Closed GoogleCodeExporter closed 8 years ago
a bit more information:
- Epubcheck created epub give CRC error on 'mimetype'(but opens ok ?):
testing: mimetype bad CRC 2cab616f (should be 00000000)
- included two brands of epub:
> travail-zip.epub : 'standard' zip construction using zip.exe (Mark Adler's & Co ..)
> travail.epc3.epub: using Epubcheck3 option to create epub zipped file.
(both from same set of files, of course, abd both passing epubceck 3.0 & 1.2 tests)
- see screnshots for differences in zip information
Regards,
Joel Schvartz
Original comment by schvar...@gmail.com
on 20 Oct 2011 at 1:33
Attachments:
Hi Joel,
thanks for finding and reporting this.
My inclination is to disable the save functionality from EPUBCheck, at least
for now. This was added just "for fun" -- EPUBCheck is intended to be a
read-only tool, and we need to focus our resources on the validation aspects.
This might be a simple bug to fix, but even so, moving forward EPUBCheck needs
to remain read-only.
Original comment by markus.g...@gmail.com
on 21 Oct 2011 at 11:24
I found this a great option, if you want to "indutrialize" the build of Epub ...
A standard tool to check and build the release could be very handy.
In case of error, the compressed file could be not created ...
EpubCheck would remain 'read-only' on the source files (uncompressed), of
course.
Which is all that matters, really.
Regards,
Joel
Original comment by schvar...@gmail.com
on 21 Oct 2011 at 2:26
The problem could be that the stored method is used to archive the mimetype
file (first entry of the zip file), while the other files are archived using
the deflated method.
Original comment by biord...@gmail.com
on 23 Oct 2011 at 5:58
No, this is normal (well, at least, the same with other 'good' Epubs ...)
What is different, is the mention of filesystem OS = 'FAT', when others say
NTFS, or UNIX ?
I had the same resulting file on a Solaris 9 platform.
Original comment by schvar...@gmail.com
on 28 Oct 2011 at 2:44
I am still not convinced that this feature is a good idea. Having said that, in
r288 I committed a change to this code that uses the Apache commons-compress
library instead of java.util.zip in order to get control of filename encoding
(this is not supported in JREs lower than 7).
I tried opening the output archives in Sigil, and it seems to work for me (on
OSX). More testing needed, of course.
Original comment by markus.g...@gmail.com
on 31 Oct 2011 at 1:48
Markus,
This version is ok for me too.
Thanks.
Joel
Original comment by schvar...@gmail.com
on 13 Dec 2011 at 9:12
Original comment by markus.g...@gmail.com
on 15 Jan 2012 at 4:31
Original issue reported on code.google.com by
schvar...@gmail.com
on 19 Oct 2011 at 2:35