Am 10.04.23 um 16:44 schrieb Tim Allison:
https://corpora.tika.apache.org/base/reports/pdfbox-2.0.27-v-2.0.28-rc1.tgz
Thanks Tim!
I had a first look. No new exceptions compared to the last test run but more
than 60 fixed exceptions. Looks like the last fixes did address some other
issues as well.
I seems as if the test run doesn't reveals any reason to stop the release.
WDYT?
Andreas
On Mon, Apr 10, 2023 at 7:02 AM Andreas Lehmkuehler <[email protected]> wrote:
Sounds like one of those expected issues. I guess PDFBox now swallows the former
exception and is able to process the pdf in question. At least the exception is
gone, maybe there is some more content or just an empty page.
However, IMHO that isn't a regression, but an (small) improvement.
@Tim Thanks for running the tests
Andreas
Am 10.04.23 um 12:54 schrieb Tim Allison:
We're getting a build failure on this test now. I'll turn it off for
the build and run the full process.
https://github.com/apache/tika/blob/main/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-pdf-module/src/test/java/org/apache/tika/parser/pdf/PDFParserTest.java#L1031
org.opentest4j.AssertionFailedError: Should have thrown exception ==>
expected: <true> but was: <false>
at
org.junit.jupiter.api.AssertionFailureBuilder.build(AssertionFailureBuilder.java:151)
at
org.junit.jupiter.api.AssertionFailureBuilder.buildAndThrow(AssertionFailureBuilder.java:132)
at org.junit.jupiter.api.AssertTrue.failNotTrue(AssertTrue.java:63)
at org.junit.jupiter.api.AssertTrue.assertTrue(AssertTrue.java:36)
at org.junit.jupiter.api.Assertions.assertTrue(Assertions.java:211)
at
org.apache.tika.parser.pdf.PDFParserTest.testSkipBadPage(PDFParserTest.java:1044)
On Mon, Apr 10, 2023 at 6:41 AM Tim Allison <[email protected]> wrote:
Y. Will start process now. Thank you!
On Mon, Apr 10, 2023 at 6:20 AM Andreas Lehmkuehler <[email protected]> wrote:
Hi,
I've finished the release process and provided a releases candidate for 2.0.28
@Tim Is there any chance to re-run the tests in the next 3 days, so that we
could stop the release if there is any major regression.
I don't expect any new issue as the last changes should produces less exceptions
than before but you knows ....
Thanks in advance
Andreas
Am 10.04.23 um 11:42 schrieb Andreas Lehmkuehler:
Am 10.04.23 um 04:32 schrieb Tilman Hausherr:
On 09.04.2023 22:36, Andreas Lehmkuehler wrote:
OK, so there is one more question left: do we need to re-run the tests before
starting the release process?
Yes I prefer to have another comparison, but it can be done in parallel.
Good idea, I'm going to cut the release ...
Andreas
Tilman
Andreas
Am 09.04.23 um 20:56 schrieb Tilman Hausherr:
On 09.04.2023 17:35, Andreas Lehmkuehler wrote:
Hi,
I've fixed the issue with 2 of the 3 pdfs.
GHOSTSCRIPT-702891-0.pdf is left as the only problematic pdf. I didn't
found a solution which fixes the regressions and still fixes the origin
issue from PDFBOX-5178. The parser from the trunk is able to handle that
pdf well.
IMHO we should leave it alone, as it is malformed anmd doesn't contain any
useful content. More important, it is one pdf out of hundreds of
thoudsands, just a corner cases.
WDYT?
I agree!
Tilman
Andreas
Am 05.04.23 um 08:10 schrieb Andreas Lehmkuehler:
Am 04.04.23 um 07:40 schrieb Andreas Lehmkuehler:
Am 03.04.23 um 19:50 schrieb Tim Allison:
https://corpora.tika.apache.org/base/reports/pdfbox-2.0.27-v-2.0.28-20230403-reports.tgz
Haven't had a chance to take a look yet. :(
Thanks Tim!
There are still 5 new exceptions listed. All of them are related to the
very same change coming from PDFBOX-5178 which I've fixed the other day.
But these cases are different and the trunk is affected as well. My bad
to not have a deeper look in the first place.
I'm going to investigate those issues
All pdfs are more or less broken. Two of them are totally useless and the
new exception is just another one. The other three contain some more or
less readable content and we are hitting the well know dilemma: should we
stop reading once we hit something bad or should we try to read as much as
possible and maybe run into much bigger issues than before.
I guess these are all some special corner cases. I'm still thinking about
a solution to support both strategies.
Andreas
Andreas
On Mon, Apr 3, 2023 at 6:53 AM Tilman Hausherr <[email protected]>
wrote:
Don't wait please
Thanks
Tilman
--- Original-Nachricht ---
Von: Tim Allison
Betreff: Re: Fwd: 2.0.28 release?
Datum: 03. April 2023, 12:47
An: [email protected]
Y. I can kick that off now. Or should I wait?
On Sat, Apr 1, 2023 at 2:06 PM Andreas Lehmkuehler <[email protected]
<mailto:[email protected]> > wrote:
@Tim <mailto:@Tim>
Is there any chance to re-run the tests?
Andreas
Am 01.04.23 um 17:08 schrieb Andreas Lehmkuehler:
Am 01.04.23 um 17:05 schrieb Andreas Lehmkuehler:
I've accidentally send this to Tim only :-|
-------- Weitergeleitete Nachricht --------
Betreff: Re: 2.0.28 release?
Datum: Fri, 31 Mar 2023 07:50:10 +0200
Von: Andreas Lehmkuehler <[email protected] <mailto:[email protected]>
An: Tim Allison <[email protected] <mailto:[email protected]> >
Am 30.03.23 um 16:27 schrieb Tim Allison:
Reports are here:
<https://corpora.tika.apache.org/base/reports/pdfbox-2.0.27-v-2.0.28-SNAPSHOT.tgz>
Thanks Tim.
Looks like we have a regression. There is a handful of new exceptions.
Some of
them just replace another exception and it is unclear if the result is
better
or worse. But at least one of the pdfs works in 2.0.27 and doesn't in
2.0.28
bug_trackers/PDFBOX/PDFBOX-4424-1.pdf
I'll have a look
The regression was related to PDFBOX-5178. I've fixed it so that the
exceptions
should be gone.
Andreas
Andreas
On Tue, Mar 28, 2023 at 10:42 PM Tilman Hausherr <
[email protected] <mailto:[email protected]> > wrote:
Yes please!
Thanks
Tilman
On 28.03.2023 19:22, Tim Allison wrote:
+1
Should I run the regression tests now or is there anything else
text
related that is still being worked on?
On Tue, Mar 28, 2023 at 1:05 PM Tilman Hausherr <
[email protected] <mailto:[email protected]> > wrote:
+1
Tilman
On 28.03.2023 08:46, Andreas Lehmkuehler wrote:
Hi,
how about cutting a 2.0.28 release next week on Monday?
there is a bunch of solved tickets and the last release dates
back 6
months
Andreas
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
