Renaming boost libraries
Heya Guys, Just wondering if it's possible to rename the boost libraries after they've been built. Specifically, I want to remove the boost version (-1_33) that is appended to all the lib and dll libraries. The reason is twofold; a) our source control system will mark the libs with a label to indicate the version and b) it should make migrating to a new boost version easier. I'm pretty sure it's fine to just physically rename the static libraries but seem to remember that for dynamic libs you can't do that because the matching .lib file indicates which .dll the application should load. Am I mistaken? If not is there any other way to specify that you don't want the boost version appended to the end of the filename? Cheers, Matt
Matt wrote:
I'm pretty sure it's fine to just physically rename the static libraries but seem to remember that for dynamic libs you can't do that because the matching .lib file indicates which .dll the application should load.
Am I mistaken?
No.
If not is there any other way to specify that you don't want the boost version appended to the end of the filename?
Use the "--layout=system" option when building/installing. -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org - grafik/redshift-software.com -- 102708583/icq - grafikrobot/aim - Grafik/jabber.org
Rene Rivera wrote:
If not is there any other way to specify that you don't want the boost version appended to the end of the filename?
Use the "--layout=system" option when building/installing.
Thanks Rene, that worked a treat. I couldn't find much documentation about this feature, is there any plans to expose it? It seems like a common thing to want to do. Can I help? Also, I found some weirdness when building libraries. Specifically, the Test library built libs for the dynamic versions altough the dll's themselves were not built. I understand that the test library doesn't have a dynamically linked version (according to the faq: http://tinyurl.com/77cjd) so why are the (dynamic) .libs being generated? I also found that a couple of the libraries were producing seemingly duplicate libs. By that I mean that two versions of a lib were being created, one _with_ the boost version number appended and one without - the filesizes were the same and, according to the naming convenion (http://tinyurl.com/aycpa), I believe the ones without the version number should not exist. I was using the 'stage' option with the vc-7_1 toolkit and noticed this on at least the serialization and wave libraries. Are these two issues known? I may have been using an old version of bjam though if that helps... Cheers, Matt
Matt wrote:
Rene Rivera wrote:
If not is there any other way to specify that you don't want the boost version appended to the end of the filename?
Use the "--layout=system" option when building/installing.
Thanks Rene, that worked a treat.
Welcome.
I couldn't find much documentation about this feature, is there any plans to expose it? It seems like a common thing to want to do. Can I help?
Well it's a feature used by packagers so it's there to stay. Docs, yes you can help if you want to write up a short explanation of what it does we can incorporate that into the getting started docs.
Also, I found some weirdness when building libraries. Specifically, the Test library built libs for the dynamic versions altough the dll's themselves were not built. I understand that the test library doesn't have a dynamically linked version (according to the faq: http://tinyurl.com/77cjd) so why are the (dynamic) .libs being generated?
Not sure why, I'd have to investigate.
I also found that a couple of the libraries were producing seemingly duplicate libs. By that I mean that two versions of a lib were being created, one _with_ the boost version number appended and one without - the filesizes were the same and, according to the naming convenion (http://tinyurl.com/aycpa), I believe the ones without the version number should not exist. I was using the 'stage' option with the vc-7_1 toolkit and noticed this on at least the serialization and wave libraries.
Again I'd have to investigate... Are you sure that you aren't seeing left overs from previous compiles on this one?
Are these two issues known?
No.
I may have been using an old version of bjam though if that helps...
As long as it's 3.1.10 it should be OK. But using the newest version is always better as that's what we test with. -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org - grafik/redshift-software.com -- 102708583/icq - grafikrobot/aim - Grafik/jabber.org
But if the boost library and dll compiled with no version suffix of the end
of the file, when I use the vc71 to compile the code with MACRO "
BOOST_WHATEVER_DYN_LINK" , "BOOST_ALL_DYN_LINK", "BOOST_LIB_DIAGNOSTIC", the
Automatic Linking on Windows feature can not work, I also want to know how
to solve this problem if I do not want to version suffix?
----- Original Message -----
From: "Matt"
Rene Rivera wrote:
If not is there any other way to specify that you don't want the boost version appended to the end of the filename?
Use the "--layout=system" option when building/installing.
Thanks Rene, that worked a treat.
I couldn't find much documentation about this feature, is there any plans to expose it? It seems like a common thing to want to do. Can I help?
Also, I found some weirdness when building libraries. Specifically, the Test library built libs for the dynamic versions altough the dll's themselves were not built. I understand that the test library doesn't have a dynamically linked version (according to the faq: http://tinyurl.com/77cjd) so why are the (dynamic) .libs being generated?
I also found that a couple of the libraries were producing seemingly duplicate libs. By that I mean that two versions of a lib were being created, one _with_ the boost version number appended and one without - the filesizes were the same and, according to the naming convenion (http://tinyurl.com/aycpa), I believe the ones without the version number should not exist. I was using the 'stage' option with the vc-7_1 toolkit and noticed this on at least the serialization and wave libraries.
Are these two issues known?
I may have been using an old version of bjam though if that helps...
Cheers, Matt
_______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users
Has anyone been able to compile a recent version of Boost on AIX? I keep getting a compile error in the is_enum.hpp file. Something about "non-type template parameter of type bool". Failing that, what is the last known version to successfully compile on AIX? If anyone is interested in figuring it out, I can include a log of the failed compilation. I'm quite willing to test any ideas in order to help get Boost compiling on AIX. Regards, Dieter
??? wrote:
But if the boost library and dll compiled with no version suffix of the end of the file, when I use the vc71 to compile the code with MACRO " BOOST_WHATEVER_DYN_LINK" , "BOOST_ALL_DYN_LINK", "BOOST_LIB_DIAGNOSTIC", the Automatic Linking on Windows feature can not work, I also want to know how to solve this problem if I do not want to version suffix?
It's not currently possible. An it would take some reworking of the autolink code to allow for the possibility. For a longer discussion about this see these previous posts: http://thread.gmane.org/gmane.comp.lib.boost.user/12663 http://article.gmane.org/gmane.comp.lib.boost.user/12677 -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org - grafik/redshift-software.com -- 102708583/icq - grafikrobot/aim - Grafik/jabber.org
participants (4)
-
??? -
Dieter Rosch -
Matt -
Rene Rivera