-rw-r--r-- 1 lukasm staff 3426 7 Jul 12:14 README.md
-rw-r--r-- 1 lukasm staff 7436 24 Oct 17:54 Rakefile
drwxr-xr-x 3 lukasm staff 102 27 Oct 11:42 config
-rw-r--r-- 1 lukasm staff 5234 13 Oct 16:02 config.yml
-rw-r--r-- 1 lukasm staff 7771 29 Sep 09:46 dump.rdb
-rw-r--r-- 1 lukasm staff 366256 5 Oct 10:28 eir_access.log
-rw-r--r-- 1 lukasm staff 528 26 Sep 16:15 eir_error.log
drwxr-xr-x 10 lukasm staff 340 5 Aug 12:29 lib
-rw-r--r-- 1 lukasm staff 2168 11 Jul 11:13 results.xml
drwxr-xr-x 4 lukasm staff 136 5 Aug 12:29 scripts
drwxr-xr-x 9 lukasm staff 306 20 Oct 11:12 spec
drwxr-xr-x 2 lukasm staff 68 8 Jul 16:38 src
drwxr-xr-x 9 lukasm staff 306 16 Oct 15:04 stub_data
drwxr-xr-x 6 lukasm staff 204 27 Oct 13:55 tmp
-rw-r--r-- 1 lukasm staff 472 7 Jul 12:14 uris.yaml
Question: how can I fix that error "Run configuration error: No Gemfile found" ?
I ran into the same issue today and after many hours I was able to fix it.
The environment where this issue occurred was: Ubuntu 14.04.1, IntelliJ 13.1.5, Test::Unit Test runner UI attached for framework, RubyMine plugin 6.0.0.20140123, rbenv and Ruby 2.0.0-p594
To resolve this "Run Configuration error: No Gemfile found", I had to:
Delete the Gemfile.lock located at the root of the project
From the shell command line, navigate to the project root folder and run "bundle install" so that Bundler started fetching gem metadata from https://rubygems.org/........
Go back to Intellij and reselect an SDK for the project at File -> Project Structure
Go to edit configuration, delete and recreated the rake configuration
I hope this would work for you... I also did other steps that I don't think they were necessary like restarting Intellij, reimporting the project, checking on and off the Test::Unit and the bundler option
Last, I want to share the steps that I had to do to reproduce this issue from scratch:
Imported a ruby project into Intellij
From the shell command, installed rbenv then installed ruby 2.0.0-p594
From Intellij, navigated to File -> Project Structure -> Project and set the SDK to rbenv 2.0.0-p594
From the shell command, installed bundler as follows: gem install bundler
Then back to Intellij were I navigated again to File -> Project Structure -> Project and noticed that the SDK was set back to , I set the SDK back to rbenv 2.0.0-p594
As soon as I hit ok, a popup showed up in IntellJ asking to install Ruby Plugin Gem Manager so I clicked on "install missing gem" link
After the installation was over, I navigated to Edit Configuration, switched to Bundler tab and as soon as I selected Run the script in the context of Bundler check box, I got the "Run Configuration error: No Gemfile found"
So I think that installing bundler late in the project configuration process caused this issue to surface
I have the exact same error, but different rubymine version on Windows and I use Rack configuration tempalte, here is my answer generally it will be the same concept: RubyMine somehow only looks for Gemfile in the "Settings/Project: projectname/Project Structure" so if you copy and paste the missing Gemfile into the Parent Project directory it will get picked up. However, I think Rubymine should also scan for subdirectories where Gemfile existing to include it, or probably rubymine has a way to override the default Gemfile by specify another Gemfile location - this i do not know.
Issue resolved for me.
I had faced the same problem in RubyMine 2018.3.5 version. And I found that my project structure was
Parent Folder
-> Project Folder
-> Vendor folder
-> Gemfile (etc.)
when I changed the structure to
-> Project Folder
-> Vendor folder
-> Gemfile (etc.)
Then there was no complain about "No Gemfile Found"
The basic checks which I think is necessary to debug in Rubymine are
Set the project structure as above
Verify jruby version and its gems(Preferences -> Languages & Frameworks -> Ruby SDK & Gems) My Jruby Version is 9.1.7.0
In Run/Debug Configuration -> Bundler set the checkbox Run the script in context of bundle (bundle exec)
above points are basic ones to begin with, you might need to install the missing gem or right version of gem after this configuration.
I am trying to do a clean build of Lilypond on a clean install of Mac OS X. Dependencies have been installed using Brew.
I do ./configure && make all in the Lilypond dir. It gets through the configure and starts building but ends on
ERROR: In procedure dynamic-link:
ERROR: file: "libguile-srfi-srfi-1-v-3", message: "file not found"
I know what libguile is, I don't know if it's on my system (or where it ought to be) but the configure script completes successfully and does not identify it as a missing dependency so I assume it was able to find it.
Any suggestions? Googling shows some old hints with regards to setting $LD_LIBRARY_PATH, but it seems these don't apply to Mac OS 10.6.
There is a port of lilypond available from MacPorts.
