You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@pdfbox.apache.org by Andreas Lehmkühler <an...@lehmi.de.INVALID> on 2024/07/30 05:43:06 UTC
PDFBox 3.0.3 release
Hi,
2.0.32 is out of the door and I'd like to cut the next 3.0 release. How
about next Monday?
WDYT? Any objections?
Andreas
P.S.: I've to fix the build process first, but I'm almost done. Maybe we
have to redo the 2.0 release as there is an issue with the checksums,
which I created manually due to the broken last step of the build
process. :-(
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
For additional commands, e-mail: dev-help@pdfbox.apache.org
Re: PDFBox 3.0.3 release
Posted by Andreas Lehmkühler <an...@lehmi.de.INVALID>.
Thanks for the fast feedback. I'm going to cut the release now.
Hopefully it'll work this time :-o
Andreas
Am 05.08.24 um 14:24 schrieb Tilman Hausherr:
> https://home.snafu.de/tilman/tmp/reports_pdfbox_3.0.2_vs_3.0.3_3.tar.xz
>
> The new exceptions and differences I had a closer look on are all
> because of Flate decompression errors, so I'd say go ahead.
>
> Tilman
>
> On 05.08.2024 09:34, Tilman Hausherr wrote:
>>>
>>> @Tilman
>>> Are there any issues with the text extraction, otherwise I'm going to
>>> cut the release this evening.
>>
>> The text extraction differences seem all to be related to the Flate
>> decompression errors, I'm gonna do another test that will hopefully be
>> finished in time.
>>
>> Tilman
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
For additional commands, e-mail: dev-help@pdfbox.apache.org
Re: PDFBox 3.0.3 release
Posted by Tilman Hausherr <TH...@t-online.de>.
https://home.snafu.de/tilman/tmp/reports_pdfbox_3.0.2_vs_3.0.3_3.tar.xz
The new exceptions and differences I had a closer look on are all
because of Flate decompression errors, so I'd say go ahead.
Tilman
On 05.08.2024 09:34, Tilman Hausherr wrote:
>>
>> @Tilman
>> Are there any issues with the text extraction, otherwise I'm going to
>> cut the release this evening.
>
> The text extraction differences seem all to be related to the Flate
> decompression errors, I'm gonna do another test that will hopefully be
> finished in time.
>
> Tilman
Re: PDFBox 3.0.3 release
Posted by Tilman Hausherr <TH...@t-online.de>.
On 05.08.2024 08:19, Andreas Lehmkühler wrote:
>
>
> Am 03.08.24 um 05:48 schrieb Tilman Hausherr:
>> I thought I had posted the link, but it seems I didn't?! Here it is
>> https://home.snafu.de/tilman/tmp/reports_pdfbox_3.0.2_vs_3.0.3_2.tar.xz
> I had a look at the new exception. The changes from [1] are
> responsible for the different behavior. IMHO that issue can be
> ignored. The file is a mess and 3.0.2 isn't able to extract anything.
> 3.0.3 simply produces another error message.
>
> @Tilman
> Are there any issues with the text extraction, otherwise I'm going to
> cut the release this evening.
The text extraction differences seem all to be related to the Flate
decompression errors, I'm gonna do another test that will hopefully be
finished in time.
Tilman
>
> Andreas
>
> [1] https://issues.apache.org/jira/browse/PDFBOX-5786
>>
>> Tilman
>>
>> On 01.08.2024 10:47, Tilman Hausherr wrote:
>>> Thanks, I'll run another "B" and eval job but with the change from
>>> PDFBOX-5790 reverted, like I did for 2.0.32, so we get less noise in
>>> the content results.
>>>
>>> Tilman
>>>
>>> On 01.08.2024 07:56, Andreas Lehmkühler wrote:
>>>>
>>>>
>>>> Am 31.07.24 um 11:45 schrieb Tilman Hausherr:
>>>>> On 31.07.2024 06:47, Andreas Lehmkühler wrote:
>>>>>> Bad news is there are a lot of new exceptions. Good news is, it
>>>>>> looks like they are all the same.
