I am back again with an issue on iOS that only seems to be affecting
devices and not the simulator. I did not have access to a device until
this point so it went unnoticed. I can build the library with no compiler
issues but when I try to run it on a device(this one is arm64 architecture)
I
I meant to put undefined symbols not unreferenced
On Monday, February 20, 2017 at 1:19:52 PM UTC-6, pdoh...@cipherloc.net
wrote:
>
> I am back again with an issue on iOS that only seems to be affecting
> devices and not the simulator. I did not have access to a device until
> this point so it
Thank you for the reply Jeff that did fix the issue. I am now getting a
different issue. It looks like even when I use setenv-ios.sh arm64, it is
still building the library for x86_64 architecture. It is also not
building it for mobile according to xcode. It looks like all the .o
That fixed it as well. Thank you for that. Now the error I am getting is
about missing symbols on 4 functions. Those 4 functions are
Filter::TransferTo2, CopyRangeTo2, BufferedTransformation::Skip, and
StringStore::TransferTo2. I have the error output below. The only other
change I have
I am having an issue when trying to use the Crypto++ Library on an iOS
device. I do have the Crypto++ library bundled inside of another Library,
called libA.a, but I have tested this with other Libraries and it is
possible to do this. I did check if I get this error when just referencing
the
We are using 5.6.3 I believe. I just looked in the config.h file and it
says CRYPTOPP_VERSION 563 so believe that is correct in saying 5.6.3. I am
building it inside Xcode if that is what you meant by Master/zip/something
else. The other library it is packaged within is using Crypto++. I
Sorry about not replying to the last response, but am still having this
issue even after removing the -fembed-bitcode. I have included the output
of the errors below. I am unfortunately stuck with iOS 10 and cannot use
an older version due to a different library. Any help or an idea of where
I just double checked and we are actually using 5.6.6 for the iOS version.
On Monday, November 21, 2016 at 4:00:17 PM UTC-6, pdoh...@cipherloc.net
wrote:
>
> We are using 5.6.3 I believe. I just looked in the config.h file and it
> says CRYPTOPP_VERSION 563 so believe that is correct in saying
I have included the link command we are using below:
CXXFLAGS ?= -DNDEBUG -g2 -Os -fPIC -pipe
CXXFLAGS += -std=c++11
# The following options reduce code size, but breaks link or makes link
very slow on some systems
# CXXFLAGS += -ffunction-sections -fdata-sections
# LDFLAGS += -Wl,--gc-sections
no libA is using the 5.6.6 version.
On Monday, November 21, 2016 at 5:05:11 PM UTC-6, Jeffrey Walton wrote:
>
> On Mon, Nov 21, 2016 at 6:02 PM,
> wrote:
> > I just double checked and we are actually using 5.6.6 for the iOS
> version.
> >
>
> Does that mean libA
I am still having the issue that I had posted about last week. I have dug
deeper and still haven't found a solution. I have called the set-env
function correctly I believe as well as made sure that the line for
-stdlib=libc++ are properly set. Below I have the output of . ./setenv-ios
So just as a sanity check, I decided to pull the cryptopp lib out by itself
and see if I could use that by itself to make sure that the changes I had
made before didn't mess with that library. It does not work however and I
get completely different errors. When trying to use cryptopp on its
So after many sanity checks, everything has been sorted out and is working
now. The problem was that the library was being built incorrectly and was
fixed by erasing my version of cryptopp and using the 5.6.5 version freshly
installed. This solved the errors that I was getting. Why they did
Hey Jeff,
Sorry for not responding as quickly, I was tasked to another project but am
back now. Below is the command I believe you are looking for. This is
being used inside of xcode:
Ld
I have made that change but the error persists. The call from Xcode looks
the same as before but now the call has -llibA first instead of cryptopp
On Tuesday, December 6, 2016 at 7:36:30 PM UTC-6, Jeffrey Walton wrote:
>
>
>
> On Tuesday, December 6, 2016 at 3:19:42 PM UTC-5,
Apologies for replying so late. Work got a bit busy. I was able to solve
this by removing bench.cpp from our folder for compiling. It seems that it
was left over after an update.
Patrick
On Monday, February 20, 2017 at 1:57:49 PM UTC-6, Jeffrey Walton wrote:
>
> On Mon, Feb 20, 2017 at 2:40
I have my library, which is static, linked with a static cryptopp library.
I have cryptopp set to build lean and have the flags for -ffunctions and
-fdata-sections in my make file as well. However, when I call strip on the
final product nothing is actually stripped and the cryptopp library
Here is the build script we use to build our library with the arguments of
static: http://pastebin.com/yeH8xBNs
and we are using the cryptopp GNUmakefile-cross with static lean as the
arguments.
Patrick
On Monday, March 13, 2017 at 2:50:40 PM UTC-5, Jeffrey Walton wrote:
>
> On Mon, Mar 13,
Here is the build log from our library: http://pastebin.com/srRGzy2D
We have not modified the cryptopp script in anyway so the output would look
the same as if you called it with static lean.
The interesting thing here is that it says -dynamic not specified and
-dead_strip is ignored.
Patrick
19 matches
Mail list logo