#92455createFile with file image extension causes ImageMagick identify errors
Description
Hi,
since 9.5.22, creating file with an image extension (“jpeg”, “jpg”, “png”) with Storage->createFile() causes the following errors from ImageMagick “identify” function :
identify: insufficient image data in file `/xxxxxxxx/xxxxxx/375536226116427_1338295523173821.jpeg' @ error/jpeg.c/ReadJPEGImage/1039.
identify: improper image header `/xxxxxxxx/xxxxxx/1675979572678151_2805251863084244.png' @ error/png.c/ReadPNGImage/3954.
This is because Indexing is invoked during the file creation, and Indexer extract width and height but these metadata (via ImageMagick “identify”) don’t exist yet.
As a side question, why indexing at file creation ?
Reviews 3
History 611
-
Helmut Hummel commented 09/30/20, 23:40
-
RVVN no-lastname-given commented 10/01/20, 09:03
We have a scheduler cron job that fetches remote images then stores them in fileadmin.
Images are well created at the end, but we receive emails with the mentioned errors.TYPO3 logs warnings too :
component="TYPO3.CMS.Core.Type.File.ImageInfo": I could not retrieve the image size for file /xxxxxxxx/xxxxxx//381438182400366_776716469539200.jpeg
We use the following sequence :
- storage->createFile(filename.jpeg) => that causes identify errors because of extracting metadata on an image file that doesn’t exist yet
- storage->setFileContents(fileObject, [image data retrieved by cUrl])
Maybe that’s not the right way ?
-
Gerrit Code Review pushed patch set 1 for master to change 65968 status
NewUnder Review 10/01/20, 11:48 -
Helmut Hummel commented 10/01/20, 11:54
RVVN no-lastname-given wrote:
> We have a scheduler cron job that fetches remote images then stores them in fileadmin.
> Images are well created at the end, but we receive emails with the mentioned errors.I’m still puzzled why you get these warning emails.
> TYPO3 logs warnings too :
Yes, that is indeed wrong and unfortunate.
> We use the following sequence :
>
> # storage->createFile(filename.jpeg) => that causes identify errors because of extracting metadata on an image file that doesn’t exist yet
> # storage->setFileContents(fileObject, [image data retrieved by cUrl])
>
> Maybe that’s not the right way ?You could create a local file from CURL data and the use the addFile API instead. I will created a fix though, so that your way will not log warnings any more.
-
Gerrit Code Review pushed patch set 1 for 9.5 to change 65969 10/01/20, 11:57
-
Gerrit Code Review pushed patch set 2 for 9.5 to change 65969 10/01/20, 11:57
-
RVVN no-lastname-given commented 10/01/20, 12:09
Helmut Hummel wrote:
> You could create a local file from CURL data and the use the addFile API instead. I will created a fix though, so that your way will not log warnings any more.Thanks ;)
I’ve already tested the patch and it seems to work.
No warning in log from File.ImageInfo so I guess identify is not called.We’ll have a look at addFile method ;)
-
Gerrit Code Review pushed patch set 2 for master to change 65968 10/14/20, 19:09
-
Gerrit Code Review pushed patch set 1 for 10.4 to change 66125 10/15/20, 20:01
-
Helmut Hummel commented and changed status
Under ReviewResolved done0 %100 % 10/15/20, 20:30Applied in changeset commit:bf49d5eab37c660fce310905e18f8880ac378cae.
-
Benni Mack changed status
ResolvedClosed 12/22/20, 11:05
RVVN no-lastname-given wrote:
> since 9.5.22, creating file with an image extension (“jpeg”, “jpg”, “png”) with Storage->createFile() causes the following errors from ImageMagick “identify” function :
Can you elaborate what exactly the error is you are seeing? How do these error from Imagemagick affect your goals? Are you seeing any exceptions?
> As a side question, why indexing at file creation ?
That is how FAL works. Any file that is added will get an index entry.