>>>>>>
>>>>>> I'd a quick look and it seems to be related to [1]. I've tested
>>>>>> some of the pdfs and they all contain corrupt streams. I guess
>>>>>> the issue is a different exception handling in such cases. 3.0.2
>>>>>> catches such exceptions when reading corrupt streams and 3.0.3
>>>>>> seems to struggle and stops the parsing process.
>>>>>
>>>>> Not the ones at the bottom of new_exceptions_in_B_details.xlsx
>>>>> (e.g. 500436.pdf), although these do also seem to be related to
>>>>> PDFBOX-5675 too. They do a rewind() near the end of the stream.
>>>> I've found a fix for both cases. Please rerun the tests whenever
>>>> you have some cycles to do so.
>>>>
>>>> Thanks in advance
>>>>
>>>>
>>>>>
>>>>> Tilman
>>>>>
>>>>>
>>>>>
>>>>>>
>>>>>>
>>>>>> Andreas
>>>>>>
>>>>>> [1] https://issues.apache.org/jira/browse/PDFBOX-5675
>>>>>>
>>>>>>>
>>>>>>> Tilman
>>>>>>>
>>>>>>>
>>>>>>> ---------------------------------------------------------------------
>>>>>>>
>>>>>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>>>>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>>>>>
>>>>>>
>>>>>> ---------------------------------------------------------------------
>>>>>>
>>>>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>>>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>>>>
>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>>
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
> For additional commands, e-mail: dev-help@pdfbox.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
For additional commands, e-mail: dev-help@pdfbox.apache.org
Re: PDFBox 3.0.3 release
Posted by Andreas Lehmkühler <an...@lehmi.de.INVALID>.
Am 03.08.24 um 05:48 schrieb Tilman Hausherr:
> I thought I had posted the link, but it seems I didn't?! Here it is
> https://home.snafu.de/tilman/tmp/reports_pdfbox_3.0.2_vs_3.0.3_2.tar.xz
I had a look at the new exception. The changes from [1] are responsible
for the different behavior. IMHO that issue can be ignored. The file is
a mess and 3.0.2 isn't able to extract anything. 3.0.3 simply produces
another error message.
@Tilman
Are there any issues with the text extraction, otherwise I'm going to
cut the release this evening.
Andreas
[1] https://issues.apache.org/jira/browse/PDFBOX-5786
>
> Tilman
>
> On 01.08.2024 10:47, Tilman Hausherr wrote:
>> Thanks, I'll run another "B" and eval job but with the change from
>> PDFBOX-5790 reverted, like I did for 2.0.32, so we get less noise in
>> the content results.
>>
>> Tilman
>>
>> On 01.08.2024 07:56, Andreas Lehmkühler wrote:
>>>
>>>
>>> Am 31.07.24 um 11:45 schrieb Tilman Hausherr:
>>>> On 31.07.2024 06:47, Andreas Lehmkühler wrote:
>>>>> Bad news is there are a lot of new exceptions. Good news is, it
>>>>> looks like they are all the same.
>>>>>
>>>>> I'd a quick look and it seems to be related to [1]. I've tested
>>>>> some of the pdfs and they all contain corrupt streams. I guess the
>>>>> issue is a different exception handling in such cases. 3.0.2
>>>>> catches such exceptions when reading corrupt streams and 3.0.3
>>>>> seems to struggle and stops the parsing process.
>>>>
>>>> Not the ones at the bottom of new_exceptions_in_B_details.xlsx (e.g.
>>>> 500436.pdf), although these do also seem to be related to
>>>> PDFBOX-5675 too. They do a rewind() near the end of the stream.
>>> I've found a fix for both cases. Please rerun the tests whenever you
>>> have some cycles to do so.
>>>
>>> Thanks in advance
>>>
>>>
>>>>
>>>> Tilman
>>>>
>>>>
>>>>
>>>>>
>>>>>
>>>>> Andreas
>>>>>
>>>>> [1] https://issues.apache.org/jira/browse/PDFBOX-5675
>>>>>
>>>>>>
>>>>>> Tilman
>>>>>>
>>>>>>
>>>>>> ---------------------------------------------------------------------
>>>>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>>>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>>>
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
> For additional commands, e-mail: dev-help@pdfbox.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
For additional commands, e-mail: dev-help@pdfbox.apache.org
Re: PDFBox 3.0.3 release
Posted by Tilman Hausherr <TH...@t-online.de>.
