|
| 1 | +1.9.5-8: https://bitbucket.org/gthomas/iptcinfo/issue/4/file-permissions-for-changed-files-are-not - copy original file's permission bits on save/saveAs |
| 2 | + |
| 3 | +1.9.5-7: https://bitbucket.org/gthomas/iptcinfo/issue/3/images-w-o-iptc-data-should-not-log-errors - have silencable parse errors. |
| 4 | + |
| 5 | +1.9.5-6: to have a nice new upload (seems easy_install grabs an old version). |
| 6 | + |
| 7 | +1.9.5-5: fix some issues with "super" |
| 8 | + |
| 9 | +1.9.5-3: use logging module. |
| 10 | + |
| 11 | +1.9.5-2: Emil Stenström pinpointed some bugs/misleading (un)comments |
| 12 | + Also a new (mis)feature is implemented: if you don't specify inp_charset |
| 13 | + (and the image misses such information, too) than no conversion is made |
| 14 | + to unicode, everything stays bytestring! |
| 15 | + This way you don't need to deal with charsets, BUT it is your risk to make |
| 16 | + the modifications with the SAME charset as it is in the image! |
| 17 | + |
| 18 | +1.9.5-1: getting in sync with the Perl version 1.9.5 |
| 19 | + |
| 20 | +1.9.2-rc8: |
| 21 | + charset recognition loosened (failed with some image out of |
| 22 | + Adobe Lightroom). |
| 23 | + |
| 24 | +1.9.2-rc7: NOT READY |
| 25 | + IPTCInfo now accepts 'inp_charset' keyword for setting input charset. |
| 26 | + |
| 27 | +1.9.2-rc6: just PyLint-ed out some errors. |
| 28 | + |
| 29 | +1.9.2-rc5: Amos Latteier sent me a patch which releases the requirement of the |
| 30 | + file objects to be file objects (he uses this on jpeg files stored in |
| 31 | + databases as strings). |
| 32 | + It modifies the module in order to look for a read method on the file |
| 33 | + object. If one exists it assumes the argument is a file object, otherwise it |
| 34 | + assumes it's a filename. |
| 35 | + |
| 36 | +1.9.2-rc4: on Windows systems, tmpfile may not work correctly - now I use |
| 37 | + cStringIO on file save (to save the file without truncating it on Exception). |
| 38 | + |
| 39 | +1.9.2-rc3: some little bug fixes, some safety enhancements (now iptcinfo.py |
| 40 | + will overwrite the original image file (info.save()) only if everything goes |
| 41 | + fine (so if an exception is thrown at writing, it won't cut your original |
| 42 | + file). |
| 43 | + |
| 44 | + This is a pre-release version: needs some testing, and has an unfound bug |
| 45 | + (yet): some pictures can be enhanced with iptc data, and iptcinfo.py is able |
| 46 | + to read them, but some other iptc data readers will spit on it. |
| 47 | + |
| 48 | +1.9.1: a first release with some little encoding support |
| 49 | + |
| 50 | + The class IPTCInfo now has an inp_charset and an out_charset attribute - the |
| 51 | + first is the read image's charset (defaults to the system default charset), |
| 52 | + the second is the charset the writer will use (defaults to inp_charset). |
| 53 | + |
| 54 | + Reader will find the charset included in IPTC data (if any, defaults to the |
| 55 | + system's default charset), and use it to read to unicode strings. Writer will |
| 56 | + write using IPTCinfo.out_charset (if it is not set, will not write charset |
| 57 | + IPTC record). |
| 58 | + |
| 59 | + With this, it is possible to read and write i18n strings correctly. |
| 60 | + |
| 61 | + I haven't tested this functionality thoroughly, and that little test was only |
| 62 | + on my WinXP box only, with the only other IPTC reader: IrfanView. |
0 commit comments