[ptxdist] regression linking against libcgi after fixing patches in a53b663

[ptxdist] regression linking against libcgi after fixing patches in a53b663

[Alexander Dahl]

Hei hei, 

in commit a53b663 a typo in Makefile of LibCGI was fixed, actually in
the patches fixing various things in LibCGI. This sets the soname of the
library to libcgi.so.0 in the options passed to the linker. However this
libcgi.so.0 is never installed but libcgi.so instead. This does work at
build time when compiling and linking against libcgi but at runtime this
fails because libcgi.so is installed in both install and targetinstall
stage but there's no libcgi.so.0 present on the target.

I'm not familiar with autotools to fix this myself. Expected behaviour
would be some libcgi.so.x.y.z is installed and libcgi.so.x and libcgi.so
are symlinks of which only libcgi.so.x is installed to the target. If I
rename libcgi.so on my target to libcgi.so.0.0.0 and create a symlink
libcgi.so.0 my cgis work again.

by the way: the author of the libcgi library moved the code from
sourceforge to github [1] and kindly responses to mails, bug reports und
pull requests. I'll ask him how he plans to proceed with this long dead
project and I'll share the result here. He already integrated the
changes and fixes of the patches ptxdist carries in the master branch (I
integrated those in one single commit in my personal fork and made a
pull request in the expectation nobody would care, so sorry if you're
not getting your credits.) I'll come up with some other patches fixing
some things in the lib. I would share these with ptxdist as well and
hope we could work toward some bugfix release 1.x of the lib integrating
all the current changes to get rid of the patches. What's your opinion
about this?

Greets
Alex

[1] https://github.com/rafaelsteil/libcgi

-- 
»With the first link, the chain is forged. The first speech censured,
the first thought forbidden, the first freedom denied, chains us all
irrevocably.« (Jean-Luc Picard, quoting Judge Aaron Satie)
*** GnuPG-FP: 02C8 A590 7FE5 CA5F 3601  D1D5 8FBA 7744 CC87 10D0 ***

-- 
ptxdist mailing list
ptxdist@pengutronix.de

Re: [ptxdist] regression linking against libcgi after fixing patches in a53b663

[Michael Olbrich]

Hi,

On Wed, Oct 24, 2012 at 11:05:16AM +0200, Alexander Dahl wrote:
> in commit a53b663 a typo in Makefile of LibCGI was fixed, actually in
> the patches fixing various things in LibCGI. This sets the soname of the
> library to libcgi.so.0 in the options passed to the linker. However this
> libcgi.so.0 is never installed but libcgi.so instead. This does work at
> build time when compiling and linking against libcgi but at runtime this
> fails because libcgi.so is installed in both install and targetinstall
> stage but there's no libcgi.so.0 present on the target.
> 
> I'm not familiar with autotools to fix this myself. Expected behaviour
> would be some libcgi.so.x.y.z is installed and libcgi.so.x and libcgi.so
> are symlinks of which only libcgi.so.x is installed to the target. If I
> rename libcgi.so on my target to libcgi.so.0.0.0 and create a symlink
> libcgi.so.0 my cgis work again.

I think I fixed this now, but I can't test it. Please try.

> by the way: the author of the libcgi library moved the code from
> sourceforge to github [1] and kindly responses to mails, bug reports und
> pull requests. I'll ask him how he plans to proceed with this long dead
> project and I'll share the result here. He already integrated the
> changes and fixes of the patches ptxdist carries in the master branch (I
> integrated those in one single commit in my personal fork and made a
> pull request in the expectation nobody would care, so sorry if you're
> not getting your credits.) I'll come up with some other patches fixing
> some things in the lib. I would share these with ptxdist as well and
> hope we could work toward some bugfix release 1.x of the lib integrating
> all the current changes to get rid of the patches. What's your opinion
> about this?

That sounds good. Especially the part about a release. I always like it
when I can remove patches from ptxdist. If someone is actually working on
this, I really suggest cleaning up the build system. If you don't know
enough about this, then look at libabc[1]. This is a good place to get
started with this.

Michael

[1] http://0pointer.de/blog/projects/libabc.html

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

-- 
ptxdist mailing list
ptxdist@pengutronix.de

Re: [ptxdist] regression linking against libcgi after fixing patches in a53b663

[Alexander Dahl]

Hei hei, 

Am 2012-10-28 18:09, schrieb Michael Olbrich:
>> I'm not familiar with autotools to fix this myself. Expected behaviour
>> would be some libcgi.so.x.y.z is installed and libcgi.so.x and libcgi.so
>> are symlinks of which only libcgi.so.x is installed to the target. If I
>> rename libcgi.so on my target to libcgi.so.0.0.0 and create a symlink
>> libcgi.so.0 my cgis work again.
> 
> I think I fixed this now, but I can't test it. Please try.

Works now. File is libcgi.so.0 and symlink is libcgi.so pointing to the
file. 

> That sounds good. Especially the part about a release. I always like it
> when I can remove patches from ptxdist. If someone is actually working on
> this, I really suggest cleaning up the build system. If you don't know
> enough about this, then look at libabc[1]. This is a good place to get
> started with this.

Okay I'll try to get a release out, we'll see.

> [1] http://0pointer.de/blog/projects/libabc.html

I will have a look at this, thank you. (Had already forgotten I
stumbled over libabc a long while ago.)

Greets
Alex

-- 
»With the first link, the chain is forged. The first speech censured,
the first thought forbidden, the first freedom denied, chains us all
irrevocably.« (Jean-Luc Picard, quoting Judge Aaron Satie)
*** GnuPG-FP: 02C8 A590 7FE5 CA5F 3601  D1D5 8FBA 7744 CC87 10D0 ***

-- 
ptxdist mailing list
ptxdist@pengutronix.de