I thought I had posted the link, but it seems I didn't?! Here it is
https://home.snafu.de/tilman/tmp/reports_pdfbox_3.0.2_vs_3.0.3_2.tar.xz
Tilman
On 01.08.2024 10:47, Tilman Hausherr wrote:
> Thanks, I'll run another "B" and eval job but with the change from
> PDFBOX-5790 reverted, like I did for 2.0.32, so we get less noise in
> the content results.
>
> Tilman
>
> On 01.08.2024 07:56, Andreas Lehmkühler wrote:
>>
>>
>> Am 31.07.24 um 11:45 schrieb Tilman Hausherr:
>>> On 31.07.2024 06:47, Andreas Lehmkühler wrote:
>>>> Bad news is there are a lot of new exceptions. Good news is, it
>>>> looks like they are all the same.
>>>>
>>>> I'd a quick look and it seems to be related to [1]. I've tested
>>>> some of the pdfs and they all contain corrupt streams. I guess the
>>>> issue is a different exception handling in such cases. 3.0.2
>>>> catches such exceptions when reading corrupt streams and 3.0.3
>>>> seems to struggle and stops the parsing process.
>>>
>>> Not the ones at the bottom of new_exceptions_in_B_details.xlsx (e.g.
>>> 500436.pdf), although these do also seem to be related to
>>> PDFBOX-5675 too. They do a rewind() near the end of the stream.
>> I've found a fix for both cases. Please rerun the tests whenever you
>> have some cycles to do so.
>>
>> Thanks in advance
>>
>>
>>>
>>> Tilman
>>>
>>>
>>>
>>>>
>>>>
>>>> Andreas
>>>>
>>>> [1] https://issues.apache.org/jira/browse/PDFBOX-5675
>>>>
>>>>>
>>>>> Tilman
>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>>
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
> For additional commands, e-mail: dev-help@pdfbox.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
For additional commands, e-mail: dev-help@pdfbox.apache.org
Re: PDFBox 3.0.3 release
Posted by Tilman Hausherr <TH...@t-online.de>.
Thanks, I'll run another "B" and eval job but with the change from
PDFBOX-5790 reverted, like I did for 2.0.32, so we get less noise in the
content results.
Tilman
On 01.08.2024 07:56, Andreas Lehmkühler wrote:
>
>
> Am 31.07.24 um 11:45 schrieb Tilman Hausherr:
>> On 31.07.2024 06:47, Andreas Lehmkühler wrote:
>>> Bad news is there are a lot of new exceptions. Good news is, it
>>> looks like they are all the same.
>>>
>>> I'd a quick look and it seems to be related to [1]. I've tested some
>>> of the pdfs and they all contain corrupt streams. I guess the issue
>>> is a different exception handling in such cases. 3.0.2 catches such
>>> exceptions when reading corrupt streams and 3.0.3 seems to struggle
>>> and stops the parsing process.
>>
>> Not the ones at the bottom of new_exceptions_in_B_details.xlsx (e.g.
>> 500436.pdf), although these do also seem to be related to PDFBOX-5675
>> too. They do a rewind() near the end of the stream.
> I've found a fix for both cases. Please rerun the tests whenever you
> have some cycles to do so.
>
> Thanks in advance
>
>
>>
>> Tilman
>>
>>
>>
>>>
>>>
>>> Andreas
>>>
>>> [1] https://issues.apache.org/jira/browse/PDFBOX-5675
>>>
>>>>
>>>> Tilman
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
> For additional commands, e-mail: dev-help@pdfbox.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
For additional commands, e-mail: dev-help@pdfbox.apache.org
Re: PDFBox 3.0.3 release
Posted by Andreas Lehmkühler <an...@lehmi.de.INVALID>.
Am 31.07.24 um 11:45 schrieb Tilman Hausherr:
> On 31.07.2024 06:47, Andreas Lehmkühler wrote:
>> Bad news is there are a lot of new exceptions. Good news is, it looks
>> like they are all the same.
