On Feb 13, 2015, at 3:25 AM, "Alan W. Irwin" <ir...@beluga.phys.uvic.ca> wrote:

> On 2015-02-13 00:04-0500 Jim Dishaw wrote:
> 
>> The attached patch is a roll-up of all the changes that I have made:
>> 
>> - Moved the memory allocation routines from pdf_utils to plmem.c (new file)
>> - Bug fixes to plbuf.c
>> - Reading/Writing plot metafiles
>> - Test routine for plbuf.c
>> 
>> The plbuf.c changes should be production ready.  I'm still trying to figure 
>> out why there is an extraneous color change at the beginning on a replot, 
>> but it does not affect the appearance.
>> 
>> The plot metafiles is still is a work in progress, but I wanted to make it 
>> available for testing.  There are two new command line options, -mfo and 
>> -mfi.  The -mfi option will read in a plot metafile and almost correctly 
>> displays the output.  I have not finished the output method.
> 
> Thanks very much for your efforts.
> 
> I have found some minor issues that need fixing before I can test this
> patch series and commit it to master.
> 
> test_plbuf doesn't build because of visibility issues.  Here is the
> error message I get using the gcc option -fvisibility=hidden (which
> makes all library functions in Linux hidden unless specified otherwise
> so it is very similar to the Windows situation).
> 
> Linking C executable test_plbuf
> CMakeFiles/test_plbuf.dir/test_plbuf.c.o: In function main':
> test_plbuf.c:(.text.startup+0x32): undefined reference to
> plreadmetafile'
> 
> To fix this visibility issue (this will also fix all visibility issues
> on Windows) you must declare plreadmetafile in some PLplot header
> using the PLDLLIMPEXP macro. (Currently you don't declare
> plreadmetafile anywhere so I assume ANSI C is making some guesses
> about what that declaration is.)
> 
> Your header choices depend on whether you feel this API is worth
> having for each of our language bindings or only just for C.
> In the former case you should declare the function in include/plplot.h
> but otherwise you should declare it in include/plplotP.h.

I did not declare it anywhere because I was not sure if we wanted to expose 
this method yet and if the name was appropriate. 

> 
> When you fix this (as well as any other visibility issues you might
> have for new functions like plreadmetafile that you have introduced),
> then I suggest you should rebase your topic branch before sending
> the series of patches.  In the present case git am errored out
> the first time I tried it because your patch series was not based
> on the current tip of the master branch (i.e., your patch series
> included

My git-fu is weak.  I did a "git fetch" instead of a "git pull" on master.  I 
will fix a few things and resend the patch.

> 
> The other issue I noticed is the current test_plbuf source code exits
> almost immediately after reading in a metafile and playing the buffer.
> Did you mean that to be the complete test from now on?  If so we need
> some metafile to read in so the test can be run, and you should also
> remove all the large amount of code after the exit.
> 

That was accidentally left in while I was debugging.  The new patch will have 
an updated test routine that exercises the plot buffer and plmeta.  I thought 
it would be good to put those two together in order to detect a "drift" in 
implementation.

> I look forward to your updated patch series with the above issues fixed.
> I also hope you achieve quick turnaround because Phil's series of
> patches should arrive shortly which are based on top of yours.
> 

I will have the updates out later today.


------------------------------------------------------------------------------
Dive into the World of Parallel Programming. The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Plplot-devel mailing list
Plplot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/plplot-devel

Reply via email to