$ port info lilypond
lilypond #2.12.3, Revision 3 (textproc)
Variants: [+]docs, gui, universal
Description: Lilypond is a unix-based automated engraving system that generates beautiful sheet music from input files. Lilypond uses its
own input format, .ly, which in many ways is similar to LaTeX. Lilypond can export sheet music to PDF, EPS, SVG, and PNG
formats, and can also create MIDI files.
Homepage: http://lilypond.org/
Library Dependencies: fontforge, ghostscript, mftrace, guile, texinfo, pango, flex, urw-fonts, netpbm
Platforms: darwin
License: unknown
Maintainers: snc#macports.org, openmaintainer#macports.org
It's likely that the dynamic libraries have a .dylib extension, but need a .so extension.
I worked around this by creating links in /opt/local/lib
Owner-Users-MacBook-Pro-15:~ Carl$ ls -l /opt/local/lib | grep .so
-rwxr-xr-x 2 root admin 45152 Mar 11 12:19 libXcursor.1.dylib
-rw-r--r-- 2 root admin 50232 Mar 11 12:19 libXcursor.a
lrwxr-xr-x 1 root admin 18 Mar 11 12:19 libXcursor.dylib -> libXcursor.1.dylib
-rwxr-xr-x 2 root admin 1105 Mar 11 12:19 libXcursor.la
lrwxr-xr-x 1 root admin 32 Mar 11 21:19 libguile-srfi-srfi-1-v-3.so -> libguile-srfi-srfi-1-v-3.3.dylib
lrwxr-xr-x 1 root admin 36 Mar 11 21:19 libguile-srfi-srfi-13-14-v-3.so -> libguile-srfi-srfi-13-14-v-3.3.dylib
lrwxr-xr-x 1 root admin 32 Mar 11 21:20 libguile-srfi-srfi-4-v-3.so -> libguile-srfi-srfi-4-v-3.3.dylib
lrwxr-xr-x 1 root admin 33 Mar 11 21:20 libguile-srfi-srfi-60-v-2.so -> libguile-srfi-srfi-60-v-2.2.dylib
lrwxr-xr-x 1 root admin 17 Mar 11 21:21 libguile.so -> libguile.17.dylib
lrwxr-xr-x 1 root admin 30 Mar 11 21:22 libguilereadline-v-17.so -> libguilereadline-v-17.17.dylib
The easiest way to install Lilypond on macOS is to use lyp. Apart from installing lyp itself, which is a Ruby gem, there's no other dependency to install, no compilation step. Plus, you can install multiple versions of Lilypond and easily switch between them.
$ gem install lyp
$ lyp install lilypond
To install a specific version:
$ lyp install lilypond#2.19.53
I have enabled the Web Server on my Mac OS X (10.5.6) by going to the System Preferences -> Internet & Network -> Sharing and clicking on the Web Server option. It gives me the IP address of my system as the Web Server, but even after clicking on it, it does not show the page, but gives an error: Failed to Connect
Any help on this would be wonderful.
I have tried with MAMP, but there, I am unable to get my own PHP pages to be displayed.
Starting and/or stopping the web server from the command line, gives me the following:
dyld: Symbol not found: _apr_socket_sendfile
Referenced from: /usr/sbin/httpd
Expected in: /usr/lib/libapr-1.0.dylib
Trace/BPT trap
The files on my box are:
~ $ ls -lrt /usr/lib/libapr*
-rwxr-xr-x 1 root wheel 578640 Sep 23 2007 /usr/lib/libapr-1.0.2.7.dylib
-rwxr-xr-x 1 root wheel 393696 Sep 23 2007 /usr/lib/libaprutil-1.0.2.7.dylib
-rwxrwxr-x 1 root admin 569544 Oct 27 10:44 /usr/lib/libaprutil-1.0.2.12.dylib
-rwxrwxr-x 1 root admin 835824 Oct 27 10:44 /usr/lib/libapr-1.0.2.12.dylib
lrwxr-xr-x 1 root admin 25 Jan 4 22:28 /usr/lib/libaprutil-1.dylib -> libaprutil-1.0.2.12.dylib
lrwxr-xr-x 1 root admin 25 Jan 4 22:28 /usr/lib/libaprutil-1.0.dylib -> libaprutil-1.0.2.12.dylib
lrwxr-xr-x 1 root admin 21 Jan 4 22:28 /usr/lib/libapr-1.dylib -> libapr-1.0.2.12.dylib
lrwxr-xr-x 1 root admin 21 Jan 4 22:28 /usr/lib/libapr-1.0.dylib -> libapr-1.0.2.12.dylib
The MD5 values are:
~ $ openssl md5 /usr/sbin/httpd /usr/lib/libapr-1.0.dylib
MD5(/usr/sbin/httpd)= d5194790e9cfaca881d0e8fc3db24889
MD5(/usr/lib/libapr-1.0.dylib)= f7f7fd3cd315ca4e7c6ad2a2f222998e
The libapr-1.0.dylib is linked to the newer 1.0.2.12 version.
I'm having a similar problem and got to this page cause I feel lucky today :)
The error means the httpd binary makes reference to a symbol that cannot be found in the appropriate dynamically linked library (libapr in this case, Apache Runtime Library)
Running same OS X version on an iMac. /usr/lib/libapr-1.0.dylib points to the most recent version of the library (1.0.2.12), which does not define the method/symbol _apr_socket_sendfile
The other version of libapr library is 1.0.2.7, does have that symbol. I had the symbolic links point to this version instead (also for libaprutil):
// BEFORE (both 1.0.2.7 and 1.0.2.12 versions exist, symlinks point to .12)
bash-3.2# ls -l libapr*
-rwxr-xr-x 2 userna wheel 835824 19 Sep 2008 libapr-1.0.2.12.dylib
-rwxr-xr-x 1 root wheel 578640 24 Sep 2007 libapr-1.0.2.7.dylib
lrwxrwxrwx 1 root wheel 21 14 Oct 2008 libapr-1.0.dylib -> libapr-1.0.2.12.dylib
lrwxrwxrwx 1 root wheel 21 14 Oct 2008 libapr-1.dylib -> libapr-1.0.2.12.dylib
-rwxr-xr-x 2 userna wheel 569544 19 Sep 2008 libaprutil-1.0.2.12.dylib
-rwxr-xr-x 1 root wheel 393696 24 Sep 2007 libaprutil-1.0.2.7.dylib
lrwxrwxrwx 1 root wheel 25 14 Oct 2008 libaprutil-1.0.dylib -> libaprutil-1.0.2.12.dylib
lrwxrwxrwx 1 root wheel 25 14 Oct 2008 libaprutil-1.dylib -> libaprutil-1.0.2.12.dylib
// CHANGE symbolic links to previous version
bash-3.2# ln -sf libapr-1.0.2.7.dylib libapr-1.0.dylib
bash-3.2# ln -sf libapr-1.0.2.7.dylib libapr-1.dylib
bash-3.2# ln -sf libaprutil-1.0.2.7.dylib libaprutil-1.0.dylib
bash-3.2# ln -sf libaprutil-1.0.2.7.dylib libaprutil-1.dylib
// TESTing whether httpd is loaded and run
bash-3.2# /usr/sbin/httpd -t
Syntax OK
Not really sure whether this has been broken since last software update that included Apache Web Server, I rarely use it on this box.
Hope it helps ;)
Jon
Maybe there are errors in the apache configuration.
If you disable the webserver in the System preferences and open a terminal you can check your configuration by typing:
httpd -t
If it gives a "Syntax OK" you're one step closer to a working webserver. Start the webserver manually by typing:
sudo httpd
If there are no errors, the webserver is started. Open a browser and check http://localhost. Sometimes an extra refresh (Command+R) is needed.
Stopping the server:
sudo httpd -k stop
If the manually starting works, but the System preferences version doesn't,
Check the /var/log/apache2/error_log for errors.
You can try to do the following:
1. backup current apache
sudo cp /usr/sbin/httpd /usr/sbin/httpd.old
2. Replace link in apache binary file to the old library:
sudo install_name_tool -change /usr/lib/libapr-1.0.dylib /usr/lib/libapr-1.0.2.7.dylib /usr/sbin/httpd
After it should work as expected and other applications will still use the new version of libapr
http://foundationphp.com/tutorials/php_leopard.php
I followed this on mine and it work perfectly.
Try to follow it a see if it works
If you're trying to access the pages from the computer itself just use http://localhost/. Also, try typing the IP address into the browser (instead of just clicking the link in the interface); don't ask me to explain why, but I've seen that make a difference.
Also make sure the OS's firewall is set to allow connections to Apache. "Personal Web Sharing" should be checked.
Have you installed the subversion 1.5 client on your machine? I am seeing the same problem and it seems to be down to this. /usr/lib/libapr-1.0.dylib is normally linked to /Developer/SDKs/MacOSX10.5.sdk/usr/lib/libapr-1.0.2.7.dylib but if you install svn 1.5 it gets re-linked to /opt/subversion/lib/libapr-1.0.2.12.dylib.
The 1.0.2.12 version of the library does not contain the symbol _apr_socket_sendfile.
Having found the problem I'm still trying to figure out how to solve it.
Doug
Stupid question. Does libapr-1.0.dylib exist?
On my Macbook, /usr/lib/libapr-1.0.dylib is a symbolic link to libapr-1.0.2.7.dylib.
Running nm against that file does show that apr_socket_sendfile is defined in that library.
Also: do standard html pages work properly?
It sounds like either:
The httpd that comes first in your path is not /usr/sbin/httpd, and it was built strangely, or
You have installed something that overwrote either httpd or libapr.
What's the MD5 of your httpd and libapr? I get:
$ openssl md5 /usr/sbin/httpd /usr/lib/libapr-1.0.dylib
MD5(/usr/sbin/httpd)= d5194790e9cfaca881d0e8fc3db24889
MD5(/usr/lib/libapr-1.0.dylib)= f1f5ece10140a5c2cea2fece7a9e4918