>>
>> I'd a quick look and it seems to be related to [1]. I've tested some
>> of the pdfs and they all contain corrupt streams. I guess the issue is
>> a different exception handling in such cases. 3.0.2 catches such
>> exceptions when reading corrupt streams and 3.0.3 seems to struggle
>> and stops the parsing process.
>
> Not the ones at the bottom of new_exceptions_in_B_details.xlsx (e.g.
> 500436.pdf), although these do also seem to be related to PDFBOX-5675
> too. They do a rewind() near the end of the stream.
I've found a fix for both cases. Please rerun the tests whenever you
have some cycles to do so.
Thanks in advance
>
> Tilman
>
>
>
>>
>>
>> Andreas
>>
>> [1] https://issues.apache.org/jira/browse/PDFBOX-5675
>>
>>>
>>> Tilman
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
> For additional commands, e-mail: dev-help@pdfbox.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
For additional commands, e-mail: dev-help@pdfbox.apache.org
Re: PDFBox 3.0.3 release
Posted by Tilman Hausherr <TH...@t-online.de>.
On 31.07.2024 06:47, Andreas Lehmkühler wrote:
> Bad news is there are a lot of new exceptions. Good news is, it looks
> like they are all the same.
>
> I'd a quick look and it seems to be related to [1]. I've tested some
> of the pdfs and they all contain corrupt streams. I guess the issue is
> a different exception handling in such cases. 3.0.2 catches such
> exceptions when reading corrupt streams and 3.0.3 seems to struggle
> and stops the parsing process.
Not the ones at the bottom of new_exceptions_in_B_details.xlsx (e.g.
500436.pdf), although these do also seem to be related to PDFBOX-5675
too. They do a rewind() near the end of the stream.
Tilman
>
>
> Andreas
>
> [1] https://issues.apache.org/jira/browse/PDFBOX-5675
>
>>
>> Tilman
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
>> For additional commands, e-mail: dev-help@pdfbox.apache.org
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
> For additional commands, e-mail: dev-help@pdfbox.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
For additional commands, e-mail: dev-help@pdfbox.apache.org
Re: PDFBox 3.0.3 release
Posted by Andreas Lehmkühler <an...@lehmi.de.INVALID>.
Am 30.07.24 um 21:11 schrieb Tilman Hausherr:
> regression test reports:
Thanks for running the tests
>
> https://home.snafu.de/tilman/tmp/reports_pdfbox_3.0.2_vs_3.0.3.tar.xz
>
> I haven't investigated the details, but it looks like work to do :-(
Bad news is there are a lot of new exceptions. Good news is, it looks
like they are all the same.
I'd a quick look and it seems to be related to [1]. I've tested some of
the pdfs and they all contain corrupt streams. I guess the issue is a
different exception handling in such cases. 3.0.2 catches such
exceptions when reading corrupt streams and 3.0.3 seems to struggle and
stops the parsing process.
Andreas
[1] https://issues.apache.org/jira/browse/PDFBOX-5675
>
> Tilman
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
> For additional commands, e-mail: dev-help@pdfbox.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
For additional commands, e-mail: dev-help@pdfbox.apache.org
Re: PDFBox 3.0.3 release
Posted by Tilman Hausherr <TH...@t-online.de>.
regression test reports:
https://home.snafu.de/tilman/tmp/reports_pdfbox_3.0.2_vs_3.0.3.tar.xz
I haven't investigated the details, but it looks like work to do :-(
Tilman
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
For additional commands, e-mail: dev-help@pdfbox.apache.org
Re: PDFBox 3.0.3 release
Posted by Tilman Hausherr <TH...@t-online.de>.
+1
Tilman
On 30.07.2024 07:43, Andreas Lehmkühler wrote:
> Hi,
>
> 2.0.32 is out of the door and I'd like to cut the next 3.0 release.
> How about next Monday?
>
> WDYT? Any objections?
>
> Andreas
>
> P.S.: I've to fix the build process first, but I'm almost done. Maybe
> we have to redo the 2.0 release as there is an issue with the
> checksums, which I created manually due to the broken last step of the
> build process. :-(
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
> For additional commands, e-mail: dev-help@pdfbox.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@pdfbox.apache.org
For additional commands, e-mail: dev-help@pdfbox.apache.org