[OpenJDK 2D-Dev] [9] RFR JDK-6574279: validatePage from PrinterJob returns argument under linux
Phil Race
philip.race at oracle.com
Fri Nov 4 20:21:07 UTC 2016
526 protected void validatePaper(Paper origPaper, Paper newPaper) { ...
I think this method needs some careful inspection.
533 Class<?> psClass = Class.forName("sun.print.IPPPrintService");
I am sure this reflection code is because the source class is shared but
the target class is platform-specific.
Why can't you just use the standard API for this ?
PrintAttributeSet aset = new HashPrintAttributeSet();
aset.add(MediaSizeName.NA_LETTER) ; // I'm being lazy - you need to derive
the correct Media from the Paper size you have.
MediaPrintableArea[] mpas = (MediaPrintableArea[])
PrintService.getSupportedAttributeValues(MediaPrintableArea.class, null,
aset);
The paper context is important as otherwise you have no idea how
an MPA corresponds to a particular size.
We also have the hwMargins to look at. This also needs to be massaged
so we don't need reflection but hwMargins should be exposed via
the service as the MediaPrintableArea and has no place here.
It seems that you don't think you can get the info from the PPD that
the spec. says should be there .. per my email :-
http://mail.openjdk.java.net/pipermail/2d-dev/2016-August/007471.html
I think we should put this one on the back-burner as there's some
work and investigation to be done and its looking too risky to fix in
the time left in 9.
-phil.
On 09/02/2016 03:00 AM, Prasanta Sadhukhan wrote:
> We match the CUPS returned media to standard media here
> http://hg.openjdk.java.net/jdk9/client/jdk/file/03c248ab7484/src/java.desktop/unix/classes/sun/print/CUPSPrinter.java#l218
>
> As per getPageSizes(), it returns the size of the media as
> mentioned(including margin) but the testcase actually sets
> imageablearea to (0,0,paperwidth,paperheight) which matches one of the
> many page size(s)
> as I earlier mentioned, (0,0,8.26",11.69") is within supported mpa of
> a paper size 0,0,11.69",16.52" (which is one of the pagesize returned
> by getPageSizes()) so it means 0,0,8",11" comes out to be supported mpa
> and so it was allowed (eventhough printer cannot print from 0,0)
> so we should also take care of the printer-allowable (hardware) margin
> which in windows was taken care of by GetDeviceCaps(printDC,
> PHYSICALOFFSETX) function so I had to add getHWMargin() in linux here
> in this fix.
>
> Modified webrev which passes JPRT is here
> http://cr.openjdk.java.net/~psadhukhan/6574279/webrev.04/
>
> Regards
> Prasanta
> On 8/19/2016 3:56 AM, Philip Race wrote:
>> PS as I just mentioned in another review thread, I have
>> a suspicion that the Media and MPA for CUPS are never matched
>> to "standard" (pre-defined in API) ones. I don't know how much this
>> matters - A4
>> when sent to printing should (he says) get matched on size
>> to the right media but you should check that, but this also
>> may mean that if you try to get the MPA for the "standard"
>> MediaSizeName.ISO_A4
>> none will be found .. all this sets back this fix until you work
>> through that
>> but I don't think the fix can be completely right until we have an
>> understanding
>> and resolution on that.
>>
>> -phil.
>>
>> On 8/17/16, 12:07 PM, Phil Race wrote:
>>> The cups docs say you can get physical margins as described here :
>>> http://opensource.apple.com//source/cups/cups-62/doc/spm.shtml?txt#CUPS_API
>>> ------------------------
>>> PPD files specify all of the available pages sizes and the physical
>>> margins associated with them. ...
>>> ppd_sizet_ size = ppdPageSize(ppd, "/size/"); ...
>>> The|||ppd_size_t| structure contains the width, length, and margin
>>> information:
>>>
>>> typedef struct /**** Page Sizes ****/
>>> {
>>> int marked; /* Page size selected? */
>>> char name[41]; /* Media size option */
>>> float width, /* Width of media in points */
>>> length, /* Length of media in points */
>>> left, /* Left printable margin in points */
>>> bottom, /* Bottom printable margin in points */
>>> right, /* Right printable margin in points */
>>> top; /* Top printable margin in points */
>>> } ppd_size_t;
>>>
>>>
>>> --------------------
>>>
>>> Moreover we are *already* calling that code : See
>>> Java_sun_print_CUPSPrinter_getPageSizes Does that not work ? Perhaps
>>> the gap is that CUPSPrinter.java does not tie the MPA to the Media
>>> it has created. Not sure of the best way to do this. Maybe the
>>> windows code solves this differently but a simple HashMap may work
>>> here. You'll have to investigate. A few specific comments on the
>>> code :- 484 if (option != NULL && option->num_choices > 0) { we are
>>> expecting an array of 4 elements here so may be >=4 is what we
>>> should require ? 500 JNU_ThrowOutOfMemoryError(env, "Could not
>>> create printer name"); comment seems to be a copy from an earlier
>>> line .. this is about allocating an array 516 @Override
>>> 517 protected void validatePaper(Paper origPaper, Paper
>>> newPaper) { 518 if (!(getPrintService() instanceof
>>> IPPPrintService)) { 519 super.validatePaper(origPaper,
>>> newPaper); 520 return; Hmm. IPPPrintService is Unix
>>> only whereas PSPrinterJob is shared code .. so I will bet you didn't
>>> try yet to compile this on Windows. -phil. On 8/16/16, 4:47 AM,
>>> Prasanta Sadhukhan wrote:
>>>> Hi All,
>>>> On 8/11/2016 2:08 PM, Prasanta Sadhukhan wrote:
>>>>> Hi Phil,
>>>>> On 8/11/2016 3:05 AM, Phil Race wrote:
>>>>>> There does not seem to be a link to the new webrev in here. Is it
>>>>>> this : http://cr.openjdk.java.net/~psadhukhan/6574279/webrev.01/
>>>>>> ? > to default Paper (Letter in US, A4 otherwise) imageable area
>>>>>> I don't think that will ever be A4. Where is that coming from ?
>>>>>> And you still have not fixed the problem with how you retrieve
>>>>>> the MPA :- 512 MediaPrintableArea mpa =
>>>>>> (MediaPrintableArea) 513
>>>>>> getPrintService().getDefaultAttributeValue(MediaPrintableArea.class);
>>>>>> It is not supposed to be the MPA of the *default* paper. You want
>>>>>> the MPA of *this* paper. Only if origPaper is not a supported
>>>>>> paper do you go look for the default paper.
>>>>> As you suggested, I have modified to check if mpa of origPaper is
>>>>> amongst supported mpa(s) of the printer, if not, then go for mpa
>>>>> of default paper. However, it seems among 61 mpas returned by
>>>>> CUPSPrinter , origPaper of 0,0,8.26",11.69" is within supported
>>>>> mpa of a paper size 0,0,11.69",16.52" so it means 0,0,8",11" comes
>>>>> out to be supported mpa. Can you please comment on this modified
>>>>> webrev on what more we can do?
>>>>> http://cr.openjdk.java.net/~psadhukhan/6574279/webrev.02/ I guess
>>>>> we need to find out the physical limitation of printer via CUPS
>>>>> same what we do in windows via GetDeviceCaps(printDC,
>>>>> PHYSICALOFFSETX); and use that in calculation for valid margin but
>>>>> I could not find equivalent cups function. Can you suggest anyways?
>>>> I tried calling ppdFindOption(ppd, "HWMargins") in CUPS but it
>>>> returns null although my printer ppd file has this line *%===
>>>> Custom Paper Support ================= *LeadingEdge Short: ""
>>>> *LeadingEdge Long: "" *DefaultLeadingEdge: Short *MaxMediaWidth:
>>>> "865" *MaxMediaHeight: "1701" **HWMargins: 12 12 12 12
>>>> *Anyways, I have updated the webrev to take care of the review
>>>> comments and also get HWMargin (if the above ppdFindOption return
>>>> null then fall back to custom margin as reported via ppd_file_t
>>>> structure https://www.cups.org/doc/api-ppd.html#ppd_file_s)
>>>> http://cr.openjdk.java.net/~psadhukhan/6574279/webrev.03/ Please
>>>> review. Tested in ubuntu and solaris11. Regards Prasanta
>>>>> Regards Prasanta
>>>>>> -phil. On 08/05/2016 03:20 AM, Prasanta Sadhukhan wrote:
>>>>>>> On 8/4/2016 5:45 PM, Prasanta Sadhukhan wrote:
>>>>>>>> I will override in PSPrinterJob but 1662
>>>>>>>> MediaPrintableArea mpa = 1663
>>>>>>>> (MediaPrintableArea)getPrintService(). 1664
>>>>>>>> getDefaultAttributeValue(MediaPrintableArea.class); is
>>>>>>>> returning imageable area of default media of chosen printer
>>>>>>>> from CUPS [ all supported media printable area is obtained from
>>>>>>>> CUPSPrinter#getPageSizes(printer) & then initMedia() populates
>>>>>>>> cupsMediaPrintables which is returned in
>>>>>>>> getMediaPrintableArea()] as
>>>>>>>> IPPPrintService#getDefaultAttributeValue() does ---------- if
>>>>>>>> (category == MediaPrintableArea.class) {
>>>>>>>> MediaPrintableArea[] mpas; if ((cps != null) &&
>>>>>>>> ((mpas = *cps.getMediaPrintableArea()*) !=
>>>>>>>> null)) { if (defaultMediaIndex == -1) {
>>>>>>>> // initializes value of defaultMediaIndex
>>>>>>>> getDefaultAttributeValue(Media.class); }
>>>>>>>> return mpas[*defaultMediaIndex*];
>>>>>>>> } -------------- by which mpas[defaultMediaIndex] returns
>>>>>>>> default media printable area. In else block, we instantiate mpa
>>>>>>>> of Letter for US locale and A4 for other locales so I was not
>>>>>>>> checking for null as in both if and else block , I was getting
>>>>>>>> MediaPrintableArea instance. Regarding considering equal
>>>>>>>> margins, I am not sure if there was any way I could get right
>>>>>>>> and left margin separately. Please let me know if there is any
>>>>>>>> way you know of.
>>>>>>> I have modified to do the change in PSPrinterJob. Regarding the
>>>>>>> equal margin, we do the same in
>>>>>>> http://hg.openjdk.java.net/jdk9/client/jdk/file/abb2a39948fe/src/java.desktop/share/classes/sun/print/RasterPrinterJob.java#l688
>>>>>>> http://cr.openjdk.java.net/~psadhukhan/6574279/webrev.01/
>>>>>> That does not seem to be the same case as this. -phil.
>>>>>>> Regards Prasanta
>>>>>>>> Regards Prasanta
>>>>>>>> On 8/4/2016 2:32 AM, Philip Race wrote:
>>>>>>>>> High-level question. Why is this not an over-ride in
>>>>>>>>> PSPrinterJob ? I'm afraid I did not get much past the first
>>>>>>>>> line of the change : 1662 MediaPrintableArea mpa =
>>>>>>>>> 1663 (MediaPrintableArea)getPrintService(). 1664
>>>>>>>>> getDefaultAttributeValue(MediaPrintableArea.class); Read the
>>>>>>>>> docs for MediaPrintableArea :- --- To query for the printable
>>>>>>>>> area, a client must supply a suitable context. Without
>>>>>>>>> specifying at the very least the size of the media being used
>>>>>>>>> no meaningful value for printable area can be obtained. --- ..
>>>>>>>>> not to mention you assume a non-null return. and then you use
>>>>>>>>> new Paper() .. which is a constructor which knows nothing
>>>>>>>>> about the printer. It is *speced* to return US letter with one
>>>>>>>>> inch margins. Another reason why this seems like it should be
>>>>>>>>> a sub-class over-ride. Also this looks wrong ---
>>>>>>>>> 1672 if ((imgX*2) + imgWid > wid) {
>>>>>>>>> 1673 imgWid = wid - imgX*2; 1674 }
>>>>>>>>> You can't assume equal margins. -phil. On 7/12/16, 5:05 AM,
>>>>>>>>> Prasanta Sadhukhan wrote:
>>>>>>>>>> Hi All, Please review a fix for an issue where it is seen
>>>>>>>>>> that if user sets invalid imageablearea via
>>>>>>>>>> Paper.setImageableArea and then calls
>>>>>>>>>> PrinterJob.validatePage() in linux, then it does not get a
>>>>>>>>>> valid pageformat Bug:
>>>>>>>>>> https://bugs.openjdk.java.net/browse/JDK-6574279 webrev:
>>>>>>>>>> http://cr.openjdk.java.net/~psadhukhan/6574279/webrev.00/ Fix
>>>>>>>>>> is to check for margin as noted via CUPS ppdPageSize() and
>>>>>>>>>> then calculate to find the valid margin in validatePaper()
>>>>>>>>>> similar to what we do in native validatePaper() in windows.
>>>>>>>>>> Regards Prasanta
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/2d-dev/attachments/20161104/62dc398d/attachment.html>
More information about the 2d-dev
mailing list