Summary: | [PATCH] CSS scanner does not tokenize correctly vendor specific CSS properties (such as -moz-border-radius) | ||
---|---|---|---|
Product: | Batik - Now in Jira | Reporter: | Nicolas Garnier <nivco> |
Component: | CSS | Assignee: | Batik Developer's Mailing list <batik-dev> |
Status: | NEW --- | ||
Severity: | normal | Keywords: | PatchAvailable |
Priority: | P2 | ||
Version: | 1.8 | ||
Target Milestone: | --- | ||
Hardware: | PC | ||
OS: | Linux | ||
Attachments: |
Corrected Scanner.java
PATCH file for Scanner.java PATCH file for Scanner.java |
(In reply to comment #0) > Created an attachment (id=24226) [details] > Corrected Scanner.java Thanks for the report and the correction proposal. :-) > Corrected version of the Scanner.java file is attached. (Also about adding the "[PATCH]" prefix...) Note that attaching a patched file is not the same as attaching a patch. ;-) I've tried to merge your file with the current trunk code but, given you provided no information about the Batik release where the file was taken, the process can harder. Please provide this information (Batik release and/or revision where this file was taken from) or attach a "real" patch (even better, if you are able to). Created attachment 24230 [details]
PATCH file for Scanner.java
Hey Helder, Thanks a lot for looking into this! Sorry about not posting a patch file, it was pretty late last night and I really had to got to bed :D The batik release I worked on is Batik 1.6 (checkout out from the tags tree) I attached a PATCH file this time. Please tell me if any regressions occurred or any other pbs. (In reply to comment #3) > The batik release I worked on is Batik 1.6 (checkout out from the tags tree) This is a pretty outdated version... :-| > Please tell me if any regressions occurred or any other pbs. Regressions in which sense? Typically, one tries to assure no regressions are made. What may be happening is that you fix is already included in the current trunk version (1.8, as of this writing). I'd invite you to compare the trunk source with the version 1.6 code and try to attach a patch to the current trunk version: as far as I know, there is no work in legacy versions. At least, that's what I would to in order to try integrating your bug fix into the trunk code... ;-) I may be able to get my hands into this, but I currently can't predict if nor when. Hey Helder, Sorry about the delay, I'm a little swamped at work but as soon as I can get some time I'll submit a patch for batik 1.8 (that should not be a pb). The bug still seems to be there in 1.8 but, just in case, I'll run some more tests when as soon as that crazy period at work is over :) Cheers! Created attachment 24359 [details]
PATCH file for Scanner.java
Hey Helder,
Here is the PATCH file for Batik 1.8
(The bug was still there in Batik 1.8)
I hope this one will be the good one :D
Cheers!
Thanks again for your time
(In reply to comment #6) > Here is the PATCH file for Batik 1.8 > (The bug was still there in Batik 1.8) Updated target version accordingly. > I hope this one will be the good one :D Thanks, Nicolas! :-) The patch now needs to be reviewed/committed by someone commit rights. Thomas, could you take a look at this and/or detail on further steps which are needed (regard test cases, maybe?) before this can be landed? |
Created attachment 24226 [details] Corrected Scanner.java CSS properties such as -moz-border-radius or -webkit-border-radius (all properties starting with '-') are not tokenized correctly by the CSS scanner. Instead of being tokenized as LexicalUnits.IDENTIFIER the first character is tokenized as LexicalUnits.MINUS and the rest of the properties fails to tokenize. Corrected version of the Scanner.java file is attached.