[argyllcms] Re: "Development in progress" update

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Tue, 11 Jul 2006 00:46:40 +1000

Klaus Karcher wrote:

>>> ------------
>>> libusb
>>> ----------
>>> ...found 152 target(s)...
>>> ...updating 8 target(s)...
>>>
>>> In file included from usb.c:14:
>>> usbi.h:4:17: error: usb.h: No such file or directory
>>> ...failed Cc usb.o ...
>>> error.c:12:17: error: usb.h: No such file or directory
>>> ...failed Cc error.o ...
>>
>>
>> I'd really love to know what version/configuration of
>> gcc it is, that fails to search the . directory when
>> it comes across a #include "file", as it seems to
>> be contrary to the whole purpose of #include "file" !

Looking through the gcc doco, the using -I- causes this
behaviour, if in fact this is the problem.

> btw.: ther were two files "usb.h" in the last snapshot:
> ./libusb/usb.h and
> ./libusbw/src/usb.h
>
> now there is only the latter.

No, there are two in the current snapshot, but ./libusb/usb.h
won't appear until ./libusb/configure is run, which should
happen when jam is run. It runs configure if ./libusb/Makefile
is non-existent or out of date. It seems like
configure didn't run. Did you update Jambase ?
When you run it after unzipping into a clean directory,
does it run configure and create usb.h and the Makefile ?

just made a first test with profile 0.60: it seems to work as expected
(unfortunately the profile still shows the "bumpy" black on the gamut hull I mentioned in December if you don't use it with icclink; see <http://digitalproof.info/argyll/k-test/>)

Yep, I've had a look at it a few times, but I don't have any idea on how to solve it yet, so I wasn't going to hold up the next release.

but icclink causes errors:

icclink: Error - Write file: 1, icmColorantTableVal_write: write of PCS
coord failed

I tried some different links out, and didn't get such a failure. You will have to send me the ICC profiles and the command line you used to be able to fix this.

- Federer beat Nadal
- Italy is soccer world champion

I didn't stay up to see the soccer - it's hard enough to get up in the morning as it is !

- I get nearly no warnings using gcc-3.3 ibemerkennstead of 4.0
  (see below)
- icclink still says "Error - Write file ... failed"

possibly serious warnings I noticed using gcc-3.3:

------------
rspl
----------
don't know how to make ../xicc/libxicc.a
don't know how to make ../spectro/libinsttypes.a
don't know how to make ../gamut/libgamut.a
don't know how to make ../xicc/libxcolorants.a
...found 137 target(s)...
...updating 34 target(s)...
...can't find 4 target(s)...
...can't make 2 target(s)...
Cc smtnd.o
...skipped smtnd for lack of ../xicc/libxicc.a...
Cc smtmpp.o
...skipped smtmpp for lack of ../xicc/libxicc.a...
...skipped 2 target(s)...

There is always a lot of noise because of the circular dependencies between directories. The code would need to be majorly reorganized to avoid it, or a better build system used that has a global view on the dependencies.

------------
spectro
----------
unixio.c: In function `icoms_get_paths':
unixio.c:211: warning: comparison between pointer and integer

Apples reference documentation is pretty awful a lot of the time. It seems to be very short on details like what are the valid return values. I think I can fix this.

Graeme Gill.


Other